It is the composition root that runs nine formal-math components and folds their outputs into one public record labelling every row by origin.verifier_lab_kernel is the composition root for the formal-math verifier lab. It runs or consumes nine smaller formal-math components, then folds their outputs into one public result record that labels every row by where it came from.
\n
Not a prover or benchmark runner, it reads a public packet, drives the stack, and sorts evidence into claim buckets, emitting references and verdicts without bodies. It is not a theorem prover, a benchmark runner, a private Lean import, or a frontend surface. It reads a public problem packet, drives the component stack, sorts each piece of evidence into a fixed set of claim buckets, and writes result records that carry references, hashes, counts, and verdicts. The proof text, provider output, oracle answers, and command stdout stay out of what it emits.
\n
The point is separation: one record shows which claims a Lean checker backed, which are provider hints, which are oracle comparisons, and which were rejected. The point is separation. One record shows which claims a Lean checker actually backed, which are provider hints, which are oracle comparisons, and which rows were thrown out for breaking the rules, instead of a pile of separate outputs taken on faith.
\n
Purpose
\n
How progress blurs in a proof loop
\n
In a proof loop a checker sign-off, an oracle match, a provider tactic, and a retrieval hit all look like progress, quietly inflating the count of theorems proved. In a formal-math agent loop several different things look like progress. A Lean checker can accept a term. An oracle holding a hindsight answer can say a candidate matches. A provider model can offer a plausible next tactic. A retrieval step can return a premise. Treated loosely, all of these blur into a single sense of "it worked", and oracle or provider success quietly inflates the count of theorems actually proved.
\n
One authority-class question per row
\n
It stops that blur by asking one question per evidence row, which authority class it belongs to, so a passing component cannot lend its standing elsewhere. This component exists to stop that blur. It answers one question for each row of evidence: which authority class does it belong to, and what may that class claim? A passing component cannot lend its standing to a different bucket.
\n
How it works
\n
Two entry points and the CLI
\n
run drives the fixture with five negative cases, run_kernel_bundle consumes an exported bundle as a synthetic contract, and main is the CLI over both. Two entry points run the kernel. run drives the first-wave fixture with the five negative cases included. run_kernel_bundle drives an exported bundle, where the component result records are consumed as a declared synthetic contract rather than a live run. main is the CLI over both, taking run or run-kernel-bundle with --input and --out.
\n
Both entry points delegate to _build_result, which runs the stages in order: Both entry points delegate to _build_result, which runs the stages in order:
\n
_load_payloads reads the public packet, the optional proof-lab route file, and any negative-case inputs from the input directory.
scan_paths, seeded by load_forbidden_classes, scans every input and copied source file for forbidden classes. A blocking hit fails the run.
_run_component_stack runs each of the nine components on its own fixture (or _standalone_exported_component_stack consumes bundle result records). It normalizes each component record to strip legacy redaction fields and public path roots. The stack passes only when every component status passes.
_validate_proof_lab_route checks that the declared route resolves to formal_prover_context_strategy_gate and maps to the expected component set.
_validate_packet runs the five typed row validators (below) over forward problems, oracle structured source record, provider hypotheses, CP2 candidates, and Evolve candidates. Each violation becomes a finding with an error code.
_claim_separation sorts the validated rows into the seven buckets.
\n
The nine components in the stack
\n
The stack runs nine formal-math components spanning corpus readiness, premise retrieval, tactic routing, trace repair, diagnostics, and the Lean proof witness. The nine components are corpus_readiness_mathlib_absence_gate, lean_std_premise_index, formal_math_premise_retrieval, tactic_portfolio_availability_probe, target_shape_tactic_routing_gate, ring2_premise_retrieval_precision_recall_harness, formal_math_verifier_trace_repair_loop, proof_diagnostic_evidence_spine, and formal_math_lean_proof_witness.
\n
_claim_separation sorts into seven buckets
\n
_claim_separation sorts validated rows into exactly seven buckets, with lean_verified carrying the proof-witness count and contract_rejected holding the observed negative cases._claim_separation produces exactly seven buckets: lean_verified, provider_suggested, oracle_compared, contract_rejected, retrieval_miss, cp2_translated, and evolve_candidate. The lean_verified bucket carries the proof-witness component status and its compiled declaration count. The contract_rejected bucket carries the observed negative cases, marked as contract violations, not results.
\n
Counters, not description, enforce the boundary
\n
The boundary is enforced by counters that read zero on a clean run, and the status passes only when the scan, component stack, packet validation, and proof-lab route all pass. The boundary is enforced by counters, not description. _build_result computes oracle_forward_success_increment_count (oracle rows that marked themselves as forward success), provider_results_counted (provider rows claiming proof authority), and proof_body_export_count (forward or CP2 rows carrying a proof body). On a clean run these read zero. The final status is a pass only when the secret scan has zero blocking hits, the component stack passes, packet validation passes, and the proof-lab route passes.
\n
_write_receipts emits metadata-only records
\n
_write_receipts emits the public records with body_in_receipt false, the fixture path marking a real runtime result record and the bundle path marking itself a synthetic contract._write_receipts then emits the public records. The fixture path writes a result record, a board, a validation record, and a fixture sign-off record; the bundle path writes one bundle validation result. Every record sets body_in_receipt to false and carries receipt_transparency_contract, whose omitted_payload_scope is proof, provider, oracle, private-source, and stdout/stderr bodies only. The fixture path marks real_runtime_receipt true on a pass; the bundle path marks itself a synthetic contract and not a live run.
\n
Function
Role
run
Fixture entry point; drives the packet plus five negative cases.
run_kernel_bundle
Exported-bundle entry point; consumes bundle result records as a synthetic contract.
_build_result
Loads inputs, scans for leaks, runs the component stack, validates rows, sorts buckets, sets status.
_run_component_stack
Runs the nine components and normalizes their records.
_claim_separation
Sorts validated rows into the seven claim buckets.
_write_receipts
Writes the metadata-only public result records.
\nDiagram of the mechanism (6 steps).Diagram source & refs
flowchart TD packet["Public packet plus 5 negative cases"] build["_build_result load / scan / validate"] stack["_run_component_stack 9 formal-math components"] buckets["_claim_separation 7 claim buckets"] counters["Authority counters oracle / provider / proof-body = 0"] result records["_write_receipts metadata-only result records"] packet --> build build --> stack stack --> buckets build --> counters buckets --> result records counters --> result records
\n
Negative cases
\n
The fixture carries five named inputs the kernel must reject, each handled by one typed row validator: The fixture carries five named inputs the kernel must reject, each handled by one typed row validator:
\n
forward_problem_leaks_candidate_body: a forward problem carrying a candidate, ideal, repair, oracle-premise, source-proof, or base-index field. _validate_forward_problems records VERIFIER_LAB_FORWARD_FIELD_FORBIDDEN.
oracle_counted_as_forward: an oracle structured source record with counted_as_forward_success set true. _validate_oracle_sidecars records VERIFIER_LAB_ORACLE_FORWARD_CONTAMINATION.
provider_claims_proof: a provider hypothesis asserting proof authority. _validate_provider_hypotheses records VERIFIER_LAB_PROVIDER_PROOF_AUTHORITY_FORBIDDEN.
cp2_candidate_contains_proof_body: a CP2 row carrying a proof body, raw tactic script, provider body, or oracle template. _validate_cp2_candidates records VERIFIER_LAB_CP2_PROOF_BODY_FORBIDDEN, and it also rejects action classes outside the bounded vocabulary and candidates with no disconfirmation test.
evolve_mutates_unbounded_artifact: an Evolve row touching an artifact outside the bounded policy set, or asking for arbitrary code mutation. _validate_evolve_candidates records VERIFIER_LAB_EVOLVE_SCOPE_FORBIDDEN, and it also flags candidates that cite no baseline result record.
\n
Each rejected row lands in contract_rejected as a contract violation, never in a claim bucket. Each rejected row lands in contract_rejected as a contract violation, never in a claim bucket.
\n
Prior Art Grounding
\n
LCF composition and LeanDojo boundaries
\n
It follows LCF-style kernel composition that keeps checked results apart from automation, and LeanDojo's lesson that retrieval and provider hints need explicit boundaries. This component follows small-kernel theorem-proving and proof-certificate composition. The LCF approach and HOL Light anchor the idea that a verifier lab keeps trusted checked results apart from heuristics and automation. Lean-oriented work such as LeanDojo adds the agent context: retrieval, provider hypotheses, and proof-state interaction need explicit boundaries before they can influence a proof claim. The kernel borrows the composition discipline and keeps verifier, oracle, provider, CP2, and Evolve rows in separate buckets with separate authority.
A pass means the stack runs, the five negative cases are observed, seven buckets hold, counters stay at zero, and no record leaks a body or local root. A pass means the component stack runs, the five negative cases are observed, claim_separation holds exactly the seven buckets, the authority counters stay at zero, and no public record carries a body or a leaked local root.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The strongest claim is bucket separation and metadata-only containment across the composed records, with any return code bound to its own fixture, not generalizable to benchmarks. The strongest claim this component supports is bucket separation and metadata-only containment across the composed public component records: a reader can see which row a Lean checker backed, which is an oracle comparison, which is a provider hint, which is a CP2 or Evolve candidate, and which was rejected. A Lean/Lake return code or compiled-declaration count is evidence for the public fixture or exported bundle it came from, not a license to generalize to arbitrary benchmarks.
\n
What the scope limit refuses
\n
Past that line oracle rows stay comparators, provider rows stay advisory, CP2 and Evolve rows stay bounded, and nothing authorizes Mathlib import, benchmark rates, or launch. That is the proof boundary. The scope limit refuses everything past it. Oracle structured source record stay hindsight comparators and cannot increment forward success. Provider hypotheses stay advisory until a verifier record exists. CP2 rows stay typed action candidates, bounded evidence bodies. Evolve rows stay bounded policy artifacts, not source-file changes. The component does not import private proof bodies, does not create Mathlib import authority, does not claim a benchmark solve rate, and excludes public sharing, hosted deployment, launch, or secret export.
\n",
"components": [
{
"anchor": "component-verifier_lab_kernel",
"component_id": "verifier_lab_kernel",
"public_label": "Verifier Lab Kernel"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 9,
"principle": 6
},
"link_count": 22,
"links": [
{
"id": "verifier_lab_kernel",
"kind": "organ",
"kind_label": "component",
"label": "Verifier Lab Kernel",
"relation": "explains",
"url": "component-verifier-lab-kernel.html"
},
{
"id": "mechanism.verifier_lab_kernel.composes_public_formal_math_receipts",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "composes public formal math result records",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-verifier-lab-kernel-composes-public-formal-math-receipts"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.corpus_readiness_mathlib_absence_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Corpus Readiness Mathlib Absence Gate",
"relation": "depends on",
"url": "paper-module-corpus-readiness-mathlib-absence-gate.html"
},
{
"id": "paper_module.lean_std_premise_index",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Lean/Std Premise Index",
"relation": "depends on",
"url": "paper-module-lean-std-premise-index.html"
},
{
"id": "paper_module.formal_math_premise_retrieval",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Premise Retrieval",
"relation": "depends on",
"url": "paper-module-formal-math-premise-retrieval.html"
},
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depends on",
"url": "paper-module-tactic-portfolio-availability.html"
},
{
"id": "paper_module.target_shape_tactic_routing",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Target Shape Tactic Routing",
"relation": "depends on",
"url": "paper-module-target-shape-tactic-routing.html"
},
{
"id": "paper_module.ring2_premise_precision_recall",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Ring-2 Premise Precision Recall",
"relation": "depends on",
"url": "paper-module-ring2-premise-precision-recall.html"
},
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depends on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
},
{
"id": "paper_module.proof_diagnostic_evidence_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof Diagnostic Evidence Spine",
"relation": "depends on",
"url": "paper-module-proof-diagnostic-evidence-spine.html"
},
{
"id": "paper_module.formal_math_lean_proof_witness",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Lean Proof Witness",
"relation": "depends on",
"url": "paper-module-formal-math-lean-proof-witness.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.verifier_lab_execution_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Execution Spine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[44:paper_module.verifier_lab_execution_spine].depends_on",
"url": "paper-module-verifier-lab-execution-spine.html"
}
],
"paper_module_id": "paper_module.verifier_lab_kernel",
"reader_packet": {
"abstract": "The public verifier-lab composition root folds bounded formal-math component result records into one leak-proof result record while separating verifier-backed, oracle-compared, provider-suggested, retrieval-miss, CP2, Evolve, and contract-rejected rows.",
"authority_ceiling": "Public fixture and exported-bundle result records only; no theorem-prover authority, Mathlib import authority, oracle-to-forward success, provider proof authority, benchmark solve-rate claim, launch-scope decision, publishing-scope decision, source-file changes, or secret export.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.verifier_lab_kernel.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-verifier-lab-kernel-diagram-1",
"html_url": "docs/paper-module-verifier-lab-kernel.html#paper-module-verifier-lab-kernel-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD packet[\"Public packet plus 5 negative cases\"] build[\"_build_result load / scan / validate\"] stack[\"_run_component_stack 9 formal-math components\"] buckets[\"_claim_separation 7 claim buckets\"] counters[\"Authority counters oracle / provider / proof-body = 0\"] result records[\"_write_receipts metadata-only result records\"] packet --> build build --> stack stack --> buckets build --> counters buckets --> result records counters --> result records",
"source_format": "mermaid",
"source_hash": "sha256:2ee2bede8fab98f746a6a9a4ad427ba8db27028878945ec52320ddb139b565e2",
"source_line_end": 157,
"source_line_start": 142
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
}
],
"relation": "abides by"
},
{
"count": 9,
"links": [
{
"id": "paper_module.corpus_readiness_mathlib_absence_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Corpus Readiness Mathlib Absence Gate",
"relation": "depends on",
"url": "paper-module-corpus-readiness-mathlib-absence-gate.html"
},
{
"id": "paper_module.lean_std_premise_index",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Lean/Std Premise Index",
"relation": "depends on",
"url": "paper-module-lean-std-premise-index.html"
},
{
"id": "paper_module.formal_math_premise_retrieval",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Premise Retrieval",
"relation": "depends on",
"url": "paper-module-formal-math-premise-retrieval.html"
},
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depends on",
"url": "paper-module-tactic-portfolio-availability.html"
},
{
"id": "paper_module.target_shape_tactic_routing",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Target Shape Tactic Routing",
"relation": "depends on",
"url": "paper-module-target-shape-tactic-routing.html"
},
{
"id": "paper_module.ring2_premise_precision_recall",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Ring-2 Premise Precision Recall",
"relation": "depends on",
"url": "paper-module-ring2-premise-precision-recall.html"
},
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depends on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
},
{
"id": "paper_module.proof_diagnostic_evidence_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof Diagnostic Evidence Spine",
"relation": "depends on",
"url": "paper-module-proof-diagnostic-evidence-spine.html"
},
{
"id": "paper_module.formal_math_lean_proof_witness",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Lean Proof Witness",
"relation": "depends on",
"url": "paper-module-formal-math-lean-proof-witness.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "verifier_lab_kernel",
"kind": "organ",
"kind_label": "component",
"label": "Verifier Lab Kernel",
"relation": "explains",
"url": "component-verifier-lab-kernel.html"
},
{
"id": "mechanism.verifier_lab_kernel.composes_public_formal_math_receipts",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "composes public formal math result records",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-verifier-lab-kernel-composes-public-formal-math-receipts"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.verifier_lab_kernel",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 257,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/verifier_lab_kernel.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "LCF approach",
"section_id": "lcf-composition-and-leandojo-boundaries",
"url": "https://www.research.ed.ac.uk/en/publications/lcf-a-way-of-doing-proofs-with-a-machine/"
},
{
"kind": "external_reference",
"label": "HOL Light",
"section_id": "lcf-composition-and-leandojo-boundaries",
"url": "https://www.cl.cam.ac.uk/~jrh13/papers/hollight.html"
},
{
"kind": "external_reference",
"label": "LeanDojo",
"section_id": "lcf-composition-and-leandojo-boundaries",
"url": "https://arxiv.org/abs/2306.15626"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is the composition root that runs nine formal-math components and folds their outputs into one public record labelling every row by origin.** verifierlabkernel is the composition root for the formal-math verifier lab. It runs or consumes nine smaller formal-math components, then folds their outputs into one public result record that labels every row by where it came from.",
"level": 4,
"order": 1,
"section_id": "it-is-the-composition-root-that-runs-nine-formal-math-components",
"source_line_end": 11,
"source_line_start": 3,
"title": "It is the composition root that runs nine formal-math components"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Not a prover or benchmark runner, it reads a public packet, drives the stack, and sorts evidence into claim buckets, emitting references and verdicts without bodies.** It is not a theorem prover, a benchmark runner, a private Lean import, or a frontend surface. It reads a public problem packet, drives the component stack, sorts each piece of evidence into a fixed set of claim buckets, and writes result records that carry references, hashes, counts, and verdicts. The proof text, provider output, oracle answers, and command stdout stay out of what it emits.",
"level": 4,
"order": 2,
"section_id": "not-a-prover-or-benchmark-runner",
"source_line_end": 22,
"source_line_start": 12,
"title": "Not a prover or benchmark runner"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The point is separation: one record shows which claims a Lean checker backed, which are provider hints, which are oracle comparisons, and which were rejected.** The point is separation. One record shows which claims a Lean checker actually backed, which are provider hints, which are oracle comparisons, and which rows were thrown out for breaking the rules, instead of a pile of separate outputs taken on faith.",
"level": 4,
"order": 3,
"section_id": "the-point-is-separation",
"source_line_end": 31,
"source_line_start": 23,
"title": "The point is separation"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 33,
"source_line_start": 32,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**In a proof loop a checker sign-off, an oracle match, a provider tactic, and a retrieval hit all look like progress, quietly inflating the count of theorems proved.** In a formal-math agent loop several different things look like progress. A Lean checker can accept a term. An oracle holding a hindsight answer can say a candidate matches. A provider model can offer a plausible next tactic. A retrieval step can return a premise. Treated loosely, all of these blur into a single sense of \"it worked\", and oracle or provider success quietly inflates the count of theorems actually proved.",
"level": 4,
"order": 5,
"section_id": "how-progress-blurs-in-a-proof-loop",
"source_line_end": 44,
"source_line_start": 34,
"title": "How progress blurs in a proof loop"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It stops that blur by asking one question per evidence row, which authority class it belongs to, so a passing component cannot lend its standing elsewhere.** This component exists to stop that blur. It answers one question for each row of evidence: which authority class does it belong to, and what may that class claim? A passing component cannot lend its standing to a different bucket.",
"level": 4,
"order": 6,
"section_id": "one-authority-class-question-per-row",
"source_line_end": 52,
"source_line_start": 45,
"title": "One authority-class question per row"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 54,
"source_line_start": 53,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run drives the fixture with five negative cases, runkernel_bundle consumes an exported bundle as a synthetic contract, and main is the CLI over both.** Two entry points run the kernel. run drives the first-wave fixture with the five negative cases included. runkernelbundle drives an exported bundle, where the component result records are consumed as a declared synthetic contract rather than a live run. main is the CLI over both, taking run or run-kernel-bundle with --input and --out.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-and-the-cli",
"source_line_end": 64,
"source_line_start": 55,
"title": "Two entry points and the CLI"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "Both entry points delegate to buildresult, which runs the stages in order: Both entry points delegate to buildresult, which runs the stages in order: loadpayloads reads the public packet, the optional proof-lab route file, and any negative-case inputs from the input directory. scanpaths, seeded by loadforbidden_classes, scans every input and copied source file for forbidden classes. A blocking hit fails the run. runcomponent_stack runs each of the nine components on its own fixture (or standaloneexportedcomponentstack consumes bundle result records). It normalizes each component record to strip legacy redaction fields and public path roots. The stack passes only when every component status\u2026",
"level": 4,
"order": 9,
"section_id": "both-entry-points-delegate-to-build-result-which-runs-the-stages-in",
"source_line_end": 85,
"source_line_start": 65,
"title": "Both entry points delegate to `_build_result`, which runs the stages in"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The stack runs nine formal-math components spanning corpus readiness, premise retrieval, tactic routing, trace repair, diagnostics, and the Lean proof witness. The nine components are corpusreadinessmathlibabsence_gate, leanstdpremiseindex, formalmathpremiseretrieval, tacticportfolioavailabilityprobe, targetshapetacticrouting_gate, ring2premiseretrievalprecisionrecall_harness, formalmathverifiertracerepairloop, proofdiagnosticevidencespine, and formalmathleanproofwitness.",
"level": 4,
"order": 10,
"section_id": "the-nine-components-in-the-stack",
"source_line_end": 96,
"source_line_start": 86,
"title": "The nine components in the stack"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "claimseparation sorts validated rows into exactly seven buckets, with leanverified carrying the proof-witness count and contractrejected holding the observed negative cases. claimseparation produces exactly seven buckets: leanverified, providersuggested, oraclecompared, contract_rejected, retrievalmiss, cp2translated, and evolve_candidate. The lean_verified bucket carries the proof-witness component status and its compiled declaration count. The contract_rejected bucket carries the observed negative cases, marked as contract violations, not results.",
"level": 4,
"order": 11,
"section_id": "claim-separation-sorts-into-seven-buckets",
"source_line_end": 107,
"source_line_start": 97,
"title": "`_claim_separation` sorts into seven buckets"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The boundary is enforced by counters that read zero on a clean run, and the status passes only when the scan, component stack, packet validation, and proof-lab route all pass. The boundary is enforced by counters, not description. build_result computes oracleforwardsuccessincrementcount (oracle rows that marked themselves as forward success), providerresultscounted (provider rows claiming proof authority), and proofbodyexport_count (forward or CP2 rows carrying a proof body). On a clean run these read zero. The final status is a pass only when the secret scan has zero blocking hits, the component stack passes, packet validation passes, and the proof-lab route passes.",
"level": 4,
"order": 12,
"section_id": "counters-not-description-enforce-the-boundary",
"source_line_end": 119,
"source_line_start": 108,
"title": "Counters, not description, enforce the boundary"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "writereceipts emits the public records with bodyinreceipt false, the fixture path marking a real runtime result record and the bundle path marking itself a synthetic contract. write_receipts then emits the public records. The fixture path writes a result record, a board, a validation record, and a fixture sign-off record; the bundle path writes one bundle validation result. Every record sets bodyinreceipt to false and carries receipttransparencycontract, whose omittedpayloadscope is proof, provider, oracle, private-source, and stdout/stderr bodies only. The fixture path marks realruntimereceipt true on a pass; the bundle path marks itself a synthetic contract and not a live run.",
"level": 4,
"order": 13,
"section_id": "write-receipts-emits-body-free-records",
"source_line_end": 158,
"source_line_start": 120,
"title": "`_write_receipts` emits metadata-only records"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 160,
"source_line_start": 159,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The fixture carries five named inputs the kernel must reject, each handled by one typed row validator: The fixture carries five named inputs the kernel must reject, each handled by one typed row validator: forwardproblemleakscandidatebody: a forward problem carrying a candidate, ideal, repair, oracle-premise, source-proof, or base-index field. validateforwardproblems records VERIFIERLABFORWARDFIELD_FORBIDDEN. oraclecountedas_forward: an oracle structured source record with countedasforwardsuccess set true. validateoraclesidecars records VERIFIERLABORACLEFORWARDCONTAMINATION. providerclaimsproof: a provider hypothesis asserting proof authority. validateprovider_hypotheses records VERIFIERLAB\u2026",
"level": 4,
"order": 15,
"section_id": "the-fixture-carries-five-named-inputs-the-kernel-must-reject",
"source_line_end": 185,
"source_line_start": 161,
"title": "The fixture carries five named inputs the kernel must reject"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Each rejected row lands in contractrejected as a contract violation, never in a claim bucket.** Each rejected row lands in contract_rejected as a contract violation, never in a claim bucket.",
"level": 4,
"order": 16,
"section_id": "each-rejected-row-lands-in-contract-rejected-as-a-contract-violation",
"source_line_end": 191,
"source_line_start": 186,
"title": "Each rejected row lands in `contract_rejected` as a contract violation"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 193,
"source_line_start": 192,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It follows LCF-style kernel composition that keeps checked results apart from automation, and LeanDojo's lesson that retrieval and provider hints need explicit boundaries.** This component follows small-kernel theorem-proving and proof-certificate composition. The LCF approach and HOL Light anchor the idea that a verifier lab keeps trusted checked results apart from heuristics and automation. Lean-oriented work such as LeanDojo adds the agent context: retrieval, provider hypotheses, and proof-state interaction need explicit boundaries before they can influence a proof claim. The kernel borrows the composition discipline and keeps verifier, oracle, provider, CP2, and Evolve rows in separat\u2026",
"level": 4,
"order": 18,
"section_id": "lcf-composition-and-leandojo-boundaries",
"source_line_end": 208,
"source_line_start": 194,
"title": "LCF composition and LeanDojo boundaries"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 210,
"source_line_start": 209,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "result record or generated projection surfaces:",
"level": 4,
"order": 20,
"section_id": "reader-validation-without-durable-writes",
"source_line_end": 221,
"source_line_start": 211,
"title": "Reader validation without durable writes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the stack runs, the five negative cases are observed, seven buckets hold, counters stay at zero, and no record leaks a body or local root.** A pass means the component stack runs, the five negative cases are observed, claim_separation holds exactly the seven buckets, the authority counters stay at zero, and no public record carries a body or a leaked local root.",
"level": 4,
"order": 21,
"section_id": "what-a-pass-means-here",
"source_line_end": 229,
"source_line_start": 222,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 231,
"source_line_start": 230,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is bucket separation and metadata-only containment across the composed records, with any return code bound to its own fixture, not generalizable to benchmarks.** The strongest claim this component supports is bucket separation and metadata-only containment across the composed public component records: a reader can see which row a Lean checker backed, which is an oracle comparison, which is a provider hint, which is a CP2 or Evolve candidate, and which was rejected. A Lean/Lake return code or compiled-declaration count is evidence for the public fixture or exported bundle it came from, not a license to generalize to arbitrary benchmarks.",
"level": 4,
"order": 23,
"section_id": "the-strongest-supported-claim",
"source_line_end": 243,
"source_line_start": 232,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Past that line oracle rows stay comparators, provider rows stay advisory, CP2 and Evolve rows stay bounded, and nothing authorizes Mathlib import, benchmark rates, or launch.** That is the proof boundary. The scope limit refuses everything past it. Oracle structured source record stay hindsight comparators and cannot increment forward success. Provider hypotheses stay advisory until a verifier record exists. CP2 rows stay typed action candidates, bounded evidence bodies. Evolve rows stay bounded policy artifacts, not source-file changes. The component does not import private proof bodies, does not create Mathlib import authority, does not claim a benchmark solve rate, and excludes public\u2026",
"level": 4,
"order": 24,
"section_id": "what-the-authority-ceiling-refuses",
"source_line_end": 256,
"source_line_start": 244,
"title": "What the scope limit refuses"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/verifier_lab_kernel.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public verifier-lab composition root folds bounded formal-math component result records into one leak-proof result record while separating verifier-backed, oracle-compared, provider-suggested, retrieval-miss, CP2, Evolve, and contract-rejected rows."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "verifier_lab_kernel.py",
"path": "src/microcosm_core/organs/verifier_lab_kernel.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for the verifier-lab composition mechanism.",
"symbols": [
"run",
"run_kernel_bundle",
"_build_result",
"_write_receipts"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/verifier_lab_kernel.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/verifier_lab_kernel.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/verifier_lab_kernel.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Verifier Lab Kernel is the public result record compiler for the formal-math fixture cohort. It runs or consumes named component result records, classifies proof-lab rows by authority class, preserves scope boundaries, and excludes proof/provider/oracle bodies from exported result records.",
"summary_line": "The public verifier-lab composition root folds bounded formal-math component result records into one leak-proof result record while separating verifier-backed, oracle-compared, provider-suggested, retrieval-miss, CP2, Evolve, and contract-rejected rows.",
"title": "Verifier Lab Kernel"
},
{
"anchor": "paper-module-navigation-hologram-route-plane",
"authority_ceiling": "Public fixture and exported-bundle result records only; no live route freshness, source authority, provider/live-kernel execution, private operator state, later-component authorization, launch-scope decision, or whole-wave certification.",
"body_html": "
Readers trust whatever browse surface is nearest, so this validator checks that a navigation surface starts from one control entry and keeps every row a projection. Anyone who lands in a large codebase starts from whatever browse surface is nearest, treats that surface as the truth, and acts on a stale or partial view. The route plane is a validator that pushes back on this. It checks that a navigation surface starts from one control entry, that browse rows stay projections, and that no row is allowed to pose as the thing it describes.
\n
It runs offline over fixtures and one exact copy of navigation source modules, emitting plain metadata-only result records that each state only what they checked. The component runs offline over fixture files and over one exact copy of navigation source modules. It never touches a live system. What it emits is a set of plain result records: a cluster summary, a selected card, a source-coupling result, a route lease, an entry-admission record, an affordance-selection record, and a code-architecture packet. Each record states what it checked and carries no source body text.
\n
It is a rules check on a navigation surface, not a live router, a search index, or a source of navigation truth. It is not a live router, a search index, or a source of navigation truth. It is a check that a navigation surface obeys its own rules before a reader or agent trusts it.
\n
Purpose
\n
Why a browse row needs a coupling record
\n
A browse row that speaks with its source's authority lets a stale view steer the next move, so the plane demands a coupling record before granting any authority. A loose navigation surface fails in a specific way. A browse row is a compressed view of some source, and if the row is allowed to speak with the source's authority, then a stale or mislabeled row silently steers the next move. The route plane treats every browse row as a projection and demands a coupling record before that projection gets any authority.
\n
Source coupling is a SHA-256 fingerprint
\n
Source coupling hashes the route rows with SHA-256 and denies authority when the fingerprint or row count disagrees with disk, recording a stale-but-current summary as an overclaim. Source coupling is a plain SHA-256 over the route rows. The manifest declares an expected fingerprint and an expected row count. If either disagrees with the rows on disk, the projection is denied current authority. A summary that claims to be current while its coupling is stale is recorded as an overclaim.
\n
Refusals the fixture must keep catching
\n
The plane forces first contact to the control entry, forbids dropping a required control field on compaction, and demotes anti-trigger affordances, each a named negative case. The other half is what the plane refuses. First contact must begin at the control entry, so a request that starts from a browse row is replaced with the entry route. Compaction of the entry packet may not drop a required control field. An affordance row whose passport carries an anti-trigger is demoted before similarity search can select it. Each of these is a named negative case the fixture must keep catching.
\n
How it works
\n
Two entry points, one discipline
\n
Two entry points share the same discipline: run drives the fixture mode and run_route_plane_bundle drives the exported-bundle mode. The component has two entry points against the same discipline. run drives the fixture mode. run_route_plane_bundle drives the exported-bundle mode.
\n
The run fixture-mode pipeline
\n
run loads thirteen fixture inputs, scans for leakage, builds a toy option surface, and passes only when the scan is clean and every expected negative case fired.run loads thirteen fixture inputs, scans them for non-public-state leakage, builds a toy option surface with build_toy_option_surface (a cluster summary plus one selected card), then evaluates nine negative-case validators. It merges the negative cases each validator observed and computes the set of expected cases that went unobserved. The status is pass only when no expected case is missing and the non-public-state scan is clean. The fixture proves each guard still fires by requiring every expected negative case to actually appear.
\n
The run_route_plane_bundle bundle-mode gate
\n
run_route_plane_bundle loads nine bundle files, digests and secret-scans them, and passes only when a card was selected and every bundle validator finds nothing.run_route_plane_bundle loads nine bundle files, records a freshness basis and digest over them, scans them with the secret-exclusion scan, then runs the group of exported-bundle validators. Its status is pass only when the secret scan is clean, a card row was selected, no validator produced a finding, and the source-module, lease, entry, affordance, and code-architecture checks all pass.
\n
Function
Role
run
Fixture entry: builds the toy surface, runs the nine negative-case guards, requires every expected case to fire.
run_route_plane_bundle
Bundle entry: validates a copied route-plane bundle end to end and gates the pass on every component check.
validate_exported_route_rows
Checks each route row for an allowed surface role, an actionable command, no source-authority claim, and an omission record when the body is compressed; catches duplicate route ids and a missing card row.
validate_exported_source_coupling
The coupling gate: hashes the route rows and compares the fingerprint and row count against the manifest; denies authority on mismatch or stale status.
validate_source_module_manifest
Validates the five copied source modules by SHA-256 digest and by required navigation anchors, and enforces the copied-body boundary.
write_receipts
Writes the result records atomically under public-root-relative paths, with a common envelope and no source body text.
\n
Digest and anchor checks on copied modules
\n
_stable_hash fingerprints the route rows, while validate_source_module_manifest matches each copied file's digest and anchors, rejects unsafe paths, and keeps bodies out of records. The coupling gate uses _stable_hash, a SHA-256 over the route rows serialized as canonical JSON. validate_source_module_manifest reads each copied file, requires its digest to match both the declared source and target digests, requires every declared navigation anchor to be present in the body, and requires the boundary flags body_copied=true and body_in_receipt=false. A target reference that is absolute or contains .. is rejected before any read. The body text is never written into a result record.
\nDiagram of the mechanism (9 steps).Diagram source & refs
flowchart TD Entry["Control entry first a browse row that claims first contact is replaced with the entry route"] Couple["Source coupling SHA-256 over route rows vs manifest fingerprint and row count"] Rows["Route rows allowed role, actionable command, no source-authority claim, omission record when compressed"] Modules["Source-module manifest five copied bodies digest and required anchors, body never in the record"] Floor["Entry-packet floor required control fields survive compaction"] Pass["Affordance passports anti-trigger rows demoted before selection"] Verdict["Coupling current, all checks pass, card row selected?"] Result records["metadata-only result records cluster flag, card, coupling, route lease, entry admission, affordance, code-architecture packet"] Blocked["Blocked stable error codes and findings, bodies redacted"] Entry --> Couple Couple --> Rows Rows --> Modules Modules --> Floor Floor --> Pass Pass --> Verdict Verdict --> Result records Verdict --> Blocked
\n
Negative cases
\n
Eight cases carrying nine error codes
\n
Eight negative cases carry nine stable error codes, one case carrying two, and the fixture passes only when every code is observed. The fixture mode is defined by eight negative cases that carry nine stable error codes. One case carries two codes, which is why the counts differ. Every code below must be observed for the fixture to pass.
\n
Case
Error code
First contact starts from a drilldown projection
BANNED_FIRST_CONTACT_ROUTE
Source-coupling fingerprint is stale
SOURCE_COUPLING_STALE
Compressed route card carries no omission record
MISSING_OMISSION_RECEIPT
ATLAS_PROJECTION_NOT_CONTROL_ENTRY
Route card leaks a private body
ROUTE_CARD_PRIVATE_BODY_LEAK
Route summary overclaims freshness
ROUTE_SUMMARY_OVERCLAIMS_FRESHNESS
Two rows share a route id
DUPLICATE_ROUTE_ID_CONFLICT
Entry admission drops a required control field
ENTRY_ADMISSION_CONTROL_FLOOR_DROPPED
An anti-trigger affordance is not demoted
AFFORDANCE_PASSPORT_ANTITRIGGER_IGNORED
\n
Prior Art Grounding
\n
Progressive disclosure, CLI, and graph tooling
\n
The design borrows progressive disclosure, CLI discoverability, and directed-graph tooling, but keeps cards, leases, and browse rows below authority until coupling and entry records agree. The route plane sits on information-architecture and graph-navigation practice. The first-contact rule follows the usability pressure behind progressive disclosure: show the control entry and immediate affordances before deeper browse rows. The command-facing surface follows the Command Line Interface Guidelines emphasis on discoverable commands and clear next actions. The graph side is ordinary directed-graph tooling: NetworkX documents topological sorting over dependency edges, and PageRank shows the older pattern of computing route salience from graph structure. This component keeps those ideas below authority: cards, leases, and browse rows stay projections unless the coupling and entry-admission records agree.
\n
Source-Backed Doctrine Packet
\n
The authoritative sources for this module are:
\n
core/organ_registry.json::implemented_organs[navigation_hologram_route_plane] records the accepted component authority: evidence class semantic_validator, scope limit validates declared public contract only, and validator command python -m microcosm_core.organs.navigation_hologram_route_plane run --input fixtures/first_wave/navigation_hologram_route_plane/input --out receipts/first_wave/navigation_hologram_route_plane.
core/organ_atlas.json::organs[navigation_hologram_route_plane] carries the cold-reader gloss. It restates the scope limit: It validates only the declared public toy route-plane contract and its regression fixtures (plus exact copied navigation source modules in the bundle path); it does not establish live route freshness, grant source authority, authorize any later component, run any provider/live-kernel call, or certify the whole wave.
standards/std_microcosm_navigation_hologram_route_plane.json governs the authority boundary public_navigation_route_plane_runtime_and_copied_source_body_validator_not_live_source_authority and declares the runtime bundle command python -m microcosm_core.organs.navigation_hologram_route_plane validate-route-plane-bundle --input examples/navigation_hologram_route_plane/exported_route_plane_bundle.
src/microcosm_core/organs/navigation_hologram_route_plane.py is the runtime source for fixture validation, bundle validation, the coupling gate, the negative-case guards, and the source-module digest and anchor checks.
core/fixture_manifests/navigation_hologram_route_plane.fixture_manifest.json binds fixture expectations: body_copied_material_count=5, body material status copied_non_secret_macro_route_substrate_with_provenance, body_in_receipt=false, and the nine stable error codes above.
examples/navigation_hologram_route_plane/exported_route_plane_bundle/source_module_manifest.json names the five exact copied route and control bodies validated by digest and anchors.
tests/test_navigation_hologram_route_plane.py is the regression floor for fixture records, copied-source digests, source-module anchors, record redaction, bundle validation, digest-mismatch rejection, and this packet.
A pass means the fixture caught every negative case, the bundle validated its rows, coupling, and copied modules, and no record leaked a body or non-public paths. A pass means the fixture observed every expected negative case, the exported bundle validated its route rows, source coupling, and copied source modules, and no result record carried a source body or a non-public paths. The validation result records for this module are the result records those commands write.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The strongest claim is that public fixture and bundle evidence agree on the declared route-plane contract, and that is the proof boundary and scope limit. The strongest claim this module supports is public fixture and exported-bundle evidence: the route rows, copied source modules, source manifests, negative cases, and validation result records agree on the declared public route-plane contract. That is its proof boundary and its scope limit.
\n
What it refuses past that line
\n
Past that line it proves no live route freshness, grants no kernel authority, and excludes source-file changes, external model access, state export, public sharing, or whole-system equivalence. It refuses everything past that line. This module does not establish live route freshness, grant live source-kernel authority, authorize source-file changes, authorize external model access, export account or browser state, expose browser UI live access, authorize public sharing or launch, prove whole-system correctness, or assert whole-system equivalence.
\n",
"components": [
{
"anchor": "component-navigation_hologram_route_plane",
"component_id": "navigation_hologram_route_plane",
"public_label": "Navigation Hologram Route Plane"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 4,
"principle": 9
},
"link_count": 22,
"links": [
{
"id": "navigation_hologram_route_plane",
"kind": "organ",
"kind_label": "component",
"label": "Navigation Hologram Route Plane",
"relation": "explains",
"url": "component-navigation-hologram-route-plane.html"
},
{
"id": "mechanism.navigation_hologram_route_plane.validates_public_route_plane_bundle",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public route plane bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-navigation-hologram-route-plane-validates-public-route-plane-bundle"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.routing_anti_patterns_registry",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Routing Anti-Patterns Registry",
"relation": "depends on",
"url": "paper-module-routing-anti-patterns-registry.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[2:paper_module.agent_route_observability_runtime].depends_on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.batch11_saturation_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 11 Saturation Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[73:paper_module.batch11_saturation_engines_capsule].depends_on",
"url": "paper-module-batch11-saturation-engines-capsule.html"
},
{
"id": "paper_module.batch8_compliance_pipeline_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Compliance Pipeline Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[59:paper_module.batch8_compliance_pipeline_capsule].depends_on",
"url": "paper-module-batch8-compliance-pipeline-capsule.html"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[13:paper_module.cold_reader_route_map].depends_on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.doctrine_fact_claim_audit",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Doctrine Fact Claim Audit",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[48:paper_module.doctrine_fact_claim_audit].depends_on",
"url": "paper-module-doctrine-fact-claim-audit.html"
},
{
"id": "paper_module.engine_room_navigation_fitness_benchmark",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Navigation Fitness Benchmark",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[88:paper_module.engine_room_navigation_fitness_benchmark].depends_on",
"url": "paper-module-engine-room-navigation-fitness-benchmark.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[19:paper_module.macro_projection_import_protocol].depends_on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[9:paper_module.pattern_binding_contract].depends_on",
"url": "paper-module-pattern-binding-contract.html"
},
{
"id": "paper_module.routing_anti_patterns_registry",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Routing Anti-Patterns Registry",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[57:paper_module.routing_anti_patterns_registry].depends_on",
"url": "paper-module-routing-anti-patterns-registry.html"
},
{
"id": "paper_module.self_ignorance_coverage_ledger",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Self-Ignorance Coverage Ledger",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[49:paper_module.self_ignorance_coverage_ledger].depends_on",
"url": "paper-module-self-ignorance-coverage-ledger.html"
},
{
"id": "paper_module.voice_to_doctrine_self_improvement_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Voice to Doctrine Self-Improvement Loop",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[56:paper_module.voice_to_doctrine_self_improvement_loop].depends_on",
"url": "paper-module-voice-to-doctrine-self-improvement-loop.html"
}
],
"paper_module_id": "paper_module.navigation_hologram_route_plane",
"reader_packet": {
"abstract": "The public navigation route-plane fixture validates bounded route projections, source-coupling gates, entry floors, affordance passports, and copied navigation source-module digests without treating browse rows as authority.",
"authority_ceiling": "Public fixture and exported-bundle result records only; no live route freshness, source authority, provider/live-kernel execution, private operator state, later-component authorization, launch-scope decision, or whole-wave certification.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.navigation_hologram_route_plane.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-navigation-hologram-route-plane-diagram-1",
"html_url": "docs/paper-module-navigation-hologram-route-plane.html#paper-module-navigation-hologram-route-plane-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Entry[\"Control entry first a browse row that claims first contact is replaced with the entry route\"] Couple[\"Source coupling SHA-256 over route rows vs manifest fingerprint and row count\"] Rows[\"Route rows allowed role, actionable command, no source-authority claim, omission record when compressed\"] Modules[\"Source-module manifest five copied bodies digest and required anchors, body never in the record\"] Floor[\"Entry-packet floor required control fields survive compaction\"] Pass[\"Affordance passports anti-trigger rows demoted before selection\"] Verdict[\"Coupling current, all checks pass, card row selected?\"] Result records[\"metadata-only result records cluster flag, card, coupling, route lease, entry admission, affordance, code-architecture packet\"] Blocked[\"Blocked stable error codes and findings, bodies redacted\"] Entry --> Couple Couple --> Rows Rows --> Modules Modules --> Floor Floor --> Pass Pass --> Verdict Verdict --> Result records Verdict --> Blocked",
"source_format": "mermaid",
"source_hash": "sha256:4b85a0fb7692a0cbb328ee12100327a577d1657d71b250d8365bb2dc0c7f674f",
"source_line_end": 75,
"source_line_start": 56
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 4,
"links": [
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.routing_anti_patterns_registry",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Routing Anti-Patterns Registry",
"relation": "depends on",
"url": "paper-module-routing-anti-patterns-registry.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "navigation_hologram_route_plane",
"kind": "organ",
"kind_label": "component",
"label": "Navigation Hologram Route Plane",
"relation": "explains",
"url": "component-navigation-hologram-route-plane.html"
},
{
"id": "mechanism.navigation_hologram_route_plane.validates_public_route_plane_bundle",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public route plane bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-navigation-hologram-route-plane-validates-public-route-plane-bundle"
}
],
"relation": "explains"
},
{
"count": 10,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.navigation_hologram_route_plane",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "re-entry-conditions",
"source_line_start": 158,
"title": "Re-entry Conditions"
},
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 175,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/navigation_hologram_route_plane.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "progressive disclosure",
"section_id": "progressive-disclosure-cli-and-graph-tooling",
"url": "https://www.nngroup.com/articles/progressive-disclosure/"
},
{
"kind": "external_reference",
"label": "Command Line Interface Guidelines",
"section_id": "progressive-disclosure-cli-and-graph-tooling",
"url": "https://clig.dev/"
},
{
"kind": "external_reference",
"label": "topological sorting",
"section_id": "progressive-disclosure-cli-and-graph-tooling",
"url": "https://networkx.org/documentation/stable/reference/algorithms/generated/networkx.algorithms.dag.topological_sort.html"
},
{
"kind": "external_reference",
"label": "PageRank",
"section_id": "progressive-disclosure-cli-and-graph-tooling",
"url": "https://networkx.org/documentation/stable/reference/algorithms/generated/networkx.algorithms.link_analysis.pagerank_alg.pagerank.html"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "Readers trust whatever browse surface is nearest, so this validator checks that a navigation surface starts from one control entry and keeps every row a projection. Anyone who lands in a large codebase starts from whatever browse surface is nearest, treats that surface as the truth, and acts on a stale or partial view. The route plane is a validator that pushes back on this. It checks that a navigation surface starts from one control entry, that browse rows stay projections, and that no row is allowed to pose as the thing it describes.",
"level": 4,
"order": 1,
"section_id": "readers-trust-whatever-browse-surface-is-nearest",
"source_line_end": 6,
"source_line_start": 3,
"title": "Readers trust whatever browse surface is nearest"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It runs offline over fixtures and one exact copy of navigation source modules, emitting plain metadata-only result records that each state only what they checked. The component runs offline over fixture files and over one exact copy of navigation source modules. It never touches a live system. What it emits is a set of plain result records: a cluster summary, a selected card, a source-coupling result, a route lease, an entry-admission record, an affordance-selection record, and a code-architecture packet. Each record states what it checked and carries no source body text.",
"level": 4,
"order": 2,
"section_id": "it-runs-offline-over-fixtures",
"source_line_end": 10,
"source_line_start": 7,
"title": "It runs offline over fixtures"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It is a rules check on a navigation surface, not a live router, a search index, or a source of navigation truth. It is not a live router, a search index, or a source of navigation truth. It is a check that a navigation surface obeys its own rules before a reader or agent trusts it.",
"level": 4,
"order": 3,
"section_id": "it-is-a-rules-check-on-a-navigation-surface",
"source_line_end": 14,
"source_line_start": 11,
"title": "It is a rules check on a navigation surface"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A browse row that speaks with its source's authority lets a stale view steer the next move, so the plane demands a coupling record before granting any authority. A loose navigation surface fails in a specific way. A browse row is a compressed view of some source, and if the row is allowed to speak with the source's authority, then a stale or mislabeled row silently steers the next move. The route plane treats every browse row as a projection and demands a coupling record before that projection gets any authority.",
"level": 4,
"order": 5,
"section_id": "why-a-browse-row-needs-a-coupling-record",
"source_line_end": 20,
"source_line_start": 17,
"title": "Why a browse row needs a coupling record"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Source coupling hashes the route rows with SHA-256 and denies authority when the fingerprint or row count disagrees with disk, recording a stale-but-current summary as an overclaim. Source coupling is a plain SHA-256 over the route rows. The manifest declares an expected fingerprint and an expected row count. If either disagrees with the rows on disk, the projection is denied current authority. A summary that claims to be current while its coupling is stale is recorded as an overclaim.",
"level": 4,
"order": 6,
"section_id": "source-coupling-is-a-sha-256-fingerprint",
"source_line_end": 24,
"source_line_start": 21,
"title": "Source coupling is a SHA-256 fingerprint"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The plane forces first contact to the control entry, forbids dropping a required control field on compaction, and demotes anti-trigger affordances, each a named negative case. The other half is what the plane refuses. First contact must begin at the control entry, so a request that starts from a browse row is replaced with the entry route. Compaction of the entry packet may not drop a required control field. An affordance row whose passport carries an anti-trigger is demoted before similarity search can select it. Each of these is a named negative case the fixture must keep catching.",
"level": 4,
"order": 7,
"section_id": "refusals-the-fixture-must-keep-catching",
"source_line_end": 28,
"source_line_start": 25,
"title": "Refusals the fixture must keep catching"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 30,
"source_line_start": 29,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Two entry points share the same discipline: run drives the fixture mode and runrouteplanebundle drives the exported-bundle mode. The component has two entry points against the same discipline. run drives the fixture mode. runrouteplanebundle drives the exported-bundle mode.",
"level": 4,
"order": 9,
"section_id": "two-entry-points-one-discipline",
"source_line_end": 34,
"source_line_start": 31,
"title": "Two entry points, one discipline"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run loads thirteen fixture inputs, scans for leakage, builds a toy option surface, and passes only when the scan is clean and every expected negative case fired. run loads thirteen fixture inputs, scans them for non-public-state leakage, builds a toy option surface with buildtoyoption_surface (a cluster summary plus one selected card), then evaluates nine negative-case validators. It merges the negative cases each validator observed and computes the set of expected cases that went unobserved. The status is pass only when no expected case is missing and the non-public-state scan is clean. The fixture proves each guard still fires by requiring every expected negative case to actually appear.",
"level": 4,
"order": 10,
"section_id": "the-run-fixture-mode-pipeline",
"source_line_end": 38,
"source_line_start": 35,
"title": "The `run` fixture-mode pipeline"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "runrouteplanebundle loads nine bundle files, digests and secret-scans them, and passes only when a card was selected and every bundle validator finds nothing. runrouteplanebundle loads nine bundle files, records a freshness basis and digest over them, scans them with the secret-exclusion scan, then runs the group of exported-bundle validators. Its status is pass only when the secret scan is clean, a card row was selected, no validator produced a finding, and the source-module, lease, entry, affordance, and code-architecture checks all pass.",
"level": 4,
"order": 11,
"section_id": "the-run-route-plane-bundle-bundle-mode-gate",
"source_line_end": 51,
"source_line_start": 39,
"title": "The `run_route_plane_bundle` bundle-mode gate"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "stablehash fingerprints the route rows, while validatesourcemodulemanifest matches each copied file's digest and anchors, rejects unsafe paths, and keeps bodies out of records. The coupling gate uses stablehash, a SHA-256 over the route rows serialized as canonical JSON. validatesourcemodulemanifest reads each copied file, requires its digest to match both the declared source and target digests, requires every declared navigation anchor to be present in the body, and requires the boundary flags bodycopied=true and bodyinreceipt=false. A target reference that is absolute or contains .. is rejected before any read. The body text is never written into a result record.",
"level": 4,
"order": 12,
"section_id": "digest-and-anchor-checks-on-copied-modules",
"source_line_end": 76,
"source_line_start": 52,
"title": "Digest and anchor checks on copied modules"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 78,
"source_line_start": 77,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "Eight negative cases carry nine stable error codes, one case carrying two, and the fixture passes only when every code is observed. The fixture mode is defined by eight negative cases that carry nine stable error codes. One case carries two codes, which is why the counts differ. Every code below must be observed for the fixture to pass.",
"level": 4,
"order": 14,
"section_id": "eight-cases-carrying-nine-error-codes",
"source_line_end": 94,
"source_line_start": 79,
"title": "Eight cases carrying nine error codes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 96,
"source_line_start": 95,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The design borrows progressive disclosure, CLI discoverability, and directed-graph tooling, but keeps cards, leases, and browse rows below authority until coupling and entry records agree. The route plane sits on information-architecture and graph-navigation practice. The first-contact rule follows the usability pressure behind progressive disclosure: show the control entry and immediate affordances before deeper browse rows. The command-facing surface follows the Command Line Interface Guidelines emphasis on discoverable commands and clear next actions. The graph side is ordinary directed-graph tooling: NetworkX documents topological sorting over dependency edges, and PageRank shows the ol\u2026",
"level": 4,
"order": 16,
"section_id": "progressive-disclosure-cli-and-graph-tooling",
"source_line_end": 100,
"source_line_start": 97,
"title": "Progressive disclosure, CLI, and graph tooling"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate/, reproduce this page's proof boundary with temporary result records:",
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 111,
"source_line_start": 101,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means the fixture caught every negative case, the bundle validated its rows, coupling, and copied modules, and no record leaked a body or non-public paths. A pass means the fixture observed every expected negative case, the exported bundle validated its route rows, source coupling, and copied source modules, and no result record carried a source body or a non-public paths. The validation result records for this module are the result records those commands write.",
"level": 4,
"order": 18,
"section_id": "what-a-pass-certifies-here",
"source_line_end": 115,
"source_line_start": 112,
"title": "What a pass certifies here"
},
{
"block_kinds": [
"code",
"list",
"paragraph",
"table"
],
"excerpt": "The authoritative sources for this module are: core/organregistry.json::implementedorgans[navigationhologramrouteplane] records the accepted component authority: evidence class semanticvalidator, scope limit validates declared public contract only, and validator command python -m microcosmcore.components.navigationhologramrouteplane run --input fixtures/firstwave/navigationhologramrouteplane/input --out result records/firstwave/navigationhologramrouteplane. core/organatlas.json::components[navigationhologramrouteplane] carries the cold-reader gloss. It restates the scope limit: It validates only the declared public toy route-plane contract and its regression fixtures (plus exact copied navi\u2026",
"level": 2,
"order": 19,
"section_id": "source-backed-doctrine-packet",
"source_line_end": 157,
"source_line_start": 116,
"title": "Source-Backed Doctrine Packet"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 166,
"source_line_start": 165,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim is that public fixture and bundle evidence agree on the declared route-plane contract, and that is the proof boundary and scope limit. The strongest claim this module supports is public fixture and exported-bundle evidence: the route rows, copied source modules, source manifests, negative cases, and validation result records agree on the declared public route-plane contract. That is its proof boundary and its scope limit.",
"level": 4,
"order": 21,
"section_id": "the-strongest-supported-claim",
"source_line_end": 170,
"source_line_start": 167,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Past that line it proves no live route freshness, grants no kernel authority, and excludes source-file changes, external model access, state export, public sharing, or whole-system equivalence. It refuses everything past that line. This module does not establish live route freshness, grant live source-kernel authority, authorize source-file changes, authorize external model access, export account or browser state, expose browser UI live access, authorize public sharing or launch, prove whole-system correctness, or assert whole-system equivalence.",
"level": 4,
"order": 22,
"section_id": "what-it-refuses-past-that-line",
"source_line_end": 174,
"source_line_start": 171,
"title": "What it refuses past that line"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/navigation_hologram_route_plane.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public navigation route-plane fixture validates bounded route projections, source-coupling gates, entry floors, affordance passports, and copied navigation source-module digests without treating browse rows as authority."
}
},
"source_drilldown": {
"source_manifests": [
{
"label": "Source manifest",
"path": "examples/navigation_hologram_route_plane/exported_route_plane_bundle/source_module_manifest.json",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/examples/navigation_hologram_route_plane/exported_route_plane_bundle/source_module_manifest.json"
}
],
"source_modules": [
{
"label": "navigation_hologram_route_plane.py",
"path": "src/microcosm_core/organs/navigation_hologram_route_plane.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public route-plane fixture and exported-bundle validation.",
"symbols": [
"run",
"run_route_plane_bundle",
"validate_exported_route_rows",
"validate_source_module_manifest",
"write_receipts"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/navigation_hologram_route_plane.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/navigation_hologram_route_plane.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/navigation_hologram_route_plane.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Navigation Hologram Route Plane is the public route-surface validator for moving from a Microcosm-local control entry into browsable route projections. It checks route rows, source-coupling fingerprints, copied source-module anchors and hashes, route-lease policy, entry-packet floors, affordance-passport selection, and code-architecture projection packets while keeping full source bodies out of result records.",
"summary_line": "The public navigation route-plane fixture validates bounded route projections, source-coupling gates, entry floors, affordance passports, and copied navigation source-module digests without treating browse rows as authority.",
"title": "Navigation Hologram Route Plane"
},
{
"anchor": "paper-module-agent-route-observability-runtime",
"authority_ceiling": "Public synthetic route-observability fixtures, copied source body digests, and exported-bundle result records only; no live session introspection, provider/browser UI/account authority, live hook control, benchmark-performance proof, source-file changes, launch-scope decision, or whole-system correctness.",
"body_html": "
One recorded trail is checked end to end
\n
It replays one agent's recorded trail \u2014 route, work, events, evidence, and declared authority limit \u2014 and checks the whole thing is well-formed and self-consistent. This component takes a recorded trail of one agent's local run, the route it picked, the work it logged, the events it emitted, the evidence it pointed to, and the authority limit it declared, and checks that the trail is well-formed and self-consistent. It is a replay validator over synthetic fixtures, not a live tap on a running agent.
\n
Six checks share the declared scope limit
\n
Because the record states its authority limit up front, the runtime runs six checks and writes six metadata-only result records that each carry status, findings, and the scope limit. The record is built to state up front where the agent's authority was supposed to stop, so the limit is written down and checkable rather than taken on faith. The runtime reads that record, runs six checks over it, and writes six small JSON result records. Each result record carries a status, the negative cases it saw, the findings, and the scope limit. Large or private payloads are dropped before anything is written.
\n
Recorded evidence cannot prove live compliance
\n
It reads recorded evidence only, inspecting no live session and proving nothing about whether an agent actually stayed in bounds. It reads recorded evidence. It does not inspect live sessions, mutate a route, install a hook, authorize a provider, or prove that any agent actually stayed in bounds.
\n
Purpose
\n
A raw log reads as proof
\n
A raw agent log is hard to trust because a behavior-change claim or a quiet authority grab reads as proof while nothing checkable is recorded. A raw agent log is hard to trust. Someone reading it by hand cannot easily tell whether a row that claims "I changed how the agent behaves" is backed by anything, or whether an advisory step quietly asserted authority it never had. Loose logging fails here in a specific way: a trace that reads as proof, with nothing recorded that a reader can check.
\n
No claim counts unless the record backs it
\n
It refuses to treat a claim as evidence unless the record supports it: advisory rows cannot claim mutation authority, behavior-change needs named trace ids, and private bodies are redacted. This component closes that gap by refusing to treat a claim as evidence unless the record supports it. An advisory row cannot also claim live mutation authority. A behavior-change claim is blocked unless it names the trace ids that evidence it. A private transcript body is rejected and redacted. The useful output is a compact result record: one command, one selected route, one work-event-evidence chain, one validator handle, and one scope limit.
\n
How it works
\n
run scans, then calls six validators
\n
run loads the fixture, scans the inputs for forbidden non-public-state classes before any check, then calls six validators in order and merges what they observed. The default entrypoint is run. It loads the first-wave fixture, resolves the public root for the input path, and scans the inputs for forbidden non-public-state classes before any check runs. It then calls six validators in order and merges what they observed.
\n
A green run means every rejection fired
\n
The fixture is deliberately adversarial, so a green run means every expected negative case fired and nothing unexpected did, not that the trace was clean. The fixture holds 10 trace rows, 6 hook-shadow cases, 6 egress cases, and 2 anti-pattern debt rows. It is deliberately adversarial: it is built to trigger every planned rejection. A green run therefore means every expected negative case fired and nothing unexpected did, not that the trace was clean.
\n
validate_route_compliance classifies each row
\n
validate_route_compliance checks each trace row for duplicate ids, a missing lease, a route miss, a private payload, an advisory authority claim, or unevidenced behavior change, passing only on expected cases.validate_route_compliance classifies the trace rows. It counts duplicate event ids, and for each row it checks for a missing route lease, a route miss where the observed route differs from the expected route, a private transcript payload class, an advisory row that claims mutation authority, and a behavior-change claim with no evidence ids. Each check that fires records a typed finding and marks the row rejected. The check passes when every negative case it saw is one of the expected cases.
\n
validate_route_lease_mode_control reads leases
\n
validate_route_lease_mode_control flags kernel bloat before direct action and leases that offer static route metadata in place of consumed trace feedback.validate_route_lease_mode_control reads the same trace rows as route leases. It flags KERNEL_BLOAT_BEFORE_DIRECT_ACTION when a lease returns to broad kernel context before acting, and ROUTE_LEASE_NOT_CONSUMED when static route metadata is offered in place of consumed trace feedback.
\n
validate_hook_shadow keeps hooks projection-only
\n
validate_hook_shadow replays each hook case, requiring projection_not_authority, metadata-only fields, and every live-state field false, so any read of live operator, provider, or operator UI state is rejected.validate_hook_shadow replays the hook intervention cases. Each case must map a case id, hook id, coverage claim, repair class, and expected intervention, must declare projection_not_authority, must stay metadata-only, and must keep every live-state field false. A case that reads live operator, provider, browser, HUD, cockpit, or ledger state is rejected.
validate_debt_retirement retires a debt row only when it names behavior-change evidence trace ids, so evidence-only telemetry cannot retire debt on its own.validate_debt_retirement marks a debt row retired only when it names behavior-change evidence trace ids and carries no evidence-only trace ids. Evidence-only telemetry cannot retire debt on its own.
\n
validate_agent_principle_lens mints nothing
\n
validate_agent_principle_lens confirms the lens preserves its selected principle ids and route handles while minting no principle, promoting no axiom, and exporting no full body.validate_agent_principle_lens confirms the lens preserves its selected principle ids and its route handles back to the principle cards and the agent operating packet. The lens cannot mint principles, promote a candidate axiom, or export full principle bodies. Any of those sets a finding.
\n
validate_egress_mirror recomputes each decision
\n
validate_egress_mirror recomputes each detector decision against the declared one, requiring every private-read, provider-read, cockpit-read, and body-export flag to stay false.validate_egress_mirror recomputes each detector decision and compares it to the declared one. Every case must keep private_state_read, provider_payload_read, browser_hud_cockpit_state_read, and body_exported false.
\n
A pass writes six metadata-only result records
\n
run passes only when no expected case is missing, the scan is clean, and all six validators pass, after which write_receipts emits six metadata-only JSON records and result_card a compact card.run passes only when no expected negative case is missing, the non-public-state scan passes, and route compliance, hook shadow, route trace analytics, the principle lens, and the egress mirror all pass. It then calls write_receipts, which writes the six JSON result records under the public root and returns their public-relative paths. result_card compresses a full result into a start card and drops the full findings, private scans, source bodies, and scope-limit bodies.
\n
Bundle runners over exported evidence slices
\n
run_observability_bundle validates a larger exported bundle and blocks on any forbidden payload key or manifest drift, while companion runners check the same membrane over adjacent evidence slices. Alongside the fixture path, run_observability_bundle validates a larger exported bundle: route events, agent path observations, session diagnostics, hook-shadow rows, actor-axis checks, debt rows, process-audit rows, an observability policy, and a source-module manifest, blocking on any forbidden payload key or manifest drift. Companion runners such as run_agent_observability_store_bundle check the same membrane over adjacent public route, session, bridge, controller, and store evidence slices.
\n
Function
Role
run
Loads the fixture, runs the six checks in order, composes status, writes six result records
Reads lease consumption; flags kernel bloat before direct action and unconsumed leases
validate_hook_shadow
Replays hook cases; requires projection-only, metadata-only, and all live-state fields false
validate_debt_retirement
Retires a debt row only when it names behavior-change evidence and no evidence-only refs
validate_agent_principle_lens
Checks the lens preserves selected ids and route handles and mints nothing
validate_egress_mirror
Recomputes detector decisions; requires no private read, provider read, HUD read, or body export
write_receipts
Writes the six result records under the public root and returns their paths
result_card
Compresses a result into a card, omitting findings, private scans, and source bodies
\nDiagram of the mechanism (7 steps).Diagram source & refs
flowchart TD Fixture["Fixture inputs agent_trace, hook_shadow, debt, principle_lens, egress"] Scan["non-public-state scan forbidden classes rejected"] Checks["Six validators route, lease, hook, debt, lens, egress"] Merge["Merge observed negative cases"] Status["Status pass only if every expected case fired"] Result records["Six result records write_receipts"] Card["Compact card result_card"] Fixture --> Scan --> Checks --> Merge --> Status --> Result records --> Card
\n
Negative cases
\n
The full rejection set a green run needs
\n
A green run needs the whole rejection set to fire \u2014 eight route and lease cases plus five hook-shadow cases \u2014 and a missing or unexpected case blocks the run instead. The fixture must trigger the full rejection set for a green run. The route and lease cases are: an actor-axis mismatch on an advisory row that claims mutation authority; a missing route lease; a private transcript body; a duplicate trace id; a route-compliance overclaim of behavior change with no evidence; kernel bloat before direct action; static route metadata without trace feedback; and a route miss replaced by a public route. The hook-shadow cases are: missing authority; a banned-route intervention; command displacement; a live-state read attempt; and a hook-shadow budget overrun. If any expected case fails to fire, or an unexpected case appears, the run is marked blocked rather than green.
\n
Prior Art Grounding
\n
Distributed tracing meets agent trajectories
\n
Its shape borrows from distributed tracing \u2014 W3C Trace Context and OpenTelemetry \u2014 and from ReAct trajectories, exposing route, lease, trace id, evidence chain, validator, and scope limit together. This component borrows its shape from distributed tracing and agent trajectory work. The W3C Trace Context recommendation and OpenTelemetry set the observability pattern: propagate trace identity, collect events, and keep enough context to debug a distributed transaction. Agent work such as ReAct made the interleaved reasoning-and-action trajectory a first-class object for reading agent behavior. The traceability shape carries over: selected route, route lease, trace id, work-event-evidence chain, validator ref, and scope limit are exposed together.
\n
Source-Backed Doctrine Packet
\n
Authority rows connect claims to runtime evidence
\n
The registry, evidence-class row, and governing standard let readers trace the public claim to its exact semantic-validator boundary. A reader can move from the public claim to the runtime, standard, manifests, result records, and negative cases without guessing. The authority rows are core/organ_registry.json::implemented_organs[organ_id=agent_route_observability_runtime] and core/organ_evidence_classes.json::organ_evidence_classes[agent_route_observability_runtime]. The registry declares evidence class semantic_validator. The governing standard is standards/std_microcosm_agent_route_observability_runtime.json, whose authority boundary is public_route_observability_runtime_metadata_and_copied_macro_trace_bodies_not_live_session_provider_browser_hud_or_hook_authority.
\n
Source manifests expose copied bodies without live sessions
\n
The runtime and three source manifests keep copied bodies inspectable with body_in_receipt=false while granting no access to live sessions. The runtime source is src/microcosm_core/organs/agent_route_observability_runtime.py. Copied public source-module bodies stay inspectable through their manifests with body_in_receipt=false, including examples/agent_route_observability_runtime/exported_observability_bundle/source_module_manifest.json and the two shared projection-import manifests examples/macro_projection_import_protocol/exported_projection_import_bundle/agent_execution_trace_source_module_manifest.json and examples/macro_projection_import_protocol/exported_projection_import_bundle/agent_observability_source_module_manifest.json. The cross-module test tests/test_macro_projection_import_protocol.py::test_agent_execution_trace_body_import_is_unified_under_macro_projection_spine keeps that body import under one shared spine. This packet does not inspect live sessions.
\n
The registry validator command is:
\n
python -m microcosm_core.organs.agent_route_observability_runtime run --input fixtures/first_wave/agent_route_observability_runtime/input --out receipts/first_wave/agent_route_observability_runtime
The first artifact fits one viewport in a fixed order \u2014 command, route, work, evidence chain, authority boundary \u2014 and any browser or video view is a projection that keeps those visible. The first artifact a reader sees fits one viewport and keeps this order: the local action command, the selected route with a short reason, the work transaction with its state, the event and evidence chain with its class, and the authority boundary beside the positive claim. A browser view, screenshot, or video is a projection of that board, not a separate claim, and it must keep the command, evidence ref, and scope limit visible before any decoration.
\n
Validation Result record Path
\n
From microcosm-substrate/, reproduce this page's proof boundary with temporary result records:
A pass means the fixture and bundle produced coherent, metadata-only route feedback within the scope limit and the focused regression still holds, but the result records certify no live behavior. A pass means the fixture and exported bundle produced coherent route feedback and metadata-only result records within the scope limit, and the focused regression still holds. The focused test is the proof consumer for this section, covering streaming JSONL loaders, duplicate-key rejection, source-module manifest contracts, digest and line-count helpers, redacted result records, and the required negative cases. These validation result records do not certify live behavior.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
A green run shows only that a public synthetic fixture or bundle carries coherent, self-consistent route feedback and metadata-only records with every planned rejection firing. The strongest claim the evidence supports: a green run shows that a public synthetic fixture or exported bundle carries coherent, self-consistent route feedback and metadata-only result records, with every planned rejection firing. That is the whole proof boundary.
\n
The metadata boundary it refuses to cross
\n
It reads no live operator, provider, or account state and mutates, installs, authorizes, promotes, or releases nothing, holding to the metadata boundary rather than any live-trace authority. What it refuses: it does not read live operator, provider, browser, HUD, or account state, mutate a route or source, install a hook, authorize a provider, promote a candidate axiom, prove benchmark behavior, or approve launch. The scope limit is the metadata boundary, not a live-trace authority. The exact scope limit from the registry reads:
\n
validates only public recorded route-feedback and observability metadata fixtures, including route-lease consumption, trace attribution, hook-shadow advisory status, anti-pattern debt retirement, behavior-change evidence gates, and public source body import refs; does not read live operator/provider/browser UI/account state, mutate work log or source, install hooks, certify runtime behavior, authorize pattern assimilation, whole-system equivalence, launch, public sharing, or whole-system correctness
\n",
"components": [
{
"anchor": "component-agent_route_observability_runtime",
"component_id": "agent_route_observability_runtime",
"public_label": "Agent Route Observability Runtime"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 5,
"principle": 2
},
"link_count": 11,
"links": [
{
"id": "agent_route_observability_runtime",
"kind": "organ",
"kind_label": "component",
"label": "Agent Route Observability Runtime",
"relation": "explains",
"url": "component-agent-route-observability-runtime.html"
},
{
"id": "mechanism.agent_route_observability_runtime.validates_public_route_feedback",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public route feedback",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-route-observability-runtime-validates-public-route-feedback"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.routing_anti_patterns_registry",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Routing Anti-Patterns Registry",
"relation": "depends on",
"url": "paper-module-routing-anti-patterns-registry.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.agent_memory_temporal_conflict_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Memory Temporal-Conflict Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[32:paper_module.agent_memory_temporal_conflict_replay].depends_on",
"url": "paper-module-agent-memory-temporal-conflict-replay.html"
},
{
"id": "paper_module.batch11_saturation_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 11 Saturation Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[73:paper_module.batch11_saturation_engines_capsule].depends_on",
"url": "paper-module-batch11-saturation-engines-capsule.html"
},
{
"id": "paper_module.batch4_proof_authority_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 4 Proof, Authority, and Runtime Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[74:paper_module.batch4_proof_authority_runtime].depends_on",
"url": "paper-module-batch4-proof-authority-runtime.html"
},
{
"id": "paper_module.batch6_unsurfaced_primitives_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 6 Unsurfaced Primitives Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[75:paper_module.batch6_unsurfaced_primitives_capsule].depends_on",
"url": "paper-module-batch6-unsurfaced-primitives-capsule.html"
},
{
"id": "paper_module.batch7_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Source Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[69:paper_module.batch7_macro_engines_capsule].depends_on",
"url": "paper-module-batch7-macro-engines-capsule.html"
},
{
"id": "paper_module.batch8_compliance_pipeline_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Compliance Pipeline Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[59:paper_module.batch8_compliance_pipeline_capsule].depends_on",
"url": "paper-module-batch8-compliance-pipeline-capsule.html"
},
{
"id": "paper_module.belief_state_process_reward_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Belief-State Process Reward Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[36:paper_module.belief_state_process_reward_replay].depends_on",
"url": "paper-module-belief-state-process-reward-replay.html"
},
{
"id": "paper_module.bounded_autonomy_campaign_packet",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Bounded Autonomy Campaign Packet",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[45:paper_module.bounded_autonomy_campaign_packet].depends_on",
"url": "paper-module-bounded-autonomy-campaign-packet.html"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[13:paper_module.cold_reader_route_map].depends_on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.computer_use_action_trace_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Computer-Use Action Trace Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[46:paper_module.computer_use_action_trace_replay].depends_on",
"url": "paper-module-computer-use-action-trace-replay.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[19:paper_module.macro_projection_import_protocol].depends_on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[1:paper_module.navigation_hologram_route_plane].depends_on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[9:paper_module.pattern_binding_contract].depends_on",
"url": "paper-module-pattern-binding-contract.html"
},
{
"id": "paper_module.provider_context_recipe_budget",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Provider Context Recipe Budget",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[54:paper_module.provider_context_recipe_budget].depends_on",
"url": "paper-module-provider-context-recipe-budget.html"
},
{
"id": "paper_module.routing_anti_patterns_registry",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Routing Anti-Patterns Registry",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[57:paper_module.routing_anti_patterns_registry].depends_on",
"url": "paper-module-routing-anti-patterns-registry.html"
}
],
"paper_module_id": "paper_module.agent_route_observability_runtime",
"reader_packet": {
"abstract": "The public route-observability fixture validates synthetic route feedback, route leases, hook-shadow advisory rows, anti-pattern debt, copied source trace manifests, and metadata-only result records without claiming live session authority.",
"authority_ceiling": "Public synthetic route-observability fixtures, copied source body digests, and exported-bundle result records only; no live session introspection, provider/browser UI/account authority, live hook control, benchmark-performance proof, source-file changes, launch-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.agent_route_observability_runtime.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-agent-route-observability-runtime-diagram-1",
"html_url": "docs/paper-module-agent-route-observability-runtime.html#paper-module-agent-route-observability-runtime-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Fixture[\"Fixture inputs agent_trace, hook_shadow, debt, principle_lens, egress\"] Scan[\"non-public-state scan forbidden classes rejected\"] Checks[\"Six validators route, lease, hook, debt, lens, egress\"] Merge[\"Merge observed negative cases\"] Status[\"Status pass only if every expected case fired\"] Result records[\"Six result records write_receipts\"] Card[\"Compact card result_card\"] Fixture --> Scan --> Checks --> Merge --> Status --> Result records --> Card",
"source_format": "mermaid",
"source_hash": "sha256:f8b99567b0f7cbe945c80ba932b617082371a2ed6436097edb8cfa25a644ac5b",
"source_line_end": 190,
"source_line_start": 179
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
}
],
"relation": "abides by"
},
{
"count": 5,
"links": [
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.routing_anti_patterns_registry",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Routing Anti-Patterns Registry",
"relation": "depends on",
"url": "paper-module-routing-anti-patterns-registry.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "agent_route_observability_runtime",
"kind": "organ",
"kind_label": "component",
"label": "Agent Route Observability Runtime",
"relation": "explains",
"url": "component-agent-route-observability-runtime.html"
},
{
"id": "mechanism.agent_route_observability_runtime.validates_public_route_feedback",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public route feedback",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-route-observability-runtime-validates-public-route-feedback"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.agent_route_observability_runtime",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "re-entry-conditions",
"source_line_start": 289,
"title": "Re-entry conditions"
},
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 360,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_route_observability_runtime.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "W3C Trace Context",
"section_id": "distributed-tracing-meets-agent-trajectories",
"url": "https://www.w3.org/TR/trace-context/"
},
{
"kind": "external_reference",
"label": "OpenTelemetry",
"section_id": "distributed-tracing-meets-agent-trajectories",
"url": "https://opentelemetry.io/docs/reference/specification/overview/"
},
{
"kind": "external_reference",
"label": "ReAct",
"section_id": "distributed-tracing-meets-agent-trajectories",
"url": "https://arxiv.org/abs/2210.03629"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It replays one agent's recorded trail \u2014 route, work, events, evidence, and declared authority limit \u2014 and checks the whole thing is well-formed and self-consistent.** This component takes a recorded trail of one agent's local run, the route it picked, the work it logged, the events it emitted, the evidence it pointed to, and the authority limit it declared, and checks that the trail is well-formed and self-consistent. It is a replay validator over synthetic fixtures, not a live tap on a running agent.",
"level": 4,
"order": 1,
"section_id": "one-recorded-trail-is-checked-end-to-end",
"source_line_end": 13,
"source_line_start": 3,
"title": "One recorded trail is checked end to end"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Because the record states its authority limit up front, the runtime runs six checks and writes six metadata-only result records that each carry status, findings, and the scope limit.** The record is built to state up front where the agent's authority was supposed to stop, so the limit is written down and checkable rather than taken on faith. The runtime reads that record, runs six checks over it, and writes six small JSON result records. Each result record carries a status, the negative cases it saw, the findings, and the scope limit. Large or private payloads are dropped before anything is written.",
"level": 4,
"order": 2,
"section_id": "six-checks-share-the-declared-scope-limit",
"source_line_end": 25,
"source_line_start": 14,
"title": "Six checks share the declared scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It reads recorded evidence only, inspecting no live session and proving nothing about whether an agent actually stayed in bounds.** It reads recorded evidence. It does not inspect live sessions, mutate a route, install a hook, authorize a provider, or prove that any agent actually stayed in bounds.",
"level": 4,
"order": 3,
"section_id": "recorded-evidence-cannot-prove-live-compliance",
"source_line_end": 33,
"source_line_start": 26,
"title": "Recorded evidence cannot prove live compliance"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 35,
"source_line_start": 34,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A raw agent log is hard to trust because a behavior-change claim or a quiet authority grab reads as proof while nothing checkable is recorded.** A raw agent log is hard to trust. Someone reading it by hand cannot easily tell whether a row that claims \"I changed how the agent behaves\" is backed by anything, or whether an advisory step quietly asserted authority it never had. Loose logging fails here in a specific way: a trace that reads as proof, with nothing recorded that a reader can check.",
"level": 4,
"order": 5,
"section_id": "a-raw-log-reads-as-proof",
"source_line_end": 45,
"source_line_start": 36,
"title": "A raw log reads as proof"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It refuses to treat a claim as evidence unless the record supports it: advisory rows cannot claim mutation authority, behavior-change needs named trace ids, and private bodies are redacted.** This component closes that gap by refusing to treat a claim as evidence unless the record supports it. An advisory row cannot also claim live mutation authority. A behavior-change claim is blocked unless it names the trace ids that evidence it. A private transcript body is rejected and redacted. The useful output is a compact result record: one command, one selected route, one work-event-evidence chain, one validator handle, and one scope limit.",
"level": 4,
"order": 6,
"section_id": "no-claim-counts-unless-the-record-backs-it",
"source_line_end": 57,
"source_line_start": 46,
"title": "No claim counts unless the record backs it"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 59,
"source_line_start": 58,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**run loads the fixture, scans the inputs for forbidden non-public-state classes before any check, then calls six validators in order and merges what they observed.** The default entrypoint is run. It loads the first-wave fixture, resolves the public root for the input path, and scans the inputs for forbidden non-public-state classes before any check runs. It then calls six validators in order and merges what they observed.",
"level": 4,
"order": 8,
"section_id": "run-scans-then-calls-six-validators",
"source_line_end": 69,
"source_line_start": 60,
"title": "`run` scans, then calls six validators"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The fixture is deliberately adversarial, so a green run means every expected negative case fired and nothing unexpected did, not that the trace was clean.** The fixture holds 10 trace rows, 6 hook-shadow cases, 6 egress cases, and 2 anti-pattern debt rows. It is deliberately adversarial: it is built to trigger every planned rejection. A green run therefore means every expected negative case fired and nothing unexpected did, not that the trace was clean.",
"level": 4,
"order": 9,
"section_id": "a-green-run-means-every-rejection-fired",
"source_line_end": 78,
"source_line_start": 70,
"title": "A green run means every rejection fired"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateroute_compliance checks each trace row for duplicate ids, a missing lease, a route miss, a private payload, an advisory authority claim, or unevidenced behavior change, passing only on expected cases.** validateroutecompliance classifies the trace rows. It counts duplicate event ids, and for each row it checks for a missing route lease, a route miss where the observed route differs from the expected route, a private transcript payload class, an advisory row that claims mutation authority, and a behavior-change claim with no evidence ids. Each check that fires records a typed finding and marks the row rejected. The check passes when every negative case it saw is one of the expected c\u2026",
"level": 4,
"order": 10,
"section_id": "validate-route-compliance-classifies-each-row",
"source_line_end": 91,
"source_line_start": 79,
"title": "`validate_route_compliance` classifies each row"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validaterouteleasemode_control flags kernel bloat before direct action and leases that offer static route metadata in place of consumed trace feedback. validaterouteleasemode_control reads the same trace rows as route leases. It flags KERNELBLOATBEFOREDIRECTACTION when a lease returns to broad kernel context before acting, and ROUTELEASENOT_CONSUMED when static route metadata is offered in place of consumed trace feedback.",
"level": 4,
"order": 11,
"section_id": "validate-route-lease-mode-control-reads-leases",
"source_line_end": 101,
"source_line_start": 92,
"title": "`validate_route_lease_mode_control` reads leases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatehookshadow replays each hook case, requiring projectionnot_authority, metadata-only fields, and every live-state field false, so any read of live operator, provider, or operator UI state is rejected. validatehook_shadow replays the hook intervention cases. Each case must map a case id, hook id, coverage claim, repair class, and expected intervention, must declare projectionnotauthority, must stay metadata-only, and must keep every live-state field false. A case that reads live operator, provider, browser, HUD, cockpit, or ledger state is rejected.",
"level": 4,
"order": 12,
"section_id": "validate-hook-shadow-keeps-hooks-projection-only",
"source_line_end": 112,
"source_line_start": 102,
"title": "`validate_hook_shadow` keeps hooks projection-only"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatedebt_retirement retires a debt row only when it names behavior-change evidence trace ids, so evidence-only telemetry cannot retire debt on its own. validatedebt_retirement marks a debt row retired only when it names behavior-change evidence trace ids and carries no evidence-only trace ids. Evidence-only telemetry cannot retire debt on its own.",
"level": 4,
"order": 13,
"section_id": "validate-debt-retirement-needs-behavior-change-evidence",
"source_line_end": 121,
"source_line_start": 113,
"title": "`validate_debt_retirement` needs behavior-change evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateagentprinciplelens confirms the lens preserves its selected principle ids and route handles while minting no principle, promoting no axiom, and exporting no full body. validateagentprinciplelens confirms the lens preserves its selected principle ids and its route handles back to the principle cards and the agent operating packet. The lens cannot mint principles, promote a candidate axiom, or export full principle bodies. Any of those sets a finding.",
"level": 4,
"order": 14,
"section_id": "validate-agent-principle-lens-mints-nothing",
"source_line_end": 131,
"source_line_start": 122,
"title": "`validate_agent_principle_lens` mints nothing"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateegress_mirror recomputes each detector decision against the declared one, requiring every private-read, provider-read, cockpit-read, and body-export flag to stay false. validateegress_mirror recomputes each detector decision and compares it to the declared one. Every case must keep privatestateread, providerpayloadread, browserhudcockpitstateread, and body_exported false.",
"level": 4,
"order": 15,
"section_id": "validate-egress-mirror-recomputes-each-decision",
"source_line_end": 140,
"source_line_start": 132,
"title": "`validate_egress_mirror` recomputes each decision"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**run passes only when no expected case is missing, the scan is clean, and all six validators pass, after which write_receipts emits six metadata-only JSON records and resultcard a compact card. run passes only when no expected negative case is missing, the non-public-state scan passes, and route compliance, hook shadow, route trace analytics, the principle lens, and the egress mirror all pass. It then calls write_receipts, which writes the six JSON result records under the public root and returns their public-relative paths. result_card compresses a full result into a start card and drops the full findings, private scans, source bodies, and scope-limit bodies.",
"level": 4,
"order": 16,
"section_id": "a-pass-writes-six-body-free-receipts",
"source_line_end": 153,
"source_line_start": 141,
"title": "A pass writes six metadata-only result records"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "runobservability_bundle validates a larger exported bundle and blocks on any forbidden payload key or manifest drift, while companion runners check the same membrane over adjacent evidence slices.** Alongside the fixture path, runobservabilitybundle validates a larger exported bundle: route events, agent path observations, session diagnostics, hook-shadow rows, actor-axis checks, debt rows, process-audit rows, an observability policy, and a source-module manifest, blocking on any forbidden payload key or manifest drift. Companion runners such as runagentobservabilitystorebundle check the same membrane over adjacent public route, session, bridge, controller, and store evidence slices.",
"level": 4,
"order": 17,
"section_id": "bundle-runners-over-exported-evidence-slices",
"source_line_end": 191,
"source_line_start": 154,
"title": "Bundle runners over exported evidence slices"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "negative-cases",
"source_line_end": 193,
"source_line_start": 192,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A green run needs the whole rejection set to fire \u2014 eight route and lease cases plus five hook-shadow cases \u2014 and a missing or unexpected case blocks the run instead.** The fixture must trigger the full rejection set for a green run. The route and lease cases are: an actor-axis mismatch on an advisory row that claims mutation authority; a missing route lease; a private transcript body; a duplicate trace id; a route-compliance overclaim of behavior change with no evidence; kernel bloat before direct action; static route metadata without trace feedback; and a route miss replaced by a public route. The hook-shadow cases are: missing authority; a banned-route intervention; command displacemen\u2026",
"level": 4,
"order": 19,
"section_id": "the-full-rejection-set-a-green-run-needs",
"source_line_end": 208,
"source_line_start": 194,
"title": "The full rejection set a green run needs"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "prior-art-grounding",
"source_line_end": 210,
"source_line_start": 209,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Its shape borrows from distributed tracing \u2014 W3C Trace Context and OpenTelemetry \u2014 and from ReAct trajectories, exposing route, lease, trace id, evidence chain, validator, and scope limit together.** This component borrows its shape from distributed tracing and agent trajectory work. The W3C Trace Context recommendation and OpenTelemetry set the observability pattern: propagate trace identity, collect events, and keep enough context to debug a distributed transaction. Agent work such as ReAct made the interleaved reasoning-and-action trajectory a first-class object for reading agent behavior. The traceability shape carries over: selected route, route lease, trace id, work-event-evidence c\u2026",
"level": 4,
"order": 21,
"section_id": "distributed-tracing-meets-agent-trajectories",
"source_line_end": 226,
"source_line_start": 211,
"title": "Distributed tracing meets agent trajectories"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "source-backed-doctrine-packet",
"source_line_end": 228,
"source_line_start": 227,
"title": "Source-Backed Doctrine Packet"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The registry, evidence-class row, and governing standard let readers trace the public claim to its exact semantic-validator boundary. A reader can move from the public claim to the runtime, standard, manifests, result records, and negative cases without guessing. The authority rows are core/organregistry.json::implementedorgans[organid=agentrouteobservabilityruntime] and core/organevidenceclasses.json::organevidenceclasses[agentrouteobservability_runtime]. The registry declares evidence class semantic_validator. The governing standard is standards/stdmicrocosmagentrouteobservability_runtime.json, whose authority boundary is publicrouteobservabilityruntimemetadataandcopiedmacrotracebodiesnot\u2026",
"level": 4,
"order": 23,
"section_id": "authority-rows-connect-claims-to-runtime-evidence",
"source_line_end": 242,
"source_line_start": 229,
"title": "Authority rows connect claims to runtime evidence"
},
{
"block_kinds": [
"code",
"list",
"paragraph"
],
"excerpt": "The runtime and three source manifests keep copied bodies inspectable with bodyinreceipt=false while granting no access to live sessions. The runtime source is src/microcosmcore/components/agentrouteobservabilityruntime.py. Copied public source-module bodies stay inspectable through their manifests with bodyinreceipt=false, including examples/agentrouteobservabilityruntime/exportedobservabilitybundle/sourcemodule_manifest.json and the two shared projection-import manifests examples/macroprojectionimportprotocol/exportedprojectionimportbundle/agentexecutiontracesourcemodule_manifest.json and examples/macroprojectionimportprotocol/exportedprojectionimportbundle/agentobservabilitysourcemodulem\u2026",
"level": 4,
"order": 24,
"section_id": "source-manifests-expose-copied-bodies-without-live-sessions",
"source_line_end": 274,
"source_line_start": 243,
"title": "Source manifests expose copied bodies without live sessions"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 25,
"section_id": "observable-first-artifact-contract",
"source_line_end": 276,
"source_line_start": 275,
"title": "Observable First Artifact Contract"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The first artifact fits one viewport in a fixed order \u2014 command, route, work, evidence chain, authority boundary \u2014 and any browser or video view is a projection that keeps those visible.** The first artifact a reader sees fits one viewport and keeps this order: the local action command, the selected route with a short reason, the work transaction with its state, the event and evidence chain with its class, and the authority boundary beside the positive claim. A browser view, screenshot, or video is a projection of that board, not a separate claim, and it must keep the command, evidence ref, and scope limit visible before any decoration.",
"level": 4,
"order": 26,
"section_id": "one-viewport-in-a-fixed-order",
"source_line_end": 288,
"source_line_start": 277,
"title": "One viewport in a fixed order"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate/, reproduce this page's proof boundary with temporary result records:",
"level": 2,
"order": 27,
"section_id": "validation-receipt-path",
"source_line_end": 320,
"source_line_start": 300,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the fixture and bundle produced coherent, metadata-only route feedback within the scope limit and the focused regression still holds, but the result records certify no live behavior.** A pass means the fixture and exported bundle produced coherent route feedback and metadata-only result records within the scope limit, and the focused regression still holds. The focused test is the proof consumer for this section, covering streaming JSONL loaders, duplicate-key rejection, source-module manifest contracts, digest and line-count helpers, redacted result records, and the required negative cases. These validation result records do not certify live behavior.",
"level": 4,
"order": 28,
"section_id": "what-a-pass-certifies-and-does-not",
"source_line_end": 332,
"source_line_start": 321,
"title": "What a pass certifies, and does not"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 29,
"section_id": "claim-ceiling",
"source_line_end": 334,
"source_line_start": 333,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A green run shows only that a public synthetic fixture or bundle carries coherent, self-consistent route feedback and metadata-only records with every planned rejection firing.** The strongest claim the evidence supports: a green run shows that a public synthetic fixture or exported bundle carries coherent, self-consistent route feedback and metadata-only result records, with every planned rejection firing. That is the whole proof boundary.",
"level": 4,
"order": 30,
"section_id": "the-strongest-supported-claim",
"source_line_end": 344,
"source_line_start": 335,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "**It reads no live operator, provider, or account state and mutates, installs, authorizes, promotes, or releases nothing, holding to the metadata boundary rather than any live-trace authority.** What it refuses: it does not read live operator, provider, browser, HUD, or account state, mutate a route or source, install a hook, authorize a provider, promote a candidate axiom, prove benchmark behavior, or approve launch. The scope limit is the metadata boundary, not a live-trace authority. The exact scope limit from the registry reads:",
"level": 4,
"order": 31,
"section_id": "the-metadata-boundary-it-refuses-to-cross",
"source_line_end": 359,
"source_line_start": 345,
"title": "The metadata boundary it refuses to cross"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/agent_route_observability_runtime.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public route-observability fixture validates synthetic route feedback, route leases, hook-shadow advisory rows, anti-pattern debt, copied source trace manifests, and metadata-only result records without claiming live session authority."
}
},
"source_drilldown": {
"source_manifests": [
{
"label": "Source manifest",
"path": "examples/agent_route_observability_runtime/exported_observability_bundle/source_module_manifest.json",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/examples/agent_route_observability_runtime/exported_observability_bundle/source_module_manifest.json"
},
{
"label": "Source manifest",
"path": "examples/macro_projection_import_protocol/exported_projection_import_bundle/agent_execution_trace_source_module_manifest.json",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/examples/macro_projection_import_protocol/exported_projection_import_bundle/agent_execution_trace_source_module_manifest.json"
},
{
"label": "Source manifest",
"path": "examples/macro_projection_import_protocol/exported_projection_import_bundle/agent_observability_source_module_manifest.json",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/examples/macro_projection_import_protocol/exported_projection_import_bundle/agent_observability_source_module_manifest.json"
}
],
"source_modules": [
{
"label": "agent_route_observability_runtime.py",
"path": "src/microcosm_core/organs/agent_route_observability_runtime.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public route-feedback fixtures, observability bundles, route-lease mode control, agent-principle-lens admission, source-manifest checks, and metadata-only result records.",
"symbols": [
"run",
"run_observability_bundle",
"run_agent_observability_store_bundle",
"validate_route_compliance",
"validate_route_lease_mode_control",
"validate_agent_principle_lens"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/agent_route_observability_runtime.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_route_observability_runtime.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_route_observability_runtime.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Agent Route Observability Runtime is the public evidence membrane for recorded agent route feedback. It checks actor-axis boundaries, selected and replacement routes, route-lease consumption, duplicate trace ids, hook-shadow advisory status, anti-pattern debt retirement, agent-principle-lens admission, egress-mirror boundaries, source-module manifests, and non-public-state exclusion while keeping transcript, provider, browser, HUD, account, account secret, and live-hook bodies out of result records.",
"summary_line": "The public route-observability fixture validates synthetic route feedback, route leases, hook-shadow advisory rows, anti-pattern debt, copied source trace manifests, and metadata-only result records without claiming live session authority.",
"title": "Agent Route Observability Runtime"
},
{
"anchor": "paper-module-agent-benchmark-integrity-anti-gaming-replay",
"authority_ceiling": "Copied public source pattern provenance bodies and metadata-only synthetic benchmark-integrity replay result records only; no benchmark claims, SWE-bench performance claim, hidden-gold access, oracle patch body export, private issue body export, external model access, live repository mutation, launch-scope decision, publishing-scope decision, source-file changes, or product-progress evidence.",
"body_html": "
Every claimed pass is recomputed
\n
A claimed repair-benchmark "pass" is downgraded to a claim and recomputed from public references before this component will repeat it. Agent benchmark numbers are easy to state and hard to trust. This component takes a set of claimed repository-repair passes and refuses to repeat the word "pass" until each one can be replayed from public references that name its evaluator, its frozen configuration, and the evidence that the run was not gamed. It recomputes every verdict from scratch. A row that simply declares integrity_pass is treated as a claim to be checked, not as proof.
\n
Public references in, metadata-only verdicts out
\n
The inputs are public case ids, evaluator ids, configuration hashes, and evidence references; the output is a metadata-only record of which replays cleared the boundary and why. The component reads a fixture that models a repair benchmark with public case ids, evaluator ids, configuration hashes, and reference paths to file-access logs, contamination checks, trusted-reference scores, and output replays. It keeps issue bodies, oracle patches, hidden answers, and model-output data out of every result record. What it emits is a metadata-only record: which case replays cleared the boundary, which were quarantined, and why.
\n
A consistency checker, not a leaderboard
\n
It is a consistency checker over public references, not a benchmark runner, a leaderboard, or a score. It checks that a set of public references is internally consistent and traces back to a command that actually ran.
\n
Purpose
\n
What a bare pass-rate headline hides
\n
A headline pass-rate hides the evaluator, configuration, leakage, and selection decisions that produced it; this gate makes the claim carry them or be quarantined. A single headline like "passes N percent of repair tasks" hides every decision that produced it: which evaluator ran, whether its configuration was frozen, whether the agent could see held-out answers, whether the test cases leaked into training, and whether one lucky attempt was promoted as the result. A loose benchmark claim carries none of that. This component exists to make the claim carry it, or be quarantined.
\n
Declaring success is not a way through
\n
Every reference the policy names must exist and pass, or the row recomputes to quarantine no matter what it declared. The gate is deliberately hostile to the easy path. The evaluator id must be on a locked list. The configuration hash must be one the policy declared in advance. File-access, contamination, trusted-reference, and output-replay references must exist and pass. The case id must be registered up front. If any of those is missing or contradicted, the row is recomputed as quarantine no matter what it declared.
\n
A real command run backs every pass
\n
Each integrity_pass must cite a digest-bound artifact from a completed, exit-zero run of this component's focused test, so no claim rests on prose alone. An integrity_pass must be backed by a sanitised real command run, not only by hand-written references. Each passing row cites a real_benchmark_trace_ref that has to resolve to a copied artifact carrying a completed, exit-zero run of this component's focused test, with sha256 digests bound to a recorded run id and an explicit list of omitted live material. This is what stops a benchmark claim from resting on prose.
\n
How it works
\n
One orchestrator, two entry points
\n
_build_result loads the four fixture inputs and runs every validation stage in order; run and run_benchmark_integrity_bundle are thin wrappers over it._build_result is the orchestrator. It loads the four fixture inputs (projection_protocol.json, locked_evaluator_policy.json, benchmark_cases.json, replay_observations.json), then runs each validation stage in order and returns one metadata-only result. The entry points run and run_benchmark_integrity_bundle wrap it. run validates the fixture with the twelve negative-case files included. run_benchmark_integrity_bundle validates the exported public bundle without them, and can reuse a prior result record when _freshness_basis shows the inputs, validator source, and evidence digests are unchanged.
\n
Function
What it does
validate_source_module_imports
Reads source_module_manifest.json, checks each copied body is copied_non_secret_macro_body with body_in_receipt=false, and recomputes its sha256 against disk.
_real_trace_artifact_findings
Validates the sanitised real-run artifact: schema, completed status with exit_code 0, the focused pytest argument shape, a passing test summary, sha256 digests, and declared omissions.
validate_locked_evaluator_policy
Indexes the evaluator ids and configuration hashes the policy allows.
validate_benchmark_cases
Confirms the case roster and held-out guard ids.
_validate_replay_row
Recomputes one row's verdict from every check below.
validate_public_trace
Recomputes each verdict a second time from independent trace spans and flags any mismatch.
\n
Per-row recomputation is the centre of the mechanism
\n
A row passes only when every required field is present, its evaluator and configuration hash are policy-locked, its real-run trace verifies, and no evidence check fails. The center of the mechanism is _validate_replay_row. It collects the required fields listed in REQUIRED_REPLAY_FIELDS, checks the evaluator id is locked and its configuration hash is policy-declared, checks the case id is in the roster, and confirms the real_benchmark_trace_ref resolves to a digest-verified real-run artifact that is also cited among the row's evidence references. For a positive row it opens the file-access, contamination, trusted-reference, and output-replay evidence and folds any failing check or contamination flag into a reason list. The rule at the end is one line: the computed verdict is integrity_pass only when no required field is missing and the reason list is empty, otherwise quarantine.
\n
An independent second pass over the public trace
\n
validate_public_trace recomputes every verdict a second time from independent trace spans and turns any disagreement into a blocking finding. It reads the trace built by build_public_benchmark_integrity_anti_gaming_trace, which recomputes each verdict from contamination, file-access, and locked-evaluator spans. If a span's recomputed verdict disagrees with the declared one, or its configuration hash is not policy-declared, or its evaluator is not locked, the disagreement becomes a finding. _build_result returns status=pass only when the non-public-state scan finds no blocking hit, every sub-validator passes, and no expected negative case is missing.
\nDiagram of the mechanism (6 steps).Diagram source & refs
Three public cases recompute to two passes and one quarantine, and the public trace has to agree before the run can pass. On the shipped fixture the three cases repo_issue_public_001, repo_issue_public_002, and repo_issue_public_003 recompute to two integrity_pass rows and one quarantine.
\n
Negative cases
\n
Twelve gaming attempts the fixture must catch
\n
Each named negative case maps to a required reason code, and a run fails if any expected case is not observed.EXPECTED_NEGATIVE_CASES names twelve gaming attempts the fixture must catch, each mapped to the reason code the component has to raise. They are an evaluator edit attempt, train/test leakage, a real-trace variant of train/test leakage, an oracle patch body leak, hidden-gold access, final-answer-only grading, a model-output data leak, a score overclaim, pass-k cherry-picking, an admitted misleading test, a private issue body leak, and an unregistered case replay. The negative fixtures are a required floor rather than decoration.
\n
Prior Art Grounding
\n
Goodhart, reward hacking, and leakage accounting
\n
The design follows the reward-hacking and benchmark-leakage literature: a metric stops being evidence once it becomes a target, so the references must carry the proof. The component draws on the long-standing observation that an optimized metric loses its evidential force once it becomes a target, and on the AI-safety literature on reward hacking and specification gaming. Concrete Problems in AI Safety frames reward hacking as a practical accident risk, DeepMind's specification-gaming survey collects concrete examples, and Benchmarking Benchmark Leakage in Large Language Models motivates explicit leakage documentation. The engineering pattern is simple accounting: evaluator ids, configuration hashes, case rosters, and evidence references must be present and consistent before benchmark-style language is allowed.
\n
Validation Result record Path
\n
Fixture output stays outside the working tree
\n
The fixture validator runs from the repository root and writes its result to a disposable path outside the working tree. Run the fixture validator from the repository root and write its result record outside the working tree:
A pass means every verdict was recomputed from the public references and agreed with the independent trace; it does not mean a benchmark claims was produced. A pass means the fixture and bundle validators recomputed every verdict from public case ids, locked evaluator references, configuration hashes, evidence references, manifest digests, and the twelve negative cases, and agreed with the independent trace.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
On this fixture and bundle, a metadata-only replay respected the anti-gaming boundary, and AUTHORITY_CEILING stops the proof there. The strongest claim the evidence supports is narrow: on this public fixture and the exported bundle, a metadata-only replay respected an anti-gaming boundary over public case ids and locked evaluator references, and the two positive rows each trace back to a passing command run. The AUTHORITY_CEILING constant sets every outward authorization flag to false, so the proof boundary stops there.
\n
What stays false even on a pass
\n
A green run still claims no benchmark claims, no repair capability, and no access to hidden or private material. It claims no benchmark claims, no SWE-bench result, and no repair capability. It does not access hidden-gold answers, oracle patches, or private issue text, does not call a provider, does not mutate a live repository, and excludes public sharing or launch. Those remain false even when the fixture passes.
\n",
"components": [
{
"anchor": "component-agent_benchmark_integrity_anti_gaming_replay",
"component_id": "agent_benchmark_integrity_anti_gaming_replay",
"public_label": "Agent Benchmark Integrity Anti Gaming Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 2
},
"link_count": 7,
"links": [
{
"id": "agent_benchmark_integrity_anti_gaming_replay",
"kind": "organ",
"kind_label": "component",
"label": "Agent Benchmark Integrity Anti Gaming Replay",
"relation": "explains",
"url": "component-agent-benchmark-integrity-anti-gaming-replay.html"
},
{
"id": "mechanism.agent_benchmark_integrity_anti_gaming_replay.validates_public_benchmark_integrity_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public benchmark integrity replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-benchmark-integrity-anti-gaming-replay-validates-public-benchmark-integrity-replay"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch10_cold_eval_honesty_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 10 Cold Eval Honesty Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[67:paper_module.batch10_cold_eval_honesty_capsule].depends_on",
"url": "paper-module-batch10-cold-eval-honesty-capsule.html"
},
{
"id": "paper_module.research_replication_rubric_artifact_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Research Replication Rubric Artifact Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[4:paper_module.research_replication_rubric_artifact_replay].depends_on",
"url": "paper-module-research-replication-rubric-artifact-replay.html"
}
],
"paper_module_id": "paper_module.agent_benchmark_integrity_anti_gaming_replay",
"reader_packet": {
"abstract": "The agent benchmark integrity anti-gaming replay validates copied public source pattern provenance bodies and metadata-only benchmark replay rows before any score-like language is allowed.",
"authority_ceiling": "Copied public source pattern provenance bodies and metadata-only synthetic benchmark-integrity replay result records only; no benchmark claims, SWE-bench performance claim, hidden-gold access, oracle patch body export, private issue body export, external model access, live repository mutation, launch-scope decision, publishing-scope decision, source-file changes, or product-progress evidence.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.agent_benchmark_integrity_anti_gaming_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-agent-benchmark-integrity-anti-gaming-replay-diagram-1",
"html_url": "docs/paper-module-agent-benchmark-integrity-anti-gaming-replay.html#paper-module-agent-benchmark-integrity-anti-gaming-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Inputs[\"projection, policy, cases, replay rows\"] --> Source[\"validate_source_module_imports manifest digest + real-run floor\"] Inputs --> Rows[\"_validate_replay_row recompute per row\"] Source --> Rows Rows --> Verdict[\"integrity_pass or quarantine\"] Verdict --> Trace[\"validate_public_trace independent recompute\"] Trace --> Result[\"metadata-only result record\"]",
"source_format": "mermaid",
"source_hash": "sha256:67fbd9e1ad20cc652ed572d35e3636503380a58e77948d8c85da18893e96c98a",
"source_line_end": 126,
"source_line_start": 118
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "agent_benchmark_integrity_anti_gaming_replay",
"kind": "organ",
"kind_label": "component",
"label": "Agent Benchmark Integrity Anti Gaming Replay",
"relation": "explains",
"url": "component-agent-benchmark-integrity-anti-gaming-replay.html"
},
{
"id": "mechanism.agent_benchmark_integrity_anti_gaming_replay.validates_public_benchmark_integrity_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public benchmark integrity replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-benchmark-integrity-anti-gaming-replay-validates-public-benchmark-integrity-replay"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.agent_benchmark_integrity_anti_gaming_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 224,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_benchmark_integrity_anti_gaming_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Concrete Problems in AI Safety",
"section_id": "goodhart-reward-hacking-and-leakage-accounting",
"url": "https://arxiv.org/abs/1606.06565"
},
{
"kind": "external_reference",
"label": "specification-gaming survey",
"section_id": "goodhart-reward-hacking-and-leakage-accounting",
"url": "https://deepmind.google/blog/specification-gaming-the-flip-side-of-ai-ingenuity/"
},
{
"kind": "external_reference",
"label": "Benchmarking Benchmark Leakage in Large Language Models",
"section_id": "goodhart-reward-hacking-and-leakage-accounting",
"url": "https://arxiv.org/abs/2404.18824"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A claimed repair-benchmark \"pass\" is downgraded to a claim and recomputed from public references before this component will repeat it.** Agent benchmark numbers are easy to state and hard to trust. This component takes a set of claimed repository-repair passes and refuses to repeat the word \"pass\" until each one can be replayed from public references that name its evaluator, its frozen configuration, and the evidence that the run was not gamed. It recomputes every verdict from scratch. A row that simply declares integrity_pass is treated as a claim to be checked, not as proof.",
"level": 4,
"order": 1,
"section_id": "every-claimed-pass-is-recomputed",
"source_line_end": 13,
"source_line_start": 3,
"title": "Every claimed pass is recomputed"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The inputs are public case ids, evaluator ids, configuration hashes, and evidence references; the output is a metadata-only record of which replays cleared the boundary and why.** The component reads a fixture that models a repair benchmark with public case ids, evaluator ids, configuration hashes, and reference paths to file-access logs, contamination checks, trusted-reference scores, and output replays. It keeps issue bodies, oracle patches, hidden answers, and model-output data out of every result record. What it emits is a metadata-only record: which case replays cleared the boundary, which were quarantined, and why.",
"level": 4,
"order": 2,
"section_id": "public-references-in-body-free-verdicts-out",
"source_line_end": 25,
"source_line_start": 14,
"title": "Public references in, metadata-only verdicts out"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is a consistency checker over public references, not a benchmark runner, a leaderboard, or a score.** It checks that a set of public references is internally consistent and traces back to a command that actually ran.",
"level": 4,
"order": 3,
"section_id": "a-consistency-checker-not-a-leaderboard",
"source_line_end": 31,
"source_line_start": 26,
"title": "A consistency checker, not a leaderboard"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 33,
"source_line_start": 32,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A headline pass-rate hides the evaluator, configuration, leakage, and selection decisions that produced it; this gate makes the claim carry them or be quarantined.** A single headline like \"passes N percent of repair tasks\" hides every decision that produced it: which evaluator ran, whether its configuration was frozen, whether the agent could see held-out answers, whether the test cases leaked into training, and whether one lucky attempt was promoted as the result. A loose benchmark claim carries none of that. This component exists to make the claim carry it, or be quarantined.",
"level": 4,
"order": 5,
"section_id": "what-a-bare-pass-rate-headline-hides",
"source_line_end": 44,
"source_line_start": 34,
"title": "What a bare pass-rate headline hides"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Every reference the policy names must exist and pass, or the row recomputes to quarantine no matter what it declared.** The gate is deliberately hostile to the easy path. The evaluator id must be on a locked list. The configuration hash must be one the policy declared in advance. File-access, contamination, trusted-reference, and output-replay references must exist and pass. The case id must be registered up front. If any of those is missing or contradicted, the row is recomputed as quarantine no matter what it declared.",
"level": 4,
"order": 6,
"section_id": "declaring-success-is-not-a-way-through",
"source_line_end": 54,
"source_line_start": 45,
"title": "Declaring success is not a way through"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Each integritypass must cite a digest-bound artifact from a completed, exit-zero run of this component's focused test, so no claim rests on prose alone. An integritypass must be backed by a sanitised real command run, not only by hand-written references. Each passing row cites a realbenchmarktrace_ref that has to resolve to a copied artifact carrying a completed, exit-zero run of this component's focused test, with sha256 digests bound to a recorded run id and an explicit list of omitted live material. This is what stops a benchmark claim from resting on prose.",
"level": 4,
"order": 7,
"section_id": "a-real-command-run-backs-every-pass",
"source_line_end": 65,
"source_line_start": 55,
"title": "A real command run backs every pass"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 67,
"source_line_start": 66,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "buildresult loads the four fixture inputs and runs every validation stage in order; run and runbenchmarkintegrity_bundle are thin wrappers over it. build_result is the orchestrator. It loads the four fixture inputs (projectionprotocol.json, lockedevaluator_policy.json, benchmarkcases.json, replayobservations.json), then runs each validation stage in order and returns one metadata-only result. The entry points run and runbenchmarkintegrity_bundle wrap it. run validates the fixture with the twelve negative-case files included. runbenchmarkintegrity_bundle validates the exported public bundle without them, and can reuse a prior result record when freshnessbasis shows the inputs, validator sour\u2026",
"level": 4,
"order": 9,
"section_id": "one-orchestrator-two-entry-points",
"source_line_end": 90,
"source_line_start": 68,
"title": "One orchestrator, two entry points"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A row passes only when every required field is present, its evaluator and configuration hash are policy-locked, its real-run trace verifies, and no evidence check fails. The center of the mechanism is validatereplayrow. It collects the required fields listed in REQUIREDREPLAYFIELDS, checks the evaluator id is locked and its configuration hash is policy-declared, checks the case id is in the roster, and confirms the realbenchmarktrace_ref resolves to a digest-verified real-run artifact that is also cited among the row's evidence references. For a positive row it opens the file-access, contamination, trusted-reference, and output-replay evidence and folds any failing check or contamination\u2026",
"level": 4,
"order": 10,
"section_id": "per-row-recomputation-is-the-centre-of-the-mechanism",
"source_line_end": 105,
"source_line_start": 91,
"title": "Per-row recomputation is the centre of the mechanism"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "validatepublic_trace recomputes every verdict a second time from independent trace spans and turns any disagreement into a blocking finding.** It reads the trace built by buildpublicbenchmarkintegrityantigamingtrace, which recomputes each verdict from contamination, file-access, and locked-evaluator spans. If a span's recomputed verdict disagrees with the declared one, or its configuration hash is not policy-declared, or its evaluator is not locked, the disagreement becomes a finding. buildresult returns status=pass only when the non-public-state scan finds no blocking hit, every sub-validator passes, and no expected negative case is missing.",
"level": 4,
"order": 11,
"section_id": "an-independent-second-pass-over-the-public-trace",
"source_line_end": 127,
"source_line_start": 106,
"title": "An independent second pass over the public trace"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Three public cases recompute to two passes and one quarantine, and the public trace has to agree before the run can pass.** On the shipped fixture the three cases repoissuepublic001, repoissuepublic002, and repoissuepublic003 recompute to two integritypass rows and one quarantine.",
"level": 4,
"order": 12,
"section_id": "what-the-shipped-fixture-recomputes-to",
"source_line_end": 135,
"source_line_start": 128,
"title": "What the shipped fixture recomputes to"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 137,
"source_line_start": 136,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Each named negative case maps to a required reason code, and a run fails if any expected case is not observed. EXPECTEDNEGATIVE_CASES names twelve gaming attempts the fixture must catch, each mapped to the reason code the component has to raise. They are an evaluator edit attempt, train/test leakage, a real-trace variant of train/test leakage, an oracle patch body leak, hidden-gold access, final-answer-only grading, a model-output data leak, a score overclaim, pass-k cherry-picking, an admitted misleading test, a private issue body leak, and an unregistered case replay. The negative fixtures are a required floor rather than decoration.",
"level": 4,
"order": 14,
"section_id": "twelve-gaming-attempts-the-fixture-must-catch",
"source_line_end": 149,
"source_line_start": 138,
"title": "Twelve gaming attempts the fixture must catch"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 151,
"source_line_start": 150,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The design follows the reward-hacking and benchmark-leakage literature: a metric stops being evidence once it becomes a target, so the references must carry the proof.** The component draws on the long-standing observation that an optimized metric loses its evidential force once it becomes a target, and on the AI-safety literature on reward hacking and specification gaming. Concrete Problems in AI Safety frames reward hacking as a practical accident risk, DeepMind's specification-gaming survey collects concrete examples, and Benchmarking Benchmark Leakage in Large Language Models motivates explicit leakage documentation. The engineering pattern is simple accounting: evaluator ids, configu\u2026",
"level": 4,
"order": 16,
"section_id": "goodhart-reward-hacking-and-leakage-accounting",
"source_line_end": 168,
"source_line_start": 152,
"title": "Goodhart, reward hacking, and leakage accounting"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 170,
"source_line_start": 169,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "The fixture validator runs from the repository root and writes its result to a disposable path outside the working tree. Run the fixture validator from the repository root and write its result record outside the working tree: Then run the exported-bundle validator: The focused test and corpus check are:",
"level": 4,
"order": 18,
"section_id": "fixture-output-stays-outside-the-working-tree",
"source_line_end": 193,
"source_line_start": 171,
"title": "Fixture output stays outside the working tree"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means every verdict was recomputed from the public references and agreed with the independent trace; it does not mean a benchmark claims was produced.** A pass means the fixture and bundle validators recomputed every verdict from public case ids, locked evaluator references, configuration hashes, evidence references, manifest digests, and the twelve negative cases, and agreed with the independent trace.",
"level": 4,
"order": 19,
"section_id": "what-a-pass-does-and-does-not-mean",
"source_line_end": 202,
"source_line_start": 194,
"title": "What a pass does and does not mean"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 204,
"source_line_start": 203,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**On this fixture and bundle, a metadata-only replay respected the anti-gaming boundary, and AUTHORITYCEILING stops the proof there. The strongest claim the evidence supports is narrow: on this public fixture and the exported bundle, a metadata-only replay respected an anti-gaming boundary over public case ids and locked evaluator references, and the two positive rows each trace back to a passing command run. The scope_limit constant sets every outward authorization flag to false, so the proof boundary stops there.",
"level": 4,
"order": 21,
"section_id": "the-strongest-supported-claim",
"source_line_end": 214,
"source_line_start": 205,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A green run still claims no benchmark claims, no repair capability, and no access to hidden or private material.** It claims no benchmark claims, no SWE-bench result, and no repair capability. It does not access hidden-gold answers, oracle patches, or private issue text, does not call a provider, does not mutate a live repository, and excludes public sharing or launch. Those remain false even when the fixture passes.",
"level": 4,
"order": 22,
"section_id": "what-stays-false-even-on-a-pass",
"source_line_end": 223,
"source_line_start": 215,
"title": "What stays false even on a pass"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/agent_benchmark_integrity_anti_gaming_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The agent benchmark integrity anti-gaming replay validates copied public source pattern provenance bodies and metadata-only benchmark replay rows before any score-like language is allowed."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "agent_benchmark_integrity_anti_gaming_replay.py",
"path": "src/microcosm_core/organs/agent_benchmark_integrity_anti_gaming_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for benchmark-integrity fixture validation, exported bundle validation, copied source-module digest checks, public trace verdict recomputation, negative-case enforcement, and scope limits.",
"symbols": [
"run",
"run_benchmark_integrity_bundle",
"validate_source_module_imports",
"validate_public_trace",
"_build_result",
"_freshness_basis",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/agent_benchmark_integrity_anti_gaming_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_benchmark_integrity_anti_gaming_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_benchmark_integrity_anti_gaming_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Agent Benchmark Integrity Anti-Gaming Replay is the public benchmark-claim boundary for Microcosm. It checks locked evaluator ids and config hashes, declared benchmark case rosters, replay rows, file-access and contamination refs, trusted-reference score refs, output-replay refs, computed-vs-declared integrity verdicts, three copied source pattern provenance bodies, and eleven anti-gaming negative cases while keeping private issue, oracle patch, hidden-gold, provider, raw patch, and score payload bodies out of result records.",
"summary_line": "The agent benchmark integrity anti-gaming replay validates copied public source pattern provenance bodies and metadata-only benchmark replay rows before any score-like language is allowed.",
"title": "Agent Benchmark Integrity Anti-Gaming Replay"
},
{
"anchor": "paper-module-research-replication-rubric-artifact-replay",
"authority_ceiling": "Copied public source pattern provenance bodies, exact-copy public Python internal control body, metadata-only research-replication replay result records, public agent-execution trace spans, and fixture validation only; no actual paper replication success, benchmark performance claim, private paper/data body export, hidden-rubric export, external model access, unbounded compute search, original-author code reuse, launch-scope decision, publishing-scope decision, source-file changes, or product-progress evidence.",
"body_html": "
Evidence bundle for a claimed paper replication
\n
It turns the claim that an agent replicated a paper into an evidence bundle a cold runtime can check, reruns nothing scientific, and forces the packet's numbers to reproduce.research_replication_rubric_artifact_replay is a public Microcosm component that turns the sentence "an agent replicated a paper" into a bundle a cold runtime can check. It reruns nothing scientific. It reads a replay packet and confirms the packet names every piece of evidence a real replication would have left behind, then runs one small local metric so the numbers in the packet have to reproduce rather than be taken on trust.
\n
It runs over two synthetic public papers, and a pass means the replay packet is accountable and digest-bound, not that a real paper was independently replicated. The component runs over two synthetic public papers, one machine-learning method and one computational-science study. A pass says the replay packet is structurally accountable, digest-bound, and negative-case tested. It does not say a real paper was independently replicated.
\n
Purpose
\n
Four things collapsed into one claim
\n
Demos fold the paper, the artifact, the rubric, and the rerun evidence into a single word, so this component pulls "replicated" apart and asks for each object by name. Research-agent demos collapse four separate things into one claim: the paper, the runnable artifact, the grading rubric, and the evidence that an independent rerun happened. A loose check accepts any of them as proof of all four. This component refuses to treat "replicated" as one fact. It pulls the claim apart into the objects a real replication produces and asks for each one by name.
\n
Fourteen fields, eight axes, four forbidden authorities
\n
validate_replication_policy sets the floor by requiring fourteen replay fields and eight rubric axes and forcing four escape authorities false.validate_replication_policy sets the floor. It requires the fourteen fields in REQUIRED_REPLAY_FIELDS: paper_id, contribution_decomposition_ref, rubric_tree_ref, allowed_public_input_refs, scratch_repo_scaffold_ref, experiment_dag_ref, metric_script_refs, artifact_hash_refs, declared_artifact_hash_refs, grader_report_ref, cost_runtime_budget_ref, ablation_diff_ref, failure_taxonomy_ref, and cold_rerun_receipt_ref. It also requires the eight rubric axes in REQUIRED_RUBRIC_AXES: contribution decomposition, artifact replay, experiment DAG, metric script, grader alignment, budget boundary, failure taxonomy, and cold rerun. A policy that drops any field or axis is rejected. The policy must also declare four escape authorities false: hidden-rubric export, benchmark-performance claim, unbounded compute search, and final-answer-only grading. A packet cannot grant itself any of them.
\n
How it works
\n
The two entry points are run and run_replication_bundle. Both call _build_result and differ only in what they read. The two entry points are run and run_replication_bundle. Both call _build_result and differ only in what they read.
\n
Fixture mode versus exported-bundle mode
\n
run loads the fixture with its eight negative cases, while run_replication_bundle validates the public export without them and reuses an unchanged prior record.run is fixture mode. It reads the first-wave input directory with include_negative=True, so it loads the positive replay rows plus the eight negative-case files, and writes the result, board, validation, and sign-off records. run_replication_bundle is exported-bundle mode. It reads the public example bundle with include_negative=False, then writes exported_research_replication_bundle_validation_result.json. It can reuse a prior record when _freshness_basis shows the inputs are byte-identical, so a clean rerun does not rewrite an unchanged result.
\n
_build_result runs the checks in order and folds them into one record._build_result runs the checks in order and folds them into one record.
\n
Function
What it checks
validate_source_module_imports
The source-module manifest declares copied public bodies with matching digests.
validate_replication_policy
The policy requires all fourteen fields and eight axes and forbids the escape authorities.
validate_research_replays
At least two positive rows carry every required field; the eight negative files each trip their expected error code.
validate_execution_artifact_replay
One declared metric runs over the allowed public input, and the output matches the declared hash.
validate_projection_protocol
The projection cites its source patterns, result records, target symbols, and public trace import.
\n
The metric replay that must reproduce
\n
_execute_metric_script sums one named field over an allowed public input, and validate_execution_artifact_replay fails the row if the produced numbers do not match the declared hash. The executable core is _execute_metric_script. It reads a metric spec whose operation is sum_field, sums one named field over the rows of one allowed public input table, and returns the total plus a row count. It rejects a spec that is not a bare sum, and it rejects input rows that are missing or non-numeric. validate_execution_artifact_replay wraps that call: it reads execution_artifacts/execution_artifact_manifest.json, refuses a manifest that sets body_in_receipt, authorizes external model access, or names a private body, runs the metric, then compares the produced payload against the declared output artifact and its declared sha256 file. A replay row can name every correct reference and still fail here if the numbers do not reproduce, so the pass is not a field-presence check.
\n
Verdicts from fields, not from fixture labels
\n
validate_research_replays reads verdicts from row fields rather than declared labels, and validate_source_module_imports verifies copied-body digests while result records stay metadata-only.validate_research_replays derives its verdicts from the replay-row fields, not from any label the fixture declares. It records the keys it ignored, so a file that claims status: pass or names its own expected_error_codes cannot talk its way past the check. validate_source_module_imports reads the manifest at SOURCE_MODULE_MANIFEST_REF, requires source_import_class of copied_non_secret_macro_body, requires body_in_receipt to be false, and verifies each declared target digest. For an exact-copy row it also confirms the live source file exists and that the source and target bytes match. Result records carry counts, references, digests, and verdicts. They never carry a source body.
\nDiagram of the mechanism (6 steps).Diagram source & refs
The fixture ships eight named negative files, one per key in EXPECTED_NEGATIVE_CASES. Each must trip its own error code: The fixture ships eight named negative files, one per key in EXPECTED_NEGATIVE_CASES. Each must trip its own error code:
\n
original-author code reuse
hidden-rubric leakage
report-only success
benchmark-performance overclaim
private paper or data body leakage
unbounded compute search
final-answer-only grading
undeclared artifact-hash reference
\n
Eight fakes plus mutation cases stay blocked
\n
These name the eight ways a plausible fake tries to pass, and the focused tests add mutation cases that stay blocked even when the tampered bundle is made internally consistent. These are the eight ways a plausible fake tries to pass. A packet that reuses the original authors' code, calls a run a success on the strength of a write-up alone, asserts a benchmark number, or points at an output hash it never declared is rejected by field semantics, not by filename. The focused test suite adds mutation cases on top: digest mismatch, local body tamper, rehashed source swaps, wrong execution-artifact hashes, metric perturbation, input perturbation, output body tamper, and self-consistent input/output/hash rewrites. Each mutation stays blocked even when the tampered bundle is made internally consistent.
\n
Prior Art Grounding
\n
Borrowed from ACM artifact badging
\n
The rubric-over-artifact shape follows ACM artifact badging, which separates available from functional from reproduced, made here into a runtime gate over a declared bundle. This is the rubric-over-artifact shape from systems and machine-learning artifact evaluation, where the ACM Artifact Review and Badging process separates "available" from "functional" from "reproduced". Microcosm borrows that separation and makes it a runtime gate over a declared evidence bundle. The result is fixture-bound replay evidence, not a reproducibility guarantee and not a peer-review verdict.
A pass certifies every check held, all eight negative cases were observed, and the records stayed metadata-only. A pass means the policy, replay rows, source-module digests, projection protocol, and local metric replay all held, all eight negative cases were observed, and the records stayed metadata-only.
\n
Scope boundary
\n
Scope limit
\n
The narrow strongest claim
\n
The strongest supported claim is narrow: over synthetic public inputs the packet named every evidence object, the eight overclaims were caught, and one metric reproduced against its hash. The strongest claim the evidence supports is narrow. Over public synthetic inputs, the component confirmed a replay packet named every required evidence object, the eight overclaim cases were each caught, and one declared metric reproduced against its declared output hash. That is the whole proof boundary: the bundle, mechanism row, runtime code, source-module manifest, exported record, first-wave fixture records, and focused tests.
\n
What a green run still does not claim
\n
The scope limit stops there: a green run claims no real replication, benchmark performance, external model service, or launch-scope decision, and the value is the boundary, not the science. The scope limit stops there. A green run does not claim actual paper replication, benchmark performance, external model service, hidden-rubric access, original-author code reuse, private paper or data export, unbounded compute, final-answer-only grading, launch-scope decision, or publishing-scope decision. The two subjects are synthetic public examples, and the metric is one sum over one small table. The value here is the boundary, not the science. Benchmark overclaim controls stay routed through the sibling module agent_benchmark_integrity_anti_gaming_replay rather than reinvented here.
\n",
"components": [
{
"anchor": "component-research_replication_rubric_artifact_replay",
"component_id": "research_replication_rubric_artifact_replay",
"public_label": "Research Replication Rubric Artifact Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 6
},
"link_count": 14,
"links": [
{
"id": "research_replication_rubric_artifact_replay",
"kind": "organ",
"kind_label": "component",
"label": "Research Replication Rubric Artifact Replay",
"relation": "explains",
"url": "component-research-replication-rubric-artifact-replay.html"
},
{
"id": "mechanism.research_replication_rubric_artifact_replay.validates_public_research_replication_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public research replication replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-research-replication-rubric-artifact-replay-validates-public-research-replication-replay"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
},
{
"id": "paper_module.agent_benchmark_integrity_anti_gaming_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Benchmark Integrity Anti-Gaming Replay",
"relation": "depends on",
"url": "paper-module-agent-benchmark-integrity-anti-gaming-replay.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch10_cold_eval_honesty_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 10 Cold Eval Honesty Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[67:paper_module.batch10_cold_eval_honesty_capsule].depends_on",
"url": "paper-module-batch10-cold-eval-honesty-capsule.html"
},
{
"id": "paper_module.materials_chemistry_closed_loop_lab_safety_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Materials Chemistry Closed-Loop Lab-Safety Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[6:paper_module.materials_chemistry_closed_loop_lab_safety_replay].depends_on",
"url": "paper-module-materials-chemistry-closed-loop-lab-safety-replay.html"
},
{
"id": "paper_module.mechanistic_interpretability_circuit_attribution_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mechanistic Interpretability Circuit Attribution Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[51:paper_module.mechanistic_interpretability_circuit_attribution_replay].depends_on",
"url": "paper-module-mechanistic-interpretability-circuit-attribution-replay.html"
},
{
"id": "paper_module.prediction_oracle_reconciliation",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Prediction Oracle Reconciliation",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[53:paper_module.prediction_oracle_reconciliation].depends_on",
"url": "paper-module-prediction-oracle-reconciliation.html"
},
{
"id": "paper_module.spatial_world_model_counterfactual_simulation_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Spatial World Model Counterfactual Simulation Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[52:paper_module.spatial_world_model_counterfactual_simulation_replay].depends_on",
"url": "paper-module-spatial-world-model-counterfactual-simulation-replay.html"
}
],
"paper_module_id": "paper_module.research_replication_rubric_artifact_replay",
"reader_packet": {
"abstract": "The research replication rubric-artifact replay validates source-backed public replication bundles before any paper-replication language is allowed.",
"authority_ceiling": "Copied public source pattern provenance bodies, exact-copy public Python internal control body, metadata-only research-replication replay result records, public agent-execution trace spans, and fixture validation only; no actual paper replication success, benchmark performance claim, private paper/data body export, hidden-rubric export, external model access, unbounded compute search, original-author code reuse, launch-scope decision, publishing-scope decision, source-file changes, or product-progress evidence.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.research_replication_rubric_artifact_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-research-replication-rubric-artifact-replay-diagram-1",
"html_url": "docs/paper-module-research-replication-rubric-artifact-replay.html#paper-module-research-replication-rubric-artifact-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD policy[\"Replication policy 14 fields + 8 axes + forbidden claims\"] replays[\"Replay rows 2 synthetic papers + 8 negative cases\"] artifacts[\"Execution manifest allowed input + metric spec + declared hash\"] metric[\"Local metric replay sum_field over allowed input\"] source[\"Source-module manifest 4 copied public bodies\"] result[\"metadata-only result record counts, refs, digests, scope boundaries\"] policy --> replays replays --> artifacts artifacts --> metric metric --> result source --> result",
"source_format": "mermaid",
"source_hash": "sha256:75741d3bc6989e6c18630a75aadfe0d2adbd5909f760e229e47fe2e656942f26",
"source_line_end": 138,
"source_line_start": 124
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.agent_benchmark_integrity_anti_gaming_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Benchmark Integrity Anti-Gaming Replay",
"relation": "depends on",
"url": "paper-module-agent-benchmark-integrity-anti-gaming-replay.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "research_replication_rubric_artifact_replay",
"kind": "organ",
"kind_label": "component",
"label": "Research Replication Rubric Artifact Replay",
"relation": "explains",
"url": "component-research-replication-rubric-artifact-replay.html"
},
{
"id": "mechanism.research_replication_rubric_artifact_replay.validates_public_research_replication_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public research replication replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-research-replication-rubric-artifact-replay-validates-public-research-replication-replay"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.research_replication_rubric_artifact_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 248,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/research_replication_rubric_artifact_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "ACM Artifact Review and Badging",
"section_id": "borrowed-from-acm-artifact-badging",
"url": "https://www.acm.org/publications/policies/artifact-review-and-badging-current"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It turns the claim that an agent replicated a paper into an evidence bundle a cold runtime can check, reruns nothing scientific, and forces the packet's numbers to reproduce. researchreplicationrubricartifact_replay is a public Microcosm component that turns the sentence \"an agent replicated a paper\" into a bundle a cold runtime can check. It reruns nothing scientific. It reads a replay packet and confirms the packet names every piece of evidence a real replication would have left behind, then runs one small local metric so the numbers in the packet have to reproduce rather than be taken on trust.",
"level": 4,
"order": 1,
"section_id": "evidence-bundle-for-a-claimed-paper-replication",
"source_line_end": 14,
"source_line_start": 3,
"title": "Evidence bundle for a claimed paper replication"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It runs over two synthetic public papers, and a pass means the replay packet is accountable and digest-bound, not that a real paper was independently replicated.** The component runs over two synthetic public papers, one machine-learning method and one computational-science study. A pass says the replay packet is structurally accountable, digest-bound, and negative-case tested. It does not say a real paper was independently replicated.",
"level": 4,
"order": 2,
"section_id": "it-runs-over-two-synthetic-public-papers",
"source_line_end": 24,
"source_line_start": 15,
"title": "It runs over two synthetic public papers"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 3,
"section_id": "purpose",
"source_line_end": 26,
"source_line_start": 25,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Demos fold the paper, the artifact, the rubric, and the rerun evidence into a single word, so this component pulls \"replicated\" apart and asks for each object by name.** Research-agent demos collapse four separate things into one claim: the paper, the runnable artifact, the grading rubric, and the evidence that an independent rerun happened. A loose check accepts any of them as proof of all four. This component refuses to treat \"replicated\" as one fact. It pulls the claim apart into the objects a real replication produces and asks for each one by name.",
"level": 4,
"order": 4,
"section_id": "four-things-collapsed-into-one-claim",
"source_line_end": 37,
"source_line_start": 27,
"title": "Four things collapsed into one claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatereplication_policy sets the floor by requiring fourteen replay fields and eight rubric axes and forcing four escape authorities false.** validatereplicationpolicy sets the floor. It requires the fourteen fields in REQUIREDREPLAYFIELDS: paperid, contributiondecomposition_ref, rubrictreeref, allowedpublicinputrefs, scratchreposcaffoldref, experimentdagref, metricscriptrefs, artifacthashrefs, declaredartifacthashrefs, graderreportref, costruntimebudgetref, ablationdiffref, failuretaxonomyref, and coldrerunreceipt_ref. It also requires the eight rubric axes in REQUIREDRUBRICAXES: contribution decomposition, artifact replay, experiment DAG, metric script, grader alignment, budget boundar\u2026",
"level": 4,
"order": 5,
"section_id": "fourteen-fields-eight-axes-four-forbidden-authorities",
"source_line_end": 55,
"source_line_start": 38,
"title": "Fourteen fields, eight axes, four forbidden authorities"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 6,
"section_id": "how-it-works",
"source_line_end": 57,
"source_line_start": 56,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The two entry points are run and runreplicationbundle. Both call buildresult and differ only in what they read. The two entry points are run and runreplicationbundle. Both call buildresult and differ only in what they read.",
"level": 4,
"order": 7,
"section_id": "the-two-entry-points-are-run-and-run-replication-bundle",
"source_line_end": 63,
"source_line_start": 58,
"title": "The two entry points are `run` and `run_replication_bundle`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**run loads the fixture with its eight negative cases, while runreplicationbundle validates the public export without them and reuses an unchanged prior record.** run is fixture mode. It reads the first-wave input directory with include_negative=True, so it loads the positive replay rows plus the eight negative-case files, and writes the result, board, validation, and sign-off records. runreplicationbundle is exported-bundle mode. It reads the public example bundle with include_negative=False, then writes exportedresearchreplicationbundlevalidation_result.json. It can reuse a prior record when freshnessbasis shows the inputs are byte-identical, so a clean rerun does not rewrite an unchanged\u2026",
"level": 4,
"order": 8,
"section_id": "fixture-mode-versus-exported-bundle-mode",
"source_line_end": 77,
"source_line_start": 64,
"title": "Fixture mode versus exported-bundle mode"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "buildresult runs the checks in order and folds them into one record.** buildresult runs the checks in order and folds them into one record.",
"level": 4,
"order": 9,
"section_id": "build-result-runs-the-checks-in-order-and-folds-them-into-one",
"source_line_end": 90,
"source_line_start": 78,
"title": "`_build_result` runs the checks in order and folds them into one"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "executemetric_script sums one named field over an allowed public input, and validateexecutionartifact_replay fails the row if the produced numbers do not match the declared hash. The executable core is executemetricscript. It reads a metric spec whose operation is sum_field, sums one named field over the rows of one allowed public input table, and returns the total plus a row count. It rejects a spec that is not a bare sum, and it rejects input rows that are missing or non-numeric. validateexecutionartifact_replay wraps that call: it reads executionartifacts/executionartifact_manifest.json, refuses a manifest that sets bodyinreceipt, authorizes external model access, or names a private body\u2026",
"level": 4,
"order": 10,
"section_id": "the-metric-replay-that-must-reproduce",
"source_line_end": 107,
"source_line_start": 91,
"title": "The metric replay that must reproduce"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "validateresearch_replays reads verdicts from row fields rather than declared labels, and validatesourcemodule_imports verifies copied-body digests while result records stay metadata-only. validateresearch_replays derives its verdicts from the replay-row fields, not from any label the fixture declares. It records the keys it ignored, so a file that claims status: pass or names its own expectederrorcodes cannot talk its way past the check. validatesourcemodule_imports reads the manifest at SOURCEMODULEMANIFESTREF, requires sourceimport_class of copiednonsecretmacrobody, requires bodyinreceipt to be false, and verifies each declared target digest. For an exact-copy row it also confirms the liv\u2026",
"level": 4,
"order": 11,
"section_id": "verdicts-from-fields-not-from-fixture-labels",
"source_line_end": 139,
"source_line_start": 108,
"title": "Verdicts from fields, not from fixture labels"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 141,
"source_line_start": 140,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The fixture ships eight named negative files, one per key in EXPECTEDNEGATIVECASES. Each must trip its own error code: The fixture ships eight named negative files, one per key in EXPECTEDNEGATIVECASES. Each must trip its own error code: original-author code reuse hidden-rubric leakage report-only success benchmark-performance overclaim private paper or data body leakage unbounded compute search final-answer-only grading undeclared artifact-hash reference",
"level": 4,
"order": 13,
"section_id": "the-fixture-ships-eight-named-negative-files-one-per-key-in-expected-negative-cases",
"source_line_end": 156,
"source_line_start": 142,
"title": "The fixture ships eight named negative files, one per key in `EXPECTED_NEGATIVE_CASES`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**These name the eight ways a plausible fake tries to pass, and the focused tests add mutation cases that stay blocked even when the tampered bundle is made internally consistent.** These are the eight ways a plausible fake tries to pass. A packet that reuses the original authors' code, calls a run a success on the strength of a write-up alone, asserts a benchmark number, or points at an output hash it never declared is rejected by field semantics, not by filename. The focused test suite adds mutation cases on top: digest mismatch, local body tamper, rehashed source swaps, wrong execution-artifact hashes, metric perturbation, input perturbation, output body tamper, and self-consistent input\u2026",
"level": 4,
"order": 14,
"section_id": "eight-fakes-plus-mutation-cases-stay-blocked",
"source_line_end": 171,
"source_line_start": 157,
"title": "Eight fakes plus mutation cases stay blocked"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 173,
"source_line_start": 172,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The rubric-over-artifact shape follows ACM artifact badging, which separates available from functional from reproduced, made here into a runtime gate over a declared bundle.** This is the rubric-over-artifact shape from systems and machine-learning artifact evaluation, where the ACM Artifact Review and Badging process separates \"available\" from \"functional\" from \"reproduced\". Microcosm borrows that separation and makes it a runtime gate over a declared evidence bundle. The result is fixture-bound replay evidence, not a reproducibility guarantee and not a peer-review verdict.",
"level": 4,
"order": 16,
"section_id": "borrowed-from-acm-artifact-badging",
"source_line_end": 185,
"source_line_start": 174,
"title": "Borrowed from ACM artifact badging"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Focused runtime validation: Paper-module corpus validation: The runtime commands behind the records are:",
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 212,
"source_line_start": 186,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass certifies every check held, all eight negative cases were observed, and the records stayed metadata-only.** A pass means the policy, replay rows, source-module digests, projection protocol, and local metric replay all held, all eight negative cases were observed, and the records stayed metadata-only.",
"level": 4,
"order": 18,
"section_id": "what-a-pass-over-the-records-means",
"source_line_end": 219,
"source_line_start": 213,
"title": "What a pass over the records means"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "claim-ceiling",
"source_line_end": 221,
"source_line_start": 220,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest supported claim is narrow: over synthetic public inputs the packet named every evidence object, the eight overclaims were caught, and one metric reproduced against its hash.** The strongest claim the evidence supports is narrow. Over public synthetic inputs, the component confirmed a replay packet named every required evidence object, the eight overclaim cases were each caught, and one declared metric reproduced against its declared output hash. That is the whole proof boundary: the bundle, mechanism row, runtime code, source-module manifest, exported record, first-wave fixture records, and focused tests.",
"level": 4,
"order": 20,
"section_id": "the-narrow-strongest-claim",
"source_line_end": 233,
"source_line_start": 222,
"title": "The narrow strongest claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The scope limit stops there: a green run claims no real replication, benchmark performance, external model service, or launch-scope decision, and the value is the boundary, not the science.** The scope limit stops there. A green run does not claim actual paper replication, benchmark performance, external model service, hidden-rubric access, original-author code reuse, private paper or data export, unbounded compute, final-answer-only grading, launch-scope decision, or public sharing decision. The two subjects are synthetic public examples, and the metric is one sum over one small table. The value here is the boundary, not the science. Benchmark overclaim controls stay routed through the s\u2026",
"level": 4,
"order": 21,
"section_id": "what-a-green-run-still-does-not-claim",
"source_line_end": 247,
"source_line_start": 234,
"title": "What a green run still does not claim"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/research_replication_rubric_artifact_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The research replication rubric-artifact replay validates source-backed public replication bundles before any paper-replication language is allowed."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "research_replication_rubric_artifact_replay.py",
"path": "src/microcosm_core/organs/research_replication_rubric_artifact_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for research-replication fixture validation, exported bundle validation, source-module manifest checks, public trace construction, negative-case enforcement, scope limits, and metadata-only result record writing.",
"symbols": [
"run",
"run_replication_bundle",
"validate_source_module_imports",
"validate_projection_protocol",
"validate_replication_policy",
"validate_research_replays",
"_build_result",
"_freshness_basis"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/research_replication_rubric_artifact_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/research_replication_rubric_artifact_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/research_replication_rubric_artifact_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Research Replication Rubric Artifact Replay is the public research-replication claim boundary for Microcosm. It checks contribution decomposition refs, rubric trees, allowed public inputs, scratch repo scaffolds, experiment DAG refs, metric scripts, declared artifact-hash rosters, artifact hashes, grader reports, compute/runtime budgets, ablation diffs, failure taxonomies, cold-rerun result records, public agent-execution trace spans, four copied source modules, and eight replication-overclaim negative cases while keeping private paper/data bodies, hidden rubrics, model-output data, original-author code bodies, benchmark claims, and public-sharing claims out of result records.",
"summary_line": "The research replication rubric-artifact replay validates source-backed public replication bundles before any paper-replication language is allowed.",
"title": "Research Replication Rubric Artifact Replay"
},
{
"anchor": "paper-module-agentic-vulnerability-discovery-patch-proof-replay",
"authority_ceiling": "Copied public source/control/standard/tool bodies, metadata-only synthetic patch-proof replay result records, public agent-execution trace spans, and fixture validation only; no live target testing, real CVE exploitation, weaponized payload export, account secret handling, network exfiltration, actionable exploit instructions, external model access, source-file changes, benchmark security score, launch-scope decision, publishing-scope decision, whole-system security claim, or product-progress evidence.",
"body_html": "
Patch-proof language needs more than a claim
\n
"Found and fixed a security bug" is easy to type and hard to check, standing equally for a tested repair, a patch never run, a false positive, or an empty benchmark number. An agent that says it "found and fixed a security bug" is making a claim that is easy to type and hard to check. The phrase can stand for a real, tested repair. It can also stand for a patch that was never run, a false positive promoted to a finding, or a benchmark number with nothing behind it.
\n
Seven evidence stages must agree
\n
It breaks the claim into a synthetic evidence chain, from target through issue, trace, patch, test, witness, and verifier record, and checks the pieces line up before allowing patch-proof language. This component takes that claim apart into an inspectable chain of synthetic evidence and checks that the pieces line up before any patch-proof language is allowed. The chain runs from a made-up target, through a suspected issue, a trace that backs it, an abstract exploitability reference, a patch, a regression test, an executable witness, and a verifier record. Nothing in the chain is a real target, exploit, payload, or account secret. The result record says whether the declared chain holds together, and carries no attack material.
\n
Synthetic replay grants no security authority
\n
It is a synthetic, metadata-only replay, not live security work, that runs against no real system, judges no bug or fix as genuine, and produces no benchmark claims. The code lives in src/microcosm_core/organs/agentic_vulnerability_discovery_patch_proof_replay.py. It is a synthetic, metadata-only replay, not live security work. It does not run against real systems, does not judge whether the bug or fix is genuine, and does not produce a benchmark claims.
\n
Purpose
\n
Why two recompute checks exist
\n
A summary can hide the absence of the work it summarises, so instead of trusting the fixture's labels two checks recompute the thing the fixture claims. The problem is that "found and fixed" is a summary, and a summary can hide the absence of the work it summarises. A loose check that only reads the fixture's own labels would pass a chain that asserts a green result without the work behind it. Two checks here refuse that by recomputing the thing the fixture claims, rather than trusting the field.
\n
Recompute one: running the regression witnesses
\n
The first recompute runs each witness's mini-target twice, unpatched and patched, and rejects any witness whose declared pre and post outputs do not match the computed ones. The first recompute is on the executable regression witnesses. Each witness names one of three small public mini-targets: a webhook redirect allowlist, a notebook log redactor, and a scheduler path normaliser. The validator runs that function twice, once unpatched and once patched, and compares its own computed outputs against the expected_pre_patch and expected_post_patch values the fixture declared. A witness whose declared output does not match the computed output is rejected.
\n
Recompute two: the verifier verdict
\n
The second recompute derives each verifier verdict from the joined hypothesis, proof, patch, test, and witness evidence, and requires the filename to match rather than trusting the record's own label. The second recompute is on the verifier records. Each record's pass or false_positive verdict is recomputed from the joined hypothesis, proof, patch, test, and witness evidence, and the record's filename must match that recomputed verdict. The record's own label is not taken on trust.
\n
How it works
\n
_build_result orchestrates the validators
\n
_build_result loads inputs and policy, scans every path for secrets, builds the public trace, then runs one validator per evidence kind, passing only when every check and negative case clears._build_result is the orchestrator. It loads the fixture inputs, loads the forbidden-class policy, runs a secret-exclusion scan over every input path, builds the public agent-execution trace, then runs one validator per evidence kind in dependency order. Each validator returns a status, exported rows, and a list of findings. A result reaches pass only when there are no missing negative cases, the secret scan has zero blocking hits, the public trace status is pass, every positive validator returned pass, and, for the exported bundle, the source-module manifest returned pass.
\n
Each validator joins against the preceding stage
\n
The validators form an ordered evidence join in which every stage resolves its rows against the stage immediately before it. The validators form a join, not a scanner. Each stage resolves its rows against the stage before it.
\n
Function
Role
run
Fixture entry point. Runs _build_result with negative cases on.
Loads inputs, runs every validator, computes the final status.
validate_issue_hypotheses
Each hypothesis joins to a synthetic target and cites evidence refs; needs at least four rows.
validate_trace_evidence
Each trace joins to a hypothesis; the set must cover static, dynamic, and property types.
validate_patch_diffs
Each patch-required hypothesis carries a metadata-only diff, updates tests, and claims no source-file changes.
validate_regression_tests
Each patch pairs with a test that fails before the patch and passes after it.
validate_executable_regression_witnesses
Runs the mini-target twice and compares computed output to the declared pre/post values.
validate_verifier_receipts
Recomputes each pass or false_positive verdict from joined evidence and checks the filename.
validate_sandbox_policy_verdicts
Each verdict precedes action and denies live-target and network material.
validate_cold_replay
Each replay row reproduces its trace, patch record, and sandbox verdict.
validate_negative_cases
Records the forbidden overclaims the fixture is expected to trip.
_source_module_manifest_result
Digest-checks the copied public source bodies carried by the exported bundle.
\n
Three mini-target engines run both ways
\n
Three MINI_TARGETS engines each take a patched flag: unpatched they leak or over-allow, patched they close the hole, and running both forms confirms a declared fails-before-passes-after test. The three mini-target engines are _webhook_redirect_allowlist, _notebook_log_redaction, and _scheduler_path_normalization, registered in MINI_TARGETS. Each takes an input and a patched flag. Unpatched, the webhook engine allows any redirect, the notebook engine leaks its sensitive refs, and the scheduler engine accepts a path that escapes the workspace. Patched, each closes its hole. Running both forms is what lets validate_executable_regression_witnesses confirm that a declared "fails before, passes after" test actually does.
\n
Strict join order blocks broken chains
\n
The join is strict: an unresolved hypothesis, a patch missing its proof or diff, a patch with no test, or a promoted false positive is each blocked. The join order is strict. A hypothesis that does not resolve to a synthetic target is blocked. A patch-required hypothesis missing either an exploitability proof or a metadata-only patch diff is blocked. A patch with no paired regression test is blocked. A false positive promoted to a finding is blocked, because validate_verifier_receipts will not recompute its verdict as pass.
\nDiagram of the mechanism (9 steps).Diagram source & refs
run drives the fixture with negative cases on, run_patch_proof_bundle drives the exported bundle as the public example, and both write metadata-only records with source bodies kept behind manifest hashes. The runtime keeps two modes separate. run drives the first-wave fixture with negative cases on, so it must observe every expected overclaim failure. run_patch_proof_bundle drives the exported bundle as the public runtime example, so its expected negative cases can be empty while it still proves the source-body floor, the public trace, the digest checks, and the regression witnesses. Both write metadata-only result records. Copied source bodies stay behind source_module_manifest.json refs and hashes; the exported bundle carries nine digest-verified public source bodies.
\n
Negative cases
\n
Eight overclaims the runtime must trip
\n
Eight EXPECTED_NEGATIVE_CASES overclaims must each trip a matching error code or the run is blocked, and _has_forbidden_key separately rejects any row carrying a forbidden field. The first-wave fixture includes eight negative cases, held in EXPECTED_NEGATIVE_CASES. Each one is an overclaim the runtime must trip, not a capability it offers: live_target_material, real_cve_exploitation, weaponized_payload_export, account secret material, network_exfiltration, exploit_instruction_steps, patch_without_tests, and benchmark_score_claim. validate_negative_cases records the matching error code for each, for example AGENTIC_VULN_PATCH_WITHOUT_TESTS when a patch claim arrives with no regression test. If any expected case goes unobserved, the run is blocked. Separately, _has_forbidden_key rejects any row that carries a forbidden field such as a raw exploit payload, a account secret value, or a live target URL.
\n
Prior Art Grounding
\n
Test-backed repair lineage from SWE-bench
\n
The design borrows the accountability pattern of SWE-bench, SWE-agent, and security benchmarks like CyberSecEval 2 and SecCodePLT, requiring evidence to line up before patch-proof language, without importing live exploits. This component borrows an accountability pattern from agentic software-engineering and security-evaluation work that treats a code repair as an executable, test-backed claim rather than a prose claim. SWE-bench framed repository issue resolution as a task with real codebases and test-based patch evaluation, and SWE-agent made the agent-computer interface part of the repair loop. Security benchmarks such as CyberSecEval 2 and SecCodePLT motivate separating a secure-code or vulnerability claim from an uninspected generated patch. The borrowed idea is narrow: hypotheses, traces, patches, tests, and verifier records must line up before patch-proof language is allowed. It does not import live targets, CVE exploitation, weaponized payloads, or benchmark claims.
\n
Validation Result record Path
\n
Fixture output stays outside the system tree
\n
The first-wave validator runs from the system root and writes its result record to a disposable path outside the working tree. Run the first-wave fixture validator from the system root and write its record outside the working tree:
A pass means the whole join held with both recompute checks agreeing, a clean scan, and every negative case tripped, while rejection tests mutate the chain and require a blocked result. A pass means the whole join held: the synthetic chain resolved end to end, both recompute checks agreed with the fixture, the secret scan was clean, and every expected negative case tripped. The rejection tests in the same file mutate the chain, remove a regression test, tamper a witness, cross-wire a verifier record, or alter a source-module digest, and require a blocked result with a specific error code.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
On the public fixtures the synthetic chain resolves and both recompute checks pass, with witnesses, verifier verdicts, and paired tests all agreeing and every forbidden overclaim observed. The strongest claim this component supports is narrow. On the public fixtures, the synthetic evidence chain resolves and both recompute checks pass: witnesses match their computed pre/post outputs, verifier verdicts match their recomputed result, patches pair with real tests, and every forbidden overclaim is observed. That is the scope limit, and it is where the scope limit sits.
\n
What the component refuses beyond that
\n
It refuses everything past that: no live targets, real CVE exploitation, weaponized payloads, account secrets, external model access, source-file changes, benchmark claims, or whole-system security claim, only the local synthetic records. It refuses everything past that. It does not test live targets, perform real CVE exploitation, export weaponized payloads, handle account secrets, authorise network exfiltration, or emit actionable exploit instructions. It does not make external model access, change source files, claim a benchmark security score, or judge patch correctness on a real repository. It is not launch or publishing-scope decision, and it is not a whole-system security claim. The proof boundary is the local result records, the synthetic bundle, the digest-checked source-module manifest, and the named negative cases, nothing wider.
\n",
"components": [
{
"anchor": "component-agentic_vulnerability_discovery_patch_proof_replay",
"component_id": "agentic_vulnerability_discovery_patch_proof_replay",
"public_label": "Vulnerability Patch-Proof Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 2
},
"link_count": 7,
"links": [
{
"id": "agentic_vulnerability_discovery_patch_proof_replay",
"kind": "organ",
"kind_label": "component",
"label": "Vulnerability Patch-Proof Replay",
"relation": "explains",
"url": "component-agentic-vulnerability-discovery-patch-proof-replay.html"
},
{
"id": "mechanism.agentic_vulnerability_discovery_patch_proof_replay.validates_public_agentic_vulnerability_patch_proof_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public agentic vulnerability patch proof replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agentic-vulnerability-discovery-patch-proof-replay-validates-public-agentic-vulnerability-patch-proof-replay"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.agentic_vulnerability_discovery_patch_proof_replay",
"reader_packet": {
"abstract": "The agentic vulnerability discovery patch-proof replay validates metadata-only synthetic vulnerability evidence chains before any found-and-fixed security language is allowed.",
"authority_ceiling": "Copied public source/control/standard/tool bodies, metadata-only synthetic patch-proof replay result records, public agent-execution trace spans, and fixture validation only; no live target testing, real CVE exploitation, weaponized payload export, account secret handling, network exfiltration, actionable exploit instructions, external model access, source-file changes, benchmark security score, launch-scope decision, publishing-scope decision, whole-system security claim, or product-progress evidence.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.agentic_vulnerability_discovery_patch_proof_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-agentic-vulnerability-discovery-patch-proof-replay-diagram-1",
"html_url": "docs/paper-module-agentic-vulnerability-discovery-patch-proof-replay.html#paper-module-agentic-vulnerability-discovery-patch-proof-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD targets[\"synthetic targets\"] hypotheses[\"issue hypotheses\"] traces[\"trace evidence\"] proofs[\"exploitability proofs\"] patches[\"patch diffs\"] tests[\"regression tests\"] witnesses[\"executable witnesses run mini-target twice\"] verifiers[\"verifier records recompute verdict\"] status[\"result status\"] targets --> hypotheses hypotheses --> traces hypotheses --> proofs hypotheses --> patches patches --> tests tests --> witnesses witnesses --> verifiers verifiers --> status",
"source_format": "mermaid",
"source_hash": "sha256:712ca9d5262ced5c7b06b4a8e28ac8ee96ac447ebeb4a82730eb62944b5b6b9c",
"source_line_end": 150,
"source_line_start": 130
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "agentic_vulnerability_discovery_patch_proof_replay",
"kind": "organ",
"kind_label": "component",
"label": "Vulnerability Patch-Proof Replay",
"relation": "explains",
"url": "component-agentic-vulnerability-discovery-patch-proof-replay.html"
},
{
"id": "mechanism.agentic_vulnerability_discovery_patch_proof_replay.validates_public_agentic_vulnerability_patch_proof_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public agentic vulnerability patch proof replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agentic-vulnerability-discovery-patch-proof-replay-validates-public-agentic-vulnerability-patch-proof-replay"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.agentic_vulnerability_discovery_patch_proof_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 277,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agentic_vulnerability_discovery_patch_proof_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "SWE-bench",
"section_id": "test-backed-repair-lineage-from-swe-bench",
"url": "https://arxiv.org/abs/2310.06770"
},
{
"kind": "external_reference",
"label": "SWE-agent",
"section_id": "test-backed-repair-lineage-from-swe-bench",
"url": "https://arxiv.org/abs/2405.15793"
},
{
"kind": "external_reference",
"label": "CyberSecEval 2",
"section_id": "test-backed-repair-lineage-from-swe-bench",
"url": "https://arxiv.org/abs/2404.13161"
},
{
"kind": "external_reference",
"label": "SecCodePLT",
"section_id": "test-backed-repair-lineage-from-swe-bench",
"url": "https://arxiv.org/abs/2410.11096"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**\"Found and fixed a security bug\" is easy to type and hard to check, standing equally for a tested repair, a patch never run, a false positive, or an empty benchmark number.** An agent that says it \"found and fixed a security bug\" is making a claim that is easy to type and hard to check. The phrase can stand for a real, tested repair. It can also stand for a patch that was never run, a false positive promoted to a finding, or a benchmark number with nothing behind it.",
"level": 4,
"order": 1,
"section_id": "patch-proof-language-needs-more-than-a-claim",
"source_line_end": 11,
"source_line_start": 3,
"title": "Patch-proof language needs more than a claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It breaks the claim into a synthetic evidence chain, from target through issue, trace, patch, test, witness, and verifier record, and checks the pieces line up before allowing patch-proof language.** This component takes that claim apart into an inspectable chain of synthetic evidence and checks that the pieces line up before any patch-proof language is allowed. The chain runs from a made-up target, through a suspected issue, a trace that backs it, an abstract exploitability reference, a patch, a regression test, an executable witness, and a verifier record. Nothing in the chain is a real target, exploit, payload, or account secret. The result record says whether the declared chain holds\u2026",
"level": 4,
"order": 2,
"section_id": "seven-evidence-stages-must-agree",
"source_line_end": 24,
"source_line_start": 12,
"title": "Seven evidence stages must agree"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is a synthetic, metadata-only replay, not live security work, that runs against no real system, judges no bug or fix as genuine, and produces no benchmark claims.** The code lives in src/microcosmcore/components/agenticvulnerabilitydiscoverypatchproofreplay.py. It is a synthetic, metadata-only replay, not live security work. It does not run against real systems, does not judge whether the bug or fix is genuine, and does not produce a benchmark claims.",
"level": 4,
"order": 3,
"section_id": "synthetic-replay-grants-no-security-authority",
"source_line_end": 34,
"source_line_start": 25,
"title": "Synthetic replay grants no security authority"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 36,
"source_line_start": 35,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A summary can hide the absence of the work it summarises, so instead of trusting the fixture's labels two checks recompute the thing the fixture claims.** The problem is that \"found and fixed\" is a summary, and a summary can hide the absence of the work it summarises. A loose check that only reads the fixture's own labels would pass a chain that asserts a green result without the work behind it. Two checks here refuse that by recomputing the thing the fixture claims, rather than trusting the field.",
"level": 4,
"order": 5,
"section_id": "why-two-recompute-checks-exist",
"source_line_end": 46,
"source_line_start": 37,
"title": "Why two recompute checks exist"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The first recompute runs each witness's mini-target twice, unpatched and patched, and rejects any witness whose declared pre and post outputs do not match the computed ones.** The first recompute is on the executable regression witnesses. Each witness names one of three small public mini-targets: a webhook redirect allowlist, a notebook log redactor, and a scheduler path normaliser. The validator runs that function twice, once unpatched and once patched, and compares its own computed outputs against the expectedprepatch and expectedpostpatch values the fixture declared. A witness whose declared output does not match the computed output is rejected.",
"level": 4,
"order": 6,
"section_id": "recompute-one-running-the-regression-witnesses",
"source_line_end": 58,
"source_line_start": 47,
"title": "Recompute one: running the regression witnesses"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The second recompute derives each verifier verdict from the joined hypothesis, proof, patch, test, and witness evidence, and requires the filename to match rather than trusting the record's own label.** The second recompute is on the verifier records. Each record's pass or false_positive verdict is recomputed from the joined hypothesis, proof, patch, test, and witness evidence, and the record's filename must match that recomputed verdict. The record's own label is not taken on trust.",
"level": 4,
"order": 7,
"section_id": "recompute-two-the-verifier-verdict",
"source_line_end": 68,
"source_line_start": 59,
"title": "Recompute two: the verifier verdict"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 70,
"source_line_start": 69,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult loads inputs and policy, scans every path for secrets, builds the public trace, then runs one validator per evidence kind, passing only when every check and negative case clears. build_result is the orchestrator. It loads the fixture inputs, loads the forbidden-class policy, runs a secret-exclusion scan over every input path, builds the public agent-execution trace, then runs one validator per evidence kind in dependency order. Each validator returns a status, exported rows, and a list of findings. A result reaches pass only when there are no missing negative cases, the secret scan has zero blocking hits, the public trace status is pass, every positive validator returned pass, a\u2026",
"level": 4,
"order": 9,
"section_id": "build-result-orchestrates-the-validators",
"source_line_end": 84,
"source_line_start": 71,
"title": "`_build_result` orchestrates the validators"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "The validators form an ordered evidence join in which every stage resolves its rows against the stage immediately before it. The validators form a join, not a scanner. Each stage resolves its rows against the stage before it.",
"level": 4,
"order": 10,
"section_id": "each-validator-joins-against-the-preceding-stage",
"source_line_end": 106,
"source_line_start": 85,
"title": "Each validator joins against the preceding stage"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Three MINITARGETS engines each take a patched flag: unpatched they leak or over-allow, patched they close the hole, and running both forms confirms a declared fails-before-passes-after test.** The three mini-target engines are webhookredirectallowlist, notebooklogredaction, and schedulerpathnormalization, registered in MINITARGETS. Each takes an input and a patched flag. Unpatched, the webhook engine allows any redirect, the notebook engine leaks its sensitive refs, and the scheduler engine accepts a path that escapes the workspace. Patched, each closes its hole. Running both forms is what lets validateexecutableregression_witnesses confirm that a declared \"fails before, passes after\" test\u2026",
"level": 4,
"order": 11,
"section_id": "three-mini-target-engines-run-both-ways",
"source_line_end": 119,
"source_line_start": 107,
"title": "Three mini-target engines run both ways"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "**The join is strict: an unresolved hypothesis, a patch missing its proof or diff, a patch with no test, or a promoted false positive is each blocked.** The join order is strict. A hypothesis that does not resolve to a synthetic target is blocked. A patch-required hypothesis missing either an exploitability proof or a metadata-only patch diff is blocked. A patch with no paired regression test is blocked. A false positive promoted to a finding is blocked, because validateverifierreceipts will not recompute its verdict as pass.",
"level": 4,
"order": 12,
"section_id": "strict-join-order-blocks-broken-chains",
"source_line_end": 151,
"source_line_start": 120,
"title": "Strict join order blocks broken chains"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run drives the fixture with negative cases on, runpatchproofbundle drives the exported bundle as the public example, and both write metadata-only records with source bodies kept behind manifest hashes.** The runtime keeps two modes separate. run drives the first-wave fixture with negative cases on, so it must observe every expected overclaim failure. runpatchproof_bundle drives the exported bundle as the public runtime example, so its expected negative cases can be empty while it still proves the source-body floor, the public trace, the digest checks, and the regression witnesses. Both write metadata-only result records. Copied source bodies stay behind sourcemodulemanifest.json refs and ha\u2026",
"level": 4,
"order": 13,
"section_id": "two-runtime-modes-both-body-free",
"source_line_end": 165,
"source_line_start": 152,
"title": "Two runtime modes, both metadata-only"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 167,
"source_line_start": 166,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Eight EXPECTEDNEGATIVE_CASES overclaims must each trip a matching error code or the run is blocked, and hasforbidden_key separately rejects any row carrying a forbidden field.** The first-wave fixture includes eight negative cases, held in EXPECTEDNEGATIVECASES. Each one is an overclaim the runtime must trip, not a capability it offers: livetargetmaterial, realcveexploitation, weaponizedpayloadexport, credentialmaterial, networkexfiltration, exploitinstructionsteps, patchwithouttests, and benchmarkscoreclaim. validatenegativecases records the matching error code for each, for example AGENTICVULNPATCHWITHOUTTESTS when a patch claim arrives with no regression test. If any expected case goes u\u2026",
"level": 4,
"order": 15,
"section_id": "eight-overclaims-the-runtime-must-trip",
"source_line_end": 183,
"source_line_start": 168,
"title": "Eight overclaims the runtime must trip"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 185,
"source_line_start": 184,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The design borrows the accountability pattern of SWE-bench, SWE-agent, and security benchmarks like CyberSecEval 2 and SecCodePLT, requiring evidence to line up before patch-proof language, without importing live exploits.** This component borrows an accountability pattern from agentic software-engineering and security-evaluation work that treats a code repair as an executable, test-backed claim rather than a prose claim. SWE-bench framed repository issue resolution as a task with real codebases and test-based patch evaluation, and SWE-agent made the agent-computer interface part of the repair loop. Security benchmarks such as CyberSecEval 2 and SecCodePLT motivate separating a secure-cod\u2026",
"level": 4,
"order": 17,
"section_id": "test-backed-repair-lineage-from-swe-bench",
"source_line_end": 204,
"source_line_start": 186,
"title": "Test-backed repair lineage from SWE-bench"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 206,
"source_line_start": 205,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "The first-wave validator runs from the system root and writes its result record to a disposable path outside the working tree. Run the first-wave fixture validator from the system root and write its record outside the working tree: Then run the exported bundle validator: The focused regression test is:",
"level": 4,
"order": 19,
"section_id": "fixture-output-stays-outside-the-substrate-tree",
"source_line_end": 238,
"source_line_start": 207,
"title": "Fixture output stays outside the system tree"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the whole join held with both recompute checks agreeing, a clean scan, and every negative case tripped, while rejection tests mutate the chain and require a blocked result.** A pass means the whole join held: the synthetic chain resolved end to end, both recompute checks agreed with the fixture, the secret scan was clean, and every expected negative case tripped. The rejection tests in the same file mutate the chain, remove a regression test, tamper a witness, cross-wire a verifier record, or alter a source-module digest, and require a blocked result with a specific error code.",
"level": 4,
"order": 20,
"section_id": "what-a-pass-and-the-rejection-tests-prove",
"source_line_end": 249,
"source_line_start": 239,
"title": "What a pass and the rejection tests prove"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 251,
"source_line_start": 250,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**On the public fixtures the synthetic chain resolves and both recompute checks pass, with witnesses, verifier verdicts, and paired tests all agreeing and every forbidden overclaim observed.** The strongest claim this component supports is narrow. On the public fixtures, the synthetic evidence chain resolves and both recompute checks pass: witnesses match their computed pre/post outputs, verifier verdicts match their recomputed result, patches pair with real tests, and every forbidden overclaim is observed. That is the scope limit, and it is where the scope limit sits.",
"level": 4,
"order": 22,
"section_id": "the-strongest-supported-claim",
"source_line_end": 262,
"source_line_start": 252,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It refuses everything past that: no live targets, real CVE exploitation, weaponized payloads, account secrets, external model access, source-file changes, benchmark claims, or whole-system security claim, only the local synthetic records.** It refuses everything past that. It does not test live targets, perform real CVE exploitation, export weaponized payloads, handle account secrets, authorise network exfiltration, or emit actionable exploit instructions. It does not make external model access, change source files, claim a benchmark security score, or judge patch correctness on a real repository. It is not launch or public sharing decision, and it is not a whole-system security claim. Th\u2026",
"level": 4,
"order": 23,
"section_id": "what-the-component-refuses-beyond-that",
"source_line_end": 276,
"source_line_start": 263,
"title": "What the component refuses beyond that"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/agentic_vulnerability_discovery_patch_proof_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The agentic vulnerability discovery patch-proof replay validates metadata-only synthetic vulnerability evidence chains before any found-and-fixed security language is allowed."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "agentic_vulnerability_discovery_patch_proof_replay.py",
"path": "src/microcosm_core/organs/agentic_vulnerability_discovery_patch_proof_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for agentic vulnerability patch-proof fixture validation, exported bundle validation, source-module manifest checks, public trace construction, secret-exclusion scans, negative-case enforcement, scope limits, and metadata-only result record writing.",
"symbols": [
"run",
"run_patch_proof_bundle",
"_source_module_manifest_result",
"_source_open_body_import_summary",
"_build_result",
"_freshness_basis",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/agentic_vulnerability_discovery_patch_proof_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agentic_vulnerability_discovery_patch_proof_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agentic_vulnerability_discovery_patch_proof_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Agentic Vulnerability Discovery Patch-Proof Replay is the public security-claim boundary for Microcosm. It checks projection protocol, vulnerability policy, synthetic target refs, issue hypotheses, trace evidence, abstract exploitability refs, patch diff refs, regression tests, verifier result records, sandbox verdicts, false-positive triage, cold replay, public agent-execution trace spans, secret-exclusion scan, nine copied source/control/standard/tool bodies, source-module manifest digests, metadata-only result record policy, and eight security-overclaim negative cases while keeping live targets, real CVE exploitation, weaponized payloads, account secrets, network exfiltration steps, actionable exploit instructions, model-output data, raw issue or patch bodies, benchmark claims, and source-file changes out of result records.",
"summary_line": "The agentic vulnerability discovery patch-proof replay validates metadata-only synthetic vulnerability evidence chains before any found-and-fixed security language is allowed.",
"title": "Agentic Vulnerability Discovery Patch-Proof Replay"
},
{
"anchor": "paper-module-materials-chemistry-closed-loop-lab-safety-replay",
"authority_ceiling": "Copied public Lab/Evolve source/control/result record/standard bodies, metadata-only simulator-only fixture result records, runtime bundle result records, and artifact safety/refusal validation only; no wetlab execution, hazardous synthesis guidance, reagent quantity, controlled or bioactive target, live assay, robot command, private lab notebook, external model access, discovery claim, benchmark claims, launch-scope decision, publishing-scope decision, or product-progress evidence.",
"body_html": "
"Closed-loop materials lab" is easy language to overclaim. A fixture can look like an autonomous discovery loop while carrying nothing that should be spoken aloud: wetlab steps, reagent quantities, a controlled or bioactive target, robot commands, or a bare assertion that some material was found. "Closed-loop materials lab" is easy language to overclaim. A fixture can look like an autonomous discovery loop while carrying nothing that should be spoken aloud: wetlab steps, reagent quantities, a controlled or bioactive target, robot commands, or a bare assertion that some material was found. This component sits in front of that language and answers one question. Is a closed-loop-lab-shaped fixture safe and grounded enough to be talked about at all, in a simulator-only frame, before any lab claim is allowed?
\n
Its stable id is materials_chemistry_closed_loop_lab_safety_replay, and its public surface name (PUBLIC_SURFACE_NAME) is materials_chemistry_artifact_safety_refusal_validator. Its stable id is materials_chemistry_closed_loop_lab_safety_replay, and its public surface name (PUBLIC_SURFACE_NAME) is materials_chemistry_artifact_safety_refusal_validator. The promise-sounding "closed-loop replay" name was reframed on purpose: nothing here executes a wetlab loop or commands a robot. The component reads public simulator rows, recomputes a winner, checks refusals, and writes a metadata-only result record. It emits a verdict of either public_safe_simulator_replay_accepted or blocked_public_safety_boundary, plus the findings that produced it.
\n
The part worth naming is that the component does not trust the fixture's own conclusion. A plain replay would read a declared "selected candidate" label and report it. The part worth naming is that the component does not trust the fixture's own conclusion. A plain replay would read a declared "selected candidate" label and report it. This one recomputes the winner from public numbers and treats a mismatch with the declared label as a failure, not a footnote.
\n
Purpose
\n
A loose reader of a lab fixture accepts a declared winner and a "safe" flag at face value. A loose reader of a lab fixture accepts a declared winner and a "safe" flag at face value. That fails two ways. A stale or flattering label passes even when the numbers no longer support it, and dangerous content slips through because nothing is watching for it. This component closes both gaps. It rederives the selection from score-backed rows, and it treats eight named categories of dangerous or overclaiming content as things to refuse rather than ignore. The result is fixture-bound evidence that the loop's structure is inspectable on a simulator, with no real lab, chemicals, or robot involved.
\n
How it works
\n
There are two entrypoints. run validates a first-wave fixture directory with negative cases included. There are two entrypoints. run validates a first-wave fixture directory with negative cases included. run_lab_bundle validates an exported bundle with include_negative=False and requires a copied source-module manifest. Both call _build_result to compute the payload, attach _freshness_basis, and write the result record. run_lab_bundle can reuse a still-fresh prior record through _fresh_lab_bundle_receipt when the input digests are unchanged.
\n
_build_result runs in this order.
\n
Function
Role
_load_payloads
Reads the six positive inputs (protocol, policy, candidates, experiment DAG, assays, decisions) and, in fixture mode, the eight forbidden negative inputs
Check required fields, cross-references, and safety flags on each row; each missing field or bad link becomes one finding
_numeric_replay_result
Recomputes the selected candidate from public scores; the core check
_source_module_manifest_result
Verifies copied source-module bodies by digest without putting body text into the record; required only in bundle mode
scan_paths, _secret_exclusion_scan
scan_paths sweeps the input set against the forbidden-class list and reports blocking hits; _secret_exclusion_scan drops body text from that scan record so it stays metadata-only
build_materials_lab_evolve_replay
Builds the Lab/Evolve replay graph from the same payloads and findings
\n
The numeric recompute is the heart. _numeric_replay_result reads public_safety_gate_score from each candidate, public_assay_proxy_value from each assay, and public_active_learning_score from each decision. The numeric recompute is the heart. _numeric_replay_result reads public_safety_gate_score from each candidate, public_assay_proxy_value from each assay, and public_active_learning_score from each decision. _numeric_value accepts a number only when it is finite and in [0, 1]; anything else yields MATERIALS_NUMERIC_REPLAY_SCORE_OUT_OF_RANGE or MATERIALS_NUMERIC_REPLAY_SCORE_REQUIRED. When no score field is present at all, the run blocks with MATERIALS_NUMERIC_REPLAY_POLICY_REQUIRES_SCORE_BACKED_ROWS (policy active) or MATERIALS_NUMERIC_REPLAY_REQUIRED. Scores are joined by candidate id. A candidate missing one of its three scores gives MATERIALS_NUMERIC_REPLAY_LINKAGE_REQUIRED. A safety gate below NUMERIC_REPLAY_MIN_SAFETY_GATE (0.70) gives MATERIALS_NUMERIC_REPLAY_SAFETY_GATE_FAILED.
\n
The weighted candidate score
\n
For each linked candidate the weighted score is 0.45 * assay_proxy + 0.35 * active_learning + 0.20 * safety_gate. For each linked candidate the weighted score is 0.45 * assay_proxy + 0.35 * active_learning + 0.20 * safety_gate. The winner is the maximum by score, ties broken by id. Only after that pick is computed does the component compare it to the policy's declared expected label. A mismatch is MATERIALS_NUMERIC_REPLAY_EXPECTED_LABEL_STALE. Numeric replay reports pass only when there are no findings and at least one verified row.
\n
The baseline fixture holds four candidates.
\n
Candidate
Safety gate
Assay proxy
Active-learning
Weighted score
Decision / action
mat_polymer_membrane_001
0.94
0.92
0.90
0.917
decision_membrane_001 / simulate_assay
mat_solid_electrolyte_002
0.91
0.84
0.81
0.8435
decision_electrolyte_002 / update_surrogate_model
mat_catalyst_support_003
0.85
0.78
0.74
0.780
decision_support_003 / choose_next_simulation
mat_sorbent_surface_004
0.88
0.70
0.66
0.722
decision_sorbent_004 / screen_candidate
\n
The recompute selects mat_polymer_membrane_001 at 0.917, which matches the policy's expected_selected_candidate_material_id. The recompute selects mat_polymer_membrane_001 at 0.917, which matches the policy's expected_selected_candidate_material_id. The policy declares minimum_safety_gate_score 0.70 and the selection rule max_weighted_public_assay_active_learning_and_safety_gate_score.
\n
_build_result folds every check into onepolicy_passed gate: the four row sets are non-empty, no positive finding fired, no forbidden content leaked into the encoded rows, no expected negative case is missing, every row is simulator_only, every hazard flag is false, and numeric replay passed._build_result folds every check into one policy_passed gate: the four row sets are non-empty, no positive finding fired, no forbidden content leaked into the encoded rows, no expected negative case is missing, every row is simulator_only, every hazard flag is false, and numeric replay passed. The overall status is a pass only when policy_passed holds, the secret scan passed, the replay graph passed, and, in bundle mode, the source-module manifest passed. The verdict string follows that status.
\nDiagram of the mechanism (11 steps).Diagram source & refs
Source refs
Accepted
public_safe_simulator_replay_accepted
Blocked
blocked_public_safety_boundary
flowchart TD inputs["six public inputs protocol, policy, candidates, experiments, assays, decisions"] rows["per-row field + link checks"] numeric["numeric recompute 0.45 assay + 0.35 active + 0.20 gate"] labelcheck["recomputed pick vs declared label gate >= 0.70"] negatives["eight forbidden fixtures"] refuse["any MATERIALS_*_FORBIDDEN observed"] manifest["source-module manifest digest check (bundle mode)"] scan["secret exclusion scan"] status["status + safety verdict"] accepted["public_safe_simulator_replay_accepted"] blocked["blocked_public_safety_boundary"] inputs --> rows inputs --> numeric numeric --> labelcheck labelcheck --> blocked labelcheck --> status negatives --> refuse refuse --> blocked refuse --> status manifest --> status scan --> status rows --> status status --> accepted status --> blocked
\n
Negative cases
\n
EXPECTED_NEGATIVE_CASES names eight forbidden fixtures, each with the code the component must raise when it observes one.EXPECTED_NEGATIVE_CASES names eight forbidden fixtures, each with the code the component must raise when it observes one. A run that fails to refuse any of them lists the case under expected_missing and blocks.
The numeric perturbations are covered too. Lowering the membrane gate to 0.52 moves the pick to mat_solid_electrolyte_002 and blocks with a safety-gate failure. The numeric perturbations are covered too. Lowering the membrane gate to 0.52 moves the pick to mat_solid_electrolyte_002 and blocks with a safety-gate failure. Declaring a stale expected label while the numbers select a different candidate raises the stale-label code. Removing score rows while the policy is active raises the score-backed-rows code with verified_numeric_row_count at 0.
\n
Prior Art Grounding
\n
Self-driving-lab structure under fixture evidence
\n
The shape borrowed here is the self-driving laboratory, where a propose-run-measure loop is paired with safety interlocks that can refuse an unsafe experiment. The shape borrowed here is the self-driving laboratory, where a propose-run-measure loop is paired with safety interlocks that can refuse an unsafe experiment. This component keeps the loop-plus-safety-gate structure but runs it over a simulator fixture and adds a numeric recompute so the declared winner is checked rather than trusted. The engineering lineage is ordinary refusal-first input validation applied to a lab-shaped record. It is not a real laboratory controller, a chemical-safety authority, or a discovery benchmark.
\n
Validation Result record Path
\n
Run the first-wave fixture proof from the Microcosm root:
\n
Exported source-body bundle inspection
\n
Inspect the exported source-body bundle. Until the exported fixture is refreshed with score-backed numeric rows, this command can return a blocked numeric verdict while still proving the manifest and body-floor boundary: Inspect the exported source-body bundle. Until the exported fixture is refreshed with score-backed numeric rows, this command can return a blocked numeric verdict while still proving the manifest and body-floor boundary:
A pass means the runtime recomputed the winner from public numbers, matched the declared label, refused all eight forbidden cases, kept the result record metadata-only, and stayed inside the simulator-only ceiling. A pass means the runtime recomputed the winner from public numbers, matched the declared label, refused all eight forbidden cases, kept the result record metadata-only, and stayed inside the simulator-only ceiling. The validation result records named here are the authority for a current run; older archived records predate the numeric rows and are not the numeric proof.
\n
Scope boundary
\n
Scope limit
\n
The bounded winner-recomputation claim
\n
The strongest claim the evidence supports is narrow: over public synthetic inputs, this component recomputes a lab-shaped fixture's selected candidate from score-backed rows, refuses eight named dangerous or overclaiming classes, verifies copied source-module bodies by digest, and writes a metadata-only result record, all reproducible with the commands above. The strongest claim the evidence supports is narrow: over public synthetic inputs, this component recomputes a lab-shaped fixture's selected candidate from score-backed rows, refuses eight named dangerous or overclaiming classes, verifies copied source-module bodies by digest, and writes a metadata-only result record, all reproducible with the commands above. That is the proof boundary. Fixture numbers are public replay coordinates for a safety-gated contract, not experimental measurements.
\n
Claims outside the public fixture
\n
It refuses to claim more. This is not a laboratory. It refuses to claim more. This is not a laboratory. It does not synthesize materials, give wetlab or hazardous-synthesis instructions, state reagent quantities, target controlled or bioactive compounds, control a robot, export a private lab notebook, use live account secrets, validate live assay data, rank real compounds, call a provider, assert a discovery, or post a benchmark claims. Its scope limit is the bundle ceiling: copied public source-module bodies, metadata-only simulator-only records, and artifact safety and refusal validation only. It cannot prove chemical safety, regulatory suitability, lab readiness, launch-scope decision, or public sharing.
\n",
"components": [
{
"anchor": "component-materials_chemistry_closed_loop_lab_safety_replay",
"component_id": "materials_chemistry_closed_loop_lab_safety_replay",
"public_label": "Materials Lab-Safety Refusal Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 6
},
"link_count": 16,
"links": [
{
"id": "materials_chemistry_closed_loop_lab_safety_replay",
"kind": "organ",
"kind_label": "component",
"label": "Materials Lab-Safety Refusal Replay",
"relation": "explains",
"url": "component-materials-chemistry-closed-loop-lab-safety-replay.html"
},
{
"id": "mechanism.materials_chemistry_closed_loop_lab_safety_replay.validates_public_materials_lab_safety_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public materials lab safety replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-materials-chemistry-closed-loop-lab-safety-replay-validates-public-materials-lab-safety-replay"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
},
{
"id": "paper_module.research_replication_rubric_artifact_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Research Replication Rubric Artifact Replay",
"relation": "depends on",
"url": "paper-module-research-replication-rubric-artifact-replay.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.spatial_world_model_counterfactual_simulation_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Spatial World Model Counterfactual Simulation Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[52:paper_module.spatial_world_model_counterfactual_simulation_replay].depends_on",
"url": "paper-module-spatial-world-model-counterfactual-simulation-replay.html"
}
],
"paper_module_id": "paper_module.materials_chemistry_closed_loop_lab_safety_replay",
"reader_packet": {
"abstract": "The materials chemistry lab-safety replay validates metadata-only simulator-only closed-loop rows before any materials-lab or discovery language is allowed.",
"authority_ceiling": "Copied public Lab/Evolve source/control/result record/standard bodies, metadata-only simulator-only fixture result records, runtime bundle result records, and artifact safety/refusal validation only; no wetlab execution, hazardous synthesis guidance, reagent quantity, controlled or bioactive target, live assay, robot command, private lab notebook, external model access, discovery claim, benchmark claims, launch-scope decision, publishing-scope decision, or product-progress evidence.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.materials_chemistry_closed_loop_lab_safety_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-materials-chemistry-closed-loop-lab-safety-replay-diagram-1",
"html_url": "docs/paper-module-materials-chemistry-closed-loop-lab-safety-replay.html#paper-module-materials-chemistry-closed-loop-lab-safety-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD inputs[\"six public inputs protocol, policy, candidates, experiments, assays, decisions\"] rows[\"per-row field + link checks\"] numeric[\"numeric recompute 0.45 assay + 0.35 active + 0.20 gate\"] labelcheck[\"recomputed pick vs declared label gate >= 0.70\"] negatives[\"eight forbidden fixtures\"] refuse[\"any MATERIALS_*_FORBIDDEN observed\"] manifest[\"source-module manifest digest check (bundle mode)\"] scan[\"secret exclusion scan\"] status[\"status + safety verdict\"] accepted[\"public_safe_simulator_replay_accepted\"] blocked[\"blocked_public_safety_boundary\"] inputs --> rows inputs --> numeric numeric --> labelcheck labelcheck --> blocked labelcheck --> status negatives --> refuse refuse --> blocked refuse --> status manifest --> status scan --> status rows --> status status --> accepted status --> blocked",
"source_format": "mermaid",
"source_hash": "sha256:a74bef88decf14aa2a964bc27f2a41aa013f3816b5078f54e41e3a5ce7c70433",
"source_line_end": 99,
"source_line_start": 72
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.research_replication_rubric_artifact_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Research Replication Rubric Artifact Replay",
"relation": "depends on",
"url": "paper-module-research-replication-rubric-artifact-replay.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "materials_chemistry_closed_loop_lab_safety_replay",
"kind": "organ",
"kind_label": "component",
"label": "Materials Lab-Safety Refusal Replay",
"relation": "explains",
"url": "component-materials-chemistry-closed-loop-lab-safety-replay.html"
},
{
"id": "mechanism.materials_chemistry_closed_loop_lab_safety_replay.validates_public_materials_lab_safety_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public materials lab safety replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-materials-chemistry-closed-loop-lab-safety-replay-validates-public-materials-lab-safety-replay"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.materials_chemistry_closed_loop_lab_safety_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 181,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/materials_chemistry_closed_loop_lab_safety_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "self-driving laboratory",
"section_id": "self-driving-lab-structure-under-fixture-evidence",
"url": "https://en.wikipedia.org/wiki/Self-driving_laboratory"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "\"Closed-loop materials lab\" is easy language to overclaim. A fixture can look like an autonomous discovery loop while carrying nothing that should be spoken aloud: wetlab steps, reagent quantities, a controlled or bioactive target, robot commands, or a bare assertion that some material was found. \"Closed-loop materials lab\" is easy language to overclaim. A fixture can look like an autonomous discovery loop while carrying nothing that should be spoken aloud: wetlab steps, reagent quantities, a controlled or bioactive target, robot commands, or a bare assertion that some material was found. This component sits in front of that language and answers one question. Is a closed-loop-lab-shaped fix\u2026",
"level": 4,
"order": 1,
"section_id": "closed-loop-materials-lab-is-easy-language-to-overclaim",
"source_line_end": 7,
"source_line_start": 3,
"title": "\"Closed-loop materials lab\" is easy language to overclaim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Its stable id is materialschemistryclosedlooplabsafetyreplay, and its public surface name (PUBLICSURFACENAME) is materialschemistryartifactsafetyrefusalvalidator.** Its stable id is materialschemistryclosedlooplabsafetyreplay, and its public surface name (PUBLICSURFACENAME) is materialschemistryartifactsafetyrefusalvalidator. The promise-sounding \"closed-loop replay\" name was reframed on purpose: nothing here executes a wetlab loop or commands a robot. The component reads public simulator rows, recomputes a winner, checks refusals, and writes a metadata-only result record. It emits a verdict of either publicsafesimulatorreplayaccepted or blockedpublicsafetyboundary, plus the findings that p\u2026",
"level": 4,
"order": 2,
"section_id": "its-stable-id-is-materials-chemistry-closed-loop-lab-safety-replay",
"source_line_end": 12,
"source_line_start": 8,
"title": "Its stable id is `materials_chemistry_closed_loop_lab_safety_replay`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The part worth naming is that the component does not trust the fixture's own conclusion. A plain replay would read a declared \"selected candidate\" label and report it. The part worth naming is that the component does not trust the fixture's own conclusion. A plain replay would read a declared \"selected candidate\" label and report it. This one recomputes the winner from public numbers and treats a mismatch with the declared label as a failure, not a footnote.",
"level": 4,
"order": 3,
"section_id": "the-part-worth-naming-is-that-the-component-does-not-trust-the-fixture-s-own-conclusion",
"source_line_end": 17,
"source_line_start": 13,
"title": "The part worth naming is that the component does not trust the fixture's own conclusion"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 19,
"source_line_start": 18,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A loose reader of a lab fixture accepts a declared winner and a \"safe\" flag at face value. A loose reader of a lab fixture accepts a declared winner and a \"safe\" flag at face value. That fails two ways. A stale or flattering label passes even when the numbers no longer support it, and dangerous content slips through because nothing is watching for it. This component closes both gaps. It rederives the selection from score-backed rows, and it treats eight named categories of dangerous or overclaiming content as things to refuse rather than ignore. The result is fixture-bound evidence that the loop's structure is inspectable on a simulator, with no real lab, chemicals, or robot involved.",
"level": 4,
"order": 5,
"section_id": "a-loose-reader-of-a-lab-fixture-accepts-a-declared-winner",
"source_line_end": 24,
"source_line_start": 20,
"title": "A loose reader of a lab fixture accepts a declared winner"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 6,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "There are two entrypoints. run validates a first-wave fixture directory with negative cases included. There are two entrypoints. run validates a first-wave fixture directory with negative cases included. runlabbundle validates an exported bundle with includenegative=False and requires a copied source-module manifest. Both call buildresult to compute the payload, attach freshnessbasis, and write the result record. runlabbundle can reuse a still-fresh prior record through freshlabbundle_receipt when the input digests are unchanged. buildresult runs in this order.",
"level": 4,
"order": 7,
"section_id": "there-are-two-entrypoints",
"source_line_end": 42,
"source_line_start": 27,
"title": "There are two entrypoints"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The numeric recompute is the heart. numericreplayresult reads publicsafetygatescore from each candidate, publicassayproxyvalue from each assay, and publicactivelearningscore from each decision. The numeric recompute is the heart. numericreplayresult reads publicsafetygatescore from each candidate, publicassayproxyvalue from each assay, and publicactivelearningscore from each decision. numericvalue accepts a number only when it is finite and in [0, 1]; anything else yields MATERIALSNUMERICREPLAYSCOREOUTOFRANGE or MATERIALSNUMERICREPLAYSCOREREQUIRED. When no score field is present at all, the run blocks with MATERIALSNUMERICREPLAYPOLICYREQUIRESSCOREBACKEDROWS (policy active) or MATERIALSNUMER\u2026",
"level": 4,
"order": 8,
"section_id": "the-numeric-recompute-is-the-heart",
"source_line_end": 47,
"source_line_start": 43,
"title": "The numeric recompute is the heart"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "For each linked candidate the weighted score is 0.45 assayproxy + 0.35 activelearning + 0.20 safetygate. For each linked candidate the weighted score is 0.45 assayproxy + 0.35 activelearning + 0.20 safetygate. The winner is the maximum by score, ties broken by id. Only after that pick is computed does the component compare it to the policy's declared expected label. A mismatch is MATERIALSNUMERICREPLAYEXPECTEDLABEL_STALE. Numeric replay reports pass only when there are no findings and at least one verified row. The baseline fixture holds four candidates.",
"level": 4,
"order": 9,
"section_id": "the-weighted-candidate-score",
"source_line_end": 61,
"source_line_start": 48,
"title": "The weighted candidate score"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The recompute selects matpolymermembrane001 at 0.917, which matches the policy's expectedselectedcandidatematerialid.** The recompute selects matpolymermembrane001 at 0.917, which matches the policy's expectedselectedcandidatematerialid. The policy declares minimumsafetygatescore 0.70 and the selection rule maxweightedpublicassayactivelearningandsafetygate_score.",
"level": 4,
"order": 10,
"section_id": "the-recompute-selects-mat-polymer-membrane-001-at-0-917",
"source_line_end": 66,
"source_line_start": 62,
"title": "The recompute selects `mat_polymer_membrane_001` at 0.917"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "buildresult folds every check into one policypassed gate: the four row sets are non-empty, no positive finding fired, no forbidden content leaked into the encoded rows, no expected negative case is missing, every row is simulatoronly, every hazard flag is false, and numeric replay passed.** buildresult folds every check into one policypassed gate: the four row sets are non-empty, no positive finding fired, no forbidden content leaked into the encoded rows, no expected negative case is missing, every row is simulatoronly, every hazard flag is false, and numeric replay passed. The overall status is a pass only when policy_passed holds, the secret scan passed, the replay graph passed, and, in\u2026",
"level": 4,
"order": 11,
"section_id": "build-result-folds-every-check-into-one-policy-passed-gate",
"source_line_end": 100,
"source_line_start": 67,
"title": "`_build_result` folds every check into one `policy_passed` gate"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 102,
"source_line_start": 101,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "EXPECTEDNEGATIVECASES names eight forbidden fixtures, each with the code the component must raise when it observes one. EXPECTEDNEGATIVECASES names eight forbidden fixtures, each with the code the component must raise when it observes one. A run that fails to refuse any of them lists the case under expected_missing and blocks.",
"level": 4,
"order": 13,
"section_id": "expected-negative-cases-names-eight-forbidden-fixtures",
"source_line_end": 118,
"source_line_start": 103,
"title": "`EXPECTED_NEGATIVE_CASES` names eight forbidden fixtures"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The numeric perturbations are covered too. Lowering the membrane gate to 0.52 moves the pick to matsolidelectrolyte002 and blocks with a safety-gate failure.** The numeric perturbations are covered too. Lowering the membrane gate to 0.52 moves the pick to matsolidelectrolyte002 and blocks with a safety-gate failure. Declaring a stale expected label while the numbers select a different candidate raises the stale-label code. Removing score rows while the policy is active raises the score-backed-rows code with verifiednumericrowcount at 0.",
"level": 4,
"order": 14,
"section_id": "the-numeric-perturbations-are-covered-too",
"source_line_end": 123,
"source_line_start": 119,
"title": "The numeric perturbations are covered too"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 125,
"source_line_start": 124,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The shape borrowed here is the self-driving laboratory, where a propose-run-measure loop is paired with safety interlocks that can refuse an unsafe experiment. The shape borrowed here is the self-driving laboratory, where a propose-run-measure loop is paired with safety interlocks that can refuse an unsafe experiment. This component keeps the loop-plus-safety-gate structure but runs it over a simulator fixture and adds a numeric recompute so the declared winner is checked rather than trusted. The engineering lineage is ordinary refusal-first input validation applied to a lab-shaped record. It is not a real laboratory controller, a chemical-safety authority, or a discovery benchmark.",
"level": 4,
"order": 16,
"section_id": "self-driving-lab-structure-under-fixture-evidence",
"source_line_end": 130,
"source_line_start": 126,
"title": "Self-driving-lab structure under fixture evidence"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run the first-wave fixture proof from the Microcosm root:",
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 139,
"source_line_start": 131,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Inspect the exported source-body bundle. Until the exported fixture is refreshed with score-backed numeric rows, this command can return a blocked numeric verdict while still proving the manifest and body-floor boundary: Inspect the exported source-body bundle. Until the exported fixture is refreshed with score-backed numeric rows, this command can return a blocked numeric verdict while still proving the manifest and body-floor boundary: Run the focused regression suite:",
"level": 4,
"order": 18,
"section_id": "exported-source-body-bundle-inspection",
"source_line_end": 163,
"source_line_start": 140,
"title": "Exported source-body bundle inspection"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means the runtime recomputed the winner from public numbers, matched the declared label, refused all eight forbidden cases, kept the result record metadata-only, and stayed inside the simulator-only ceiling. A pass means the runtime recomputed the winner from public numbers, matched the declared label, refused all eight forbidden cases, kept the result record metadata-only, and stayed inside the simulator-only ceiling. The validation result records named here are the authority for a current run; older archived records predate the numeric rows and are not the numeric proof.",
"level": 4,
"order": 19,
"section_id": "a-pass-means-the-runtime-recomputed-the-winner-from-public-numbers",
"source_line_end": 168,
"source_line_start": 164,
"title": "A pass means the runtime recomputed the winner from public numbers"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 170,
"source_line_start": 169,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim the evidence supports is narrow: over public synthetic inputs, this component recomputes a lab-shaped fixture's selected candidate from score-backed rows, refuses eight named dangerous or overclaiming classes, verifies copied source-module bodies by digest, and writes a metadata-only result record, all reproducible with the commands above. The strongest claim the evidence supports is narrow: over public synthetic inputs, this component recomputes a lab-shaped fixture's selected candidate from score-backed rows, refuses eight named dangerous or overclaiming classes, verifies copied source-module bodies by digest, and writes a metadata-only result record, all reproducible\u2026",
"level": 4,
"order": 21,
"section_id": "the-bounded-winner-recomputation-claim",
"source_line_end": 175,
"source_line_start": 171,
"title": "The bounded winner-recomputation claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It refuses to claim more. This is not a laboratory. It refuses to claim more. This is not a laboratory. It does not synthesize materials, give wetlab or hazardous-synthesis instructions, state reagent quantities, target controlled or bioactive compounds, control a robot, export a private lab notebook, use live account secrets, validate live assay data, rank real compounds, call a provider, assert a discovery, or post a benchmark claims. Its scope limit is the bundle ceiling: copied public source-module bodies, metadata-only simulator-only records, and artifact safety and refusal validation only. It cannot prove chemical safety, regulatory suitability, lab readiness, launch-scope decision, o\u2026",
"level": 4,
"order": 22,
"section_id": "claims-outside-the-public-fixture",
"source_line_end": 180,
"source_line_start": 176,
"title": "Claims outside the public fixture"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/materials_chemistry_closed_loop_lab_safety_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The materials chemistry lab-safety replay validates metadata-only simulator-only closed-loop rows before any materials-lab or discovery language is allowed."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "materials_chemistry_closed_loop_lab_safety_replay.py",
"path": "src/microcosm_core/organs/materials_chemistry_closed_loop_lab_safety_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for materials chemistry lab-safety fixture validation, exported bundle validation, source-module manifest checks, Lab/Evolve replay graph construction, secret-exclusion scans, negative-case enforcement, scope limits, and metadata-only result record writing.",
"symbols": [
"run",
"run_lab_bundle",
"_source_module_manifest_result",
"_source_open_body_import_summary",
"_build_result",
"_freshness_basis",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/materials_chemistry_closed_loop_lab_safety_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/materials_chemistry_closed_loop_lab_safety_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/materials_chemistry_closed_loop_lab_safety_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Materials Chemistry Closed-Loop Lab-Safety Replay is the public lab-safety claim boundary for Microcosm. It checks candidate material refs, safety-screen refs, simulator-only assay refs, active-learning decisions, failure taxonomy refs, cold replay refs, source bundle hashes, Lab/Evolve replay graph evidence, copied source/control/result record/standard bodies, metadata-only result record policy, and eight lab-safety overclaim negative cases while keeping wetlab protocols, hazardous synthesis steps, reagent quantities, controlled or bioactive targets, live lab account secrets, robot commands, private lab notebook bodies, live assay data, discovery claims, benchmark claims, model-output data, source notes, and launch-scope decision out of result records.",
"summary_line": "The materials chemistry lab-safety replay validates metadata-only simulator-only closed-loop rows before any materials-lab or discovery language is allowed.",
"title": "Materials Chemistry Closed-Loop Lab-Safety Replay"
},
{
"anchor": "paper-module-certificate-kernel-execution-lab",
"authority_ceiling": "Public Lean/Lake subprocess witness, copied source proof/tool/profile bodies, first-wave fixture result records, and exported bundle result records only; no general formal-result correctness, private proof body export, external model access, oracle authority, source-file changes, benchmark solve-rate, launch-scope decision, publishing-scope decision, or whole-system proof claim.",
"body_html": "
Lab answers question
\n
The lab answers one question: did a real Lean verifier just compile a small public project and accept the declared certificate rows, with the command, return code, and hashes recorded. The certificate kernel execution lab answers one narrow question. Did a real Lean verifier compile a small public project and accept the declared certificate rows, here and now, with the exact build command, the return code, and the source hashes on record? Everything else the component produces is bookkeeping that keeps that answer honest.
\n
Runs lean project
\n
It runs the Lean project in a temporary copy, checks rows against a decidable Boolean checker, and writes a metadata-only record of accepted, residual, and rejected rows with command and hashes. It runs a source-available Lean/Lake project inside a temporary copy, checks generated certificate rows against a decidable Boolean checker, and writes a structured result record. The record names which rows the verifier accepted, which it left unresolved, and which it rejected, together with the build command identity and the file hashes. Proof text, provider output, and private source stay out of the record, and that exclusion is written down rather than dropped in silence.
\n
Both public entrypoints, run and run_certificate_bundle, build one payload through _build_result and write it as metadata-only public evidence. The public entrypoints are run and run_certificate_bundle in src/microcosm_core/organs/certificate_kernel_execution_lab.py. Both build one payload through _build_result and write it as metadata-only public evidence.
\n
Purpose
\n
Facts as finite certificate plus decidable checker
\n
Proof-adjacent facts become a finite certificate plus a decidable validate : Cert -> Bool checker, so the agent supplies rows and Lean decides across two checker families. A large class of proof-adjacent facts can be written as a finite certificate plus a decidable checker shaped like validate : Cert -> Bool. The component never asks an agent to write a human proof. It asks for the right certificate rows and lets Lean decide. The fixture carries two checker families, NatSumCertificate over arithmetic and BoundedOrderCertificate over a bounded modular order, so sign-off is not one hard-coded shape.
\n
Sign-off needs exit-zero Lean, not a screenshot
\n
A row counts as accepted only when the runner shells out to Lean/Lake and gets exit code 0; sign-off without a genuine rejection is a replay, not evidence. A loose version of this would let a green screenshot or a confident model stand in for a check. This component refuses that. A row counts as accepted only when the runner shells out to Lean/Lake over a temporary copy of the public project and receives exit code 0. A result that can show only sign-off, with no genuine rejection, is treated as a replay, not as evidence.
\n
How it works
\n
Two entry points into _build_result
\n
run drives _build_result with the first-wave fixture and negative cases on; run_certificate_bundle drives it with the exported bundle and negatives off.run calls _build_result with input_mode="first_wave_fixture" and the negative cases on. run_certificate_bundle calls it with input_mode="exported_certificate_kernel_execution_lab_bundle" and the negative cases off, for the standalone public example. The stages below run in _build_result order.
\n
Fixing the source and record boundary
\n
First it enumerates and secret-scans the public inputs, verifies copied source bodies by class, anchors, and SHA-256, and projects them into a citable body floor. First it fixes the source and record boundary. _input_paths enumerates the public Lean files and JSON inputs. scan_paths checks them against core/private_state_forbidden_classes.json. _source_module_manifest_result verifies the exported bundle's copied source bodies by material class, target presence, required provenance anchors, and SHA-256 equality, and returns a finding for any row that fails one of those. _source_open_body_import_summary turns that into a body floor the record can cite by ref, hash, class, count, and verdict, with body_in_receipt: false.
\n
Gathering execution evidence from lake build
\n
Then it copies the project to a temp dir, runs lake build, and sets certificate rows through Lean: accepted rows must return 0, bad rows must fail in their predicted class. Then it gathers execution evidence. For the fixture path it copies the project into a temporary directory, records the public Lean imports, declarations, line counts, and hashes, and runs lake build MicrocosmCertificateLab. If the build returns 0 it sets the certificate transition rows through Lean. Accepted rows must return 0. Missing or bad certificate rows must fail in the residual class the fixture predicted. Typed-repair and bounded-policy rows must rerun inside their allowed action and artifact classes rather than editing arbitrary source.
\n
An obligation graph of public ids only
\n
Then it builds an obligation dependency graph citing public ids only, validated against observed transitions and rejecting any row carrying a proof body, model-output data, oracle template, or private obligation body. Then it builds a dependency graph that is weaker than a proof body and stronger than prose. Each obligation row cites public ids only: an obligation id, dependency ids, transition ids, certificate refs, an evidence kind, and an expected verifier class. The runner validates the rows as a graph over the transitions it actually observed and rejects any row that tries to carry a proof body, a model-output data, an oracle template, or a private obligation body.
\n
The pass floor in _build_result
\n
_build_result returns pass only when the secret scan is clean, Lean and Lake are present, the build returned 0, and the transition, repair, and policy counts each clear their floor; otherwise blocked._build_result then computes status. It returns pass only when the secret scan has zero blocking hits, Lean and Lake are both available, the build returned 0, at least eight declarations were analyzed, no expected negative case is missing, and the transition, repair, and policy counts each clear their floor. The pass floor requires at least five transitions, three accepted, two residual, one typed-repair effect, and one accepted policy rerun. For the bundle path it also requires the source-body check to pass. Anything short of that is blocked.
\n
Function
Role
run
Fixture entrypoint; runs the negative floor and writes the fixture records
run_certificate_bundle
Runs the standalone public bundle and writes BUNDLE_RESULT_NAME
_build_result
Loads inputs, scans, runs Lean/Lake, sets transitions, computes status
_source_module_manifest_result
Checks copied source bodies by class, presence, anchors, and SHA-256
_source_open_body_import_summary
Projects the checked bodies into a citable body floor
_receipt_freshness
Compares input and record mtimes so a current record can be reused
build_public_readout
Assembles the cold-reader readout from the written records
\nDiagram of the mechanism (8 steps).Diagram source & refs
flowchart TD inputs["public Lean files and JSON inputs"] scan["scan_paths secret gate"] sources["_source_module_manifest_result copied body check"] build["lake build MicrocosmCertificateLab"] transitions["set certificate rows through Lean"] dag["obligation graph over observed rows"] status["_build_result status pass or blocked"] record["metadata-only result record"] inputs --> scan scan --> sources sources --> build build --> transitions transitions --> dag dag --> status status --> record
\n
Negative cases
\n
Five rejections the fixture must observe
\n
EXPECTED_NEGATIVE_CASES names five rows that must each be rejected with their own error code, and a fixture run missing any of them is blocked because green-only is not evidence.EXPECTED_NEGATIVE_CASES names five rows that must be rejected, each with its own error code. A row that makes provider or oracle text visible to the verifier fails with CERTIFICATE_KERNEL_EXECUTION_PROVIDER_OR_ORACLE_VISIBLE. A typed-repair row carrying a proof body fails with CERTIFICATE_KERNEL_EXECUTION_CP2_PROOF_BODY_FORBIDDEN. A policy row that tries to change source files fails with CERTIFICATE_KERNEL_EXECUTION_EVOLVE_SCOPE_FORBIDDEN. A manifest row pointing at a non-public source refs fails with CERTIFICATE_KERNEL_EXECUTION_PRIVATE_SOURCE_REF_FORBIDDEN. An obligation row carrying a body fails with CERTIFICATE_KERNEL_EXECUTION_OBLIGATION_BODY_FORBIDDEN. A fixture run that does not observe all five is blocked, because a component that can only show green is not evidence that the rejection path is real.
\n
Prior Art Grounding
\n
Proof-carrying code and small-kernel Lean lineage
\n
The design follows the proof-carrying-code and proof-assistant traditions: an untrusted producer supplies a certificate a small trusted Lean kernel checks, with reproducible environments and a named evaluation-scope limit. This component sits in the proof-carrying and proof-assistant traditions. Necula's Proof-Carrying Code anchors the idea that an untrusted producer can supply a certificate checked by a small trusted verifier. The Lean theorem prover continues the small-kernel lineage, and LeanDojo shows why reproducible Lean environments and programmatic proof-state access matter for theorem-proving agents. PBLean is the nearest public reference for the reflective-checker research fork, where a verified Boolean checker imports certificates into Lean without building every explicit proof term. The Lean proof validation reference is the local reason that any move to native or compiler-backed evaluation must be named in the scope limit, not smuggled into an ordinary result record.
\n
Validation Result record Path
\n
Run the first-wave fixture into disposable records from the Microcosm root:
\n
Run the exported bundle through the same component:
A pass certifies only that the project compiled, accepted rows returned 0, each negative case failed in its predicted class, and no proof or provider body reached the record. A pass means the public Lean project compiled, the accepted rows returned 0, every expected negative case was rejected in its predicted class, and the result record carried no proof, provider, oracle, or private-source body.
\n
Scope boundary
\n
Scope limit
\n
The bounded strongest claim
\n
The strongest supported claim is bounded: a green run shows the public Lean project compiled and its certificate rows were accepted, rejected, or residual under a real local verifier, nothing more. The strongest claim this component supports is bounded. A green run shows that the declared public Lean project compiled and that the declared certificate rows were accepted, rejected, or left residual under a real local Lean/Lake verifier, with the command identity, return codes, hashes, and negative-case rows on record. The proof boundary is that local subprocess witness together with the fixture and bundle records, nothing more.
\n
What a green run still denies
\n
A green run still proves no theorem, claims no Mathlib coverage or solve rate, exports no proof bodies, mutates no source, and excludes launch. It does not establish the Erdos #257 theorem, claim Mathlib coverage, or report a benchmark solve rate. It does not export proof bodies, count provider or oracle output as proof, change source files, or include launch operations or public sharing. The Atlas card for this module is completed by the owning projection route and builder regeneration, not by editing this page.
\n",
"components": [
{
"anchor": "component-certificate_kernel_execution_lab",
"component_id": "certificate_kernel_execution_lab",
"public_label": "Certificate Kernel Execution Lab"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 1
},
"link_count": 6,
"links": [
{
"id": "certificate_kernel_execution_lab",
"kind": "organ",
"kind_label": "component",
"label": "Certificate Kernel Execution Lab",
"relation": "explains",
"url": "component-certificate-kernel-execution-lab.html"
},
{
"id": "mechanism.certificate_kernel_execution_lab.validates_public_certificate_kernel_execution",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public certificate kernel execution",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-certificate-kernel-execution-lab-validates-public-certificate-kernel-execution"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.verifier_lab_execution_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Execution Spine",
"relation": "depends on",
"url": "paper-module-verifier-lab-execution-spine.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.certificate_kernel_execution_lab",
"reader_packet": {
"abstract": "The certificate kernel execution lab validates bounded public Lean/Lake certificate-kernel rows before any proof-adjacent claim is allowed.",
"authority_ceiling": "Public Lean/Lake subprocess witness, copied source proof/tool/profile bodies, first-wave fixture result records, and exported bundle result records only; no general formal-result correctness, private proof body export, external model access, oracle authority, source-file changes, benchmark solve-rate, launch-scope decision, publishing-scope decision, or whole-system proof claim.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.certificate_kernel_execution_lab.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-certificate-kernel-execution-lab-diagram-1",
"html_url": "docs/paper-module-certificate-kernel-execution-lab.html#paper-module-certificate-kernel-execution-lab-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD inputs[\"public Lean files and JSON inputs\"] scan[\"scan_paths secret gate\"] sources[\"_source_module_manifest_result copied body check\"] build[\"lake build MicrocosmCertificateLab\"] transitions[\"set certificate rows through Lean\"] dag[\"obligation graph over observed rows\"] status[\"_build_result status pass or blocked\"] record[\"metadata-only result record\"] inputs --> scan scan --> sources sources --> build build --> transitions transitions --> dag dag --> status status --> record",
"source_format": "mermaid",
"source_hash": "sha256:a66da67ed01fc0f7f1659c734645b0fb546566c94e3e67b3480f175bda48c37d",
"source_line_end": 75,
"source_line_start": 57
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.verifier_lab_execution_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Execution Spine",
"relation": "depends on",
"url": "paper-module-verifier-lab-execution-spine.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "certificate_kernel_execution_lab",
"kind": "organ",
"kind_label": "component",
"label": "Certificate Kernel Execution Lab",
"relation": "explains",
"url": "component-certificate-kernel-execution-lab.html"
},
{
"id": "mechanism.certificate_kernel_execution_lab.validates_public_certificate_kernel_execution",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public certificate kernel execution",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-certificate-kernel-execution-lab-validates-public-certificate-kernel-execution"
}
],
"relation": "explains"
},
{
"count": 2,
"links": [
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.certificate_kernel_execution_lab",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 128,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/certificate_kernel_execution_lab.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Proof-Carrying Code",
"section_id": "proof-carrying-code-and-small-kernel-lean-lineage",
"url": "https://www.usenix.org/legacy/publications/library/proceedings/osdi96/full_papers/necula/html/node2.html"
},
{
"kind": "external_reference",
"label": "Lean theorem prover",
"section_id": "proof-carrying-code-and-small-kernel-lean-lineage",
"url": "https://www.microsoft.com/en-us/research/publication/the-lean-theorem-prover-system-description/"
},
{
"kind": "external_reference",
"label": "LeanDojo",
"section_id": "proof-carrying-code-and-small-kernel-lean-lineage",
"url": "https://arxiv.org/abs/2306.15626"
},
{
"kind": "external_reference",
"label": "PBLean",
"section_id": "proof-carrying-code-and-small-kernel-lean-lineage",
"url": "https://arxiv.org/abs/2602.08692"
},
{
"kind": "external_reference",
"label": "Lean proof validation reference",
"section_id": "proof-carrying-code-and-small-kernel-lean-lineage",
"url": "https://lean-lang.org/doc/reference/latest/ValidatingProofs/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "The lab answers one question: did a real Lean verifier just compile a small public project and accept the declared certificate rows, with the command, return code, and hashes recorded. The certificate kernel execution lab answers one narrow question. Did a real Lean verifier compile a small public project and accept the declared certificate rows, here and now, with the exact build command, the return code, and the source hashes on record? Everything else the component produces is bookkeeping that keeps that answer honest.",
"level": 4,
"order": 1,
"section_id": "lab-answers-question",
"source_line_end": 6,
"source_line_start": 3,
"title": "Lab answers question"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It runs the Lean project in a temporary copy, checks rows against a decidable Boolean checker, and writes a metadata-only record of accepted, residual, and rejected rows with command and hashes. It runs a source-available Lean/Lake project inside a temporary copy, checks generated certificate rows against a decidable Boolean checker, and writes a structured result record. The record names which rows the verifier accepted, which it left unresolved, and which it rejected, together with the build command identity and the file hashes. Proof text, provider output, and private source stay out of the record, and that exclusion is written down rather than dropped in silence.",
"level": 4,
"order": 2,
"section_id": "runs-lean-project",
"source_line_end": 10,
"source_line_start": 7,
"title": "Runs lean project"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Both public entrypoints, run and runcertificatebundle, build one payload through buildresult and write it as metadata-only public evidence. The public entrypoints are run and runcertificatebundle in src/microcosmcore/components/certificatekernelexecutionlab.py. Both build one payload through buildresult and write it as metadata-only public evidence.",
"level": 4,
"order": 3,
"section_id": "both-public-entrypoints",
"source_line_end": 14,
"source_line_start": 11,
"title": "Both public entrypoints"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Proof-adjacent facts become a finite certificate plus a decidable validate : Cert -> Bool checker, so the agent supplies rows and Lean decides across two checker families. A large class of proof-adjacent facts can be written as a finite certificate plus a decidable checker shaped like validate : Cert -> Bool. The component never asks an agent to write a human proof. It asks for the right certificate rows and lets Lean decide. The fixture carries two checker families, NatSumCertificate over arithmetic and BoundedOrderCertificate over a bounded modular order, so sign-off is not one hard-coded shape.",
"level": 4,
"order": 5,
"section_id": "facts-as-finite-certificate-plus-decidable-checker",
"source_line_end": 20,
"source_line_start": 17,
"title": "Facts as finite certificate plus decidable checker"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A row counts as accepted only when the runner shells out to Lean/Lake and gets exit code 0; sign-off without a genuine rejection is a replay, not evidence. A loose version of this would let a green screenshot or a confident model stand in for a check. This component refuses that. A row counts as accepted only when the runner shells out to Lean/Lake over a temporary copy of the public project and receives exit code 0. A result that can show only sign-off, with no genuine rejection, is treated as a replay, not as evidence.",
"level": 4,
"order": 6,
"section_id": "acceptance-needs-exit-zero-lean-not-a-screenshot",
"source_line_end": 24,
"source_line_start": 21,
"title": "Sign-off needs exit-zero Lean, not a screenshot"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run drives buildresult with the first-wave fixture and negative cases on; runcertificatebundle drives it with the exported bundle and negatives off. run calls buildresult with inputmode=\"firstwavefixture\" and the negative cases on. runcertificatebundle calls it with inputmode=\"exportedcertificatekernelexecutionlabbundle\" and the negative cases off, for the standalone public example. The stages below run in build_result order.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-into-build-result",
"source_line_end": 30,
"source_line_start": 27,
"title": "Two entry points into `_build_result`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "First it enumerates and secret-scans the public inputs, verifies copied source bodies by class, anchors, and SHA-256, and projects them into a citable body floor. First it fixes the source and record boundary. inputpaths enumerates the public Lean files and JSON inputs. scanpaths checks them against core/privatestateforbiddenclasses.json. sourcemodulemanifestresult verifies the exported bundle's copied source bodies by material class, target presence, required provenance anchors, and SHA-256 equality, and returns a finding for any row that fails one of those. sourceopenbodyimportsummary turns that into a body floor the record can cite by ref, hash, class, count, and verdict, with bodyin_rec\u2026",
"level": 4,
"order": 9,
"section_id": "fixing-the-source-and-record-boundary",
"source_line_end": 34,
"source_line_start": 31,
"title": "Fixing the source and record boundary"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Then it copies the project to a temp dir, runs lake build, and sets certificate rows through Lean: accepted rows must return 0, bad rows must fail in their predicted class. Then it gathers execution evidence. For the fixture path it copies the project into a temporary directory, records the public Lean imports, declarations, line counts, and hashes, and runs lake build MicrocosmCertificateLab. If the build returns 0 it sets the certificate transition rows through Lean. Accepted rows must return 0. Missing or bad certificate rows must fail in the residual class the fixture predicted. Typed-repair and bounded-policy rows must rerun inside their allowed action and artifact classes rather than\u2026",
"level": 4,
"order": 10,
"section_id": "gathering-execution-evidence-from-lake-build",
"source_line_end": 38,
"source_line_start": 35,
"title": "Gathering execution evidence from `lake build`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Then it builds an obligation dependency graph citing public ids only, validated against observed transitions and rejecting any row carrying a proof body, model-output data, oracle template, or private obligation body. Then it builds a dependency graph that is weaker than a proof body and stronger than prose. Each obligation row cites public ids only: an obligation id, dependency ids, transition ids, certificate refs, an evidence kind, and an expected verifier class. The runner validates the rows as a graph over the transitions it actually observed and rejects any row that tries to carry a proof body, a model-output data, an oracle template, or a private obligation body.",
"level": 4,
"order": 11,
"section_id": "an-obligation-graph-of-public-ids-only",
"source_line_end": 42,
"source_line_start": 39,
"title": "An obligation graph of public ids only"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "buildresult returns pass only when the secret scan is clean, Lean and Lake are present, the build returned 0, and the transition, repair, and policy counts each clear their floor; otherwise blocked. build_result then computes status. It returns pass only when the secret scan has zero blocking hits, Lean and Lake are both available, the build returned 0, at least eight declarations were analyzed, no expected negative case is missing, and the transition, repair, and policy counts each clear their floor. The pass floor requires at least five transitions, three accepted, two residual, one typed-repair effect, and one accepted policy rerun. For the bundle path it also requires the source-body ch\u2026",
"level": 4,
"order": 12,
"section_id": "the-pass-floor-in-build-result",
"source_line_end": 76,
"source_line_start": 43,
"title": "The pass floor in `_build_result`"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 78,
"source_line_start": 77,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVECASES names five rows that must each be rejected with their own error code, and a fixture run missing any of them is blocked because green-only is not evidence. EXPECTEDNEGATIVECASES names five rows that must be rejected, each with its own error code. A row that makes provider or oracle text visible to the verifier fails with CERTIFICATEKERNELEXECUTIONPROVIDERORORACLEVISIBLE. A typed-repair row carrying a proof body fails with CERTIFICATEKERNELEXECUTIONCP2PROOFBODYFORBIDDEN. A policy row that tries to change source files fails with CERTIFICATEKERNELEXECUTIONEVOLVESCOPEFORBIDDEN. A manifest row pointing at a non-public source refs fails with CERTIFICATEKERNELEXECUTIONPRIVATES\u2026",
"level": 4,
"order": 14,
"section_id": "five-rejections-the-fixture-must-observe",
"source_line_end": 82,
"source_line_start": 79,
"title": "Five rejections the fixture must observe"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 84,
"source_line_start": 83,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The design follows the proof-carrying-code and proof-assistant traditions: an untrusted producer supplies a certificate a small trusted Lean kernel checks, with reproducible environments and a named evaluation-scope limit. This component sits in the proof-carrying and proof-assistant traditions. Necula's Proof-Carrying Code anchors the idea that an untrusted producer can supply a certificate checked by a small trusted verifier. The Lean theorem prover continues the small-kernel lineage, and LeanDojo shows why reproducible Lean environments and programmatic proof-state access matter for theorem-proving agents. PBLean is the nearest public reference for the reflective-checker research fork, w\u2026",
"level": 4,
"order": 16,
"section_id": "proof-carrying-code-and-small-kernel-lean-lineage",
"source_line_end": 88,
"source_line_start": 85,
"title": "Proof-carrying code and small-kernel Lean lineage"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run the first-wave fixture into disposable records from the Microcosm root: Run the exported bundle through the same component:",
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 113,
"source_line_start": 89,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass certifies only that the project compiled, accepted rows returned 0, each negative case failed in its predicted class, and no proof or provider body reached the record. A pass means the public Lean project compiled, the accepted rows returned 0, every expected negative case was rejected in its predicted class, and the result record carried no proof, provider, oracle, or private-source body.",
"level": 4,
"order": 18,
"section_id": "what-a-pass-means-here",
"source_line_end": 117,
"source_line_start": 114,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "claim-ceiling",
"source_line_end": 119,
"source_line_start": 118,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest supported claim is bounded: a green run shows the public Lean project compiled and its certificate rows were accepted, rejected, or residual under a real local verifier, nothing more. The strongest claim this component supports is bounded. A green run shows that the declared public Lean project compiled and that the declared certificate rows were accepted, rejected, or left residual under a real local Lean/Lake verifier, with the command identity, return codes, hashes, and negative-case rows on record. The proof boundary is that local subprocess witness together with the fixture and bundle records, nothing more.",
"level": 4,
"order": 20,
"section_id": "the-bounded-strongest-claim",
"source_line_end": 123,
"source_line_start": 120,
"title": "The bounded strongest claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A green run still proves no theorem, claims no Mathlib coverage or solve rate, exports no proof bodies, mutates no source, and excludes launch. It does not establish the Erdos #257 theorem, claim Mathlib coverage, or report a benchmark solve rate. It does not export proof bodies, count provider or oracle output as proof, change source files, or include launch operations or public sharing. The Atlas card for this module is completed by the owning projection route and builder regeneration, not by editing this page.",
"level": 4,
"order": 21,
"section_id": "what-a-green-run-still-denies",
"source_line_end": 127,
"source_line_start": 124,
"title": "What a green run still denies"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/certificate_kernel_execution_lab.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The certificate kernel execution lab validates bounded public Lean/Lake certificate-kernel rows before any proof-adjacent claim is allowed."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "certificate_kernel_execution_lab.py",
"path": "src/microcosm_core/organs/certificate_kernel_execution_lab.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for certificate-kernel fixture validation, exported bundle validation, Lean/Lake subprocess execution, source-module manifest checks, public readout construction, negative-case enforcement, scope limits, and metadata-only result record writing.",
"symbols": [
"run",
"run_certificate_bundle",
"_source_module_manifest_result",
"_source_open_body_import_summary",
"_build_result",
"_receipt_freshness",
"build_public_readout",
"EXPECTED_NEGATIVE_CASES"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/certificate_kernel_execution_lab.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/certificate_kernel_execution_lab.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/certificate_kernel_execution_lab.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Certificate Kernel Execution Lab is the public proof-adjacent execution boundary for Microcosm. It checks a Lean/Lake certificate-kernel fixture, generated certificate rows, analyzer metadata, transition traces, typed CP2 action translations, bounded Evolve reruns, source-module manifest digests, copied Lean/tool/profile bodies, metadata-only result record policy, and four negative cases while keeping proof bodies, raw tactic scripts, model-output data, oracle ideal answers, oracle-needed premise ids, private source paths, stdout/stderr bodies, account secrets, private Erdos #257 proof bodies, launch-scope decision, benchmark solve-rate, and general theorem-proof authority out of result records. Its upgrade posture is certificate-engine first: shard and compress public certificate families while preserving the Lean/Lake result record boundary; treat reflective/native checker work as a named trust-register fork.",
"summary_line": "The certificate kernel execution lab validates bounded public Lean/Lake certificate-kernel rows before any proof-adjacent claim is allowed.",
"title": "Certificate Kernel Execution Lab"
},
{
"anchor": "paper-module-corpus-readiness-mathlib-absence-gate",
"authority_ceiling": "Public algorithmic projection over copied corpus/toolchain readiness system, first-wave fixture result records, and exported bundle result records only; no Lean/Lake rerun, Mathlib availability claim, formal-result correctness, proof body export, external model access, benchmark/corpus completeness, launch-scope decision, publishing-scope decision, or whole-system proof claim.",
"body_html": "
From recorded lean
\n
From a recorded Lean readiness report it answers, per corpus, whether the library is present and whether the Mathlib import lane resolves, pairing copied rows with a live probe.corpus_readiness_mathlib_absence_gate reads a recorded readiness report from a Lean math toolchain run and turns one question into an inspectable answer: for each math corpus, is it actually present on this host, and is the Mathlib import lane actually available, or not? It carries copied corpus and toolchain rows from a 2026-05-11 proof-state curriculum smoke run and pairs them with a live import check.
\n
Emits metadata-only readiness
\n
It emits a metadata-only readiness board and consumer verdicts: which corpora are absent, which are smoke-test only, and which tasks stay blocked until a Mathlib probe passes. The output is a readiness board and a set of consumer verdicts. It records which corpora are absent, which are usable only for translation smoke tests, and which downstream tasks stay blocked until a passing Mathlib probe exists. Everything it emits is metadata-only: no proof text, no provider output, no non-public paths.
\n
Is not prover
\n
It is not a prover or Lake build but a gate that forces a formal-math route to expose corpus readiness before any proof route counts as usable. It is not a theorem prover, a Lake build, or a Mathlib availability claim. It is a gate that forces a formal-math route to expose corpus readiness before any proof route is treated as usable.
\n
Purpose
\n
Roster presence mistaken for usability
\n
Formal-math agents conflate a corpus appearing on a roster with it being usable for their proof route, so a missing library surfaces only later as a broken import. Formal-math agents fail in one specific way. They treat "there is a corpus" as if it meant "this corpus is usable for the proof route I am about to take". A roster lists miniF2F, PutnamBench, ProofNet, LeanDojo and Mathlib, the agent assumes the libraries are present, and the failure only surfaces later as a broken import or a tactic that needs a premise the host cannot resolve.
\n
Alias field must agree with a live probe
\n
The loose check trusts a mathlib_available field or PASS status; this gate refuses that drift, flagging any row whose alias claim a live probe does not confirm. The loose version of this check trusts the roster. It reads a mathlib_available field, or a PASS lean status, and marks the corpus ready. That is the drift the gate is built to stop. A stale alias field cannot turn the gate green on its own. It has to agree with a live probe or the row is flagged.
\n
How it works
\n
Two entrypoints funnel into _build_result
\n
run and run_projection_bundle both call _build_result, which loads two inputs, scans paths against the exclusion policy, then folds four ordered checks into one result and board. The two entrypoints run and run_projection_bundle both call _build_result. run handles the first-wave fixture with its negative cases. run_projection_bundle handles an exported bundle without negative cases and requires a source-module manifest. _build_result loads the two input files, scans every input path against the non-public-state exclusion policy, then runs four checks in order and folds them into one result and a readiness board.
\nDiagram of the mechanism (7 steps).Diagram source & refs
Source refs
four source digests, parse JSON, run probe
validate_runtime_source_artifacts
Std compiles, Mathlib import rejected
runtime_lean_import_probe
seven corpus rows checked against the probe
validate_corpus_readiness
verdicts derived from readiness facts
validate_consumer_gate_cases
copied source digests for the bundle
validate_source_module_imports
flowchart TD input["Fixture or exported bundle input corpus rows plus consumer gate cases"] artifacts["validate_runtime_source_artifacts four source digests, parse JSON, run probe"] probe["runtime_lean_import_probe Std compiles, Mathlib import rejected"] corpus["validate_corpus_readiness seven corpus rows checked against the probe"] gates["validate_consumer_gate_cases verdicts derived from readiness facts"] imports["validate_source_module_imports copied source digests for the bundle"] result["_build_result one metadata-only result and board"] input --> artifacts artifacts --> probe probe --> corpus corpus --> gates input --> imports gates --> result imports --> result
\n
The live import probe: Std compiles, Mathlib rejected
\n
validate_runtime_source_artifacts digest-checks four source refs and runs a probe that expects Std to compile and Mathlib to be rejected, setting mathlib_lake_project_import_available false when the lane fails.validate_runtime_source_artifacts anchors the run to four source refs: the corpus readiness rows, a tactic-affordance probe, the Mathlib import probe Lean file, and a tactic portfolio availability file. It checks each file's SHA-256 against a recorded digest, parses the JSON ones, and calls runtime_lean_import_probe. That probe writes two small Lean files into a temporary directory. One imports Std and is expected to compile. One imports Mathlib and is expected to be rejected with the toolchain's own unknown module prefix 'Mathlib' error. The probe also confirms lake is on the path. It never runs lake build and never exports any Lean body. Its result sets mathlib_lake_project_import_available, which stays false whenever the Mathlib lane does not resolve.
\n
Corpus rows normalised against probe evidence
\n
validate_corpus_readiness marks a corpus usable only when it exists, carries a Lake file, and the probe agrees; contradicting alias claims are flagged and every unavailable lane is blocked.validate_corpus_readiness normalises the seven corpus rows against that runtime evidence. A corpus is marked usable for Mathlib-dependent work only when the runtime row says it exists, carries a Lake file, and the Mathlib import lane is available. If a row claims mathlib_lake_project_import_available or any present alias in MATHLIB_PRESENT_ALIAS_FIELDS while the probe disagrees, the row is flagged with CORPUS_READINESS_RUNTIME_PROBE_CONTRADICTION. Absent corpora and every unavailable Mathlib lane land in blocked_capabilities. In the current fixture the probe stays false, so all seven Mathlib lanes are blocked.
\n
Consumer verdicts derived from readiness facts
\n
validate_consumer_gate_cases recomputes each verdict from readiness facts, not the case's own label, blocking cases needing an unavailable Mathlib lane or an absent corpus; only the miniF2F smoke passes.validate_consumer_gate_cases derives each verdict from the readiness facts instead of trusting a case's own expected-decision label. A case is blocked if it requires the Mathlib lake project import while that lane is unavailable, or if its target corpus is absent. Of the seven cases, one passes: the miniF2F Lean3 translation smoke, which does not require a Mathlib project. The other six stay blocked. A case that never set readiness_gate_checked is flagged with CONSUMER_READINESS_GATE_UNCHECKED.
\n
Exported-bundle source floor and final status
\n
validate_source_module_imports enforces copied-body digests and byte equality, and _build_result passes only when no negative case is missing, the secret scan is clean, and the source floor holds.validate_source_module_imports adds the exported-bundle floor. It requires the manifest to declare copied_non_secret_macro_body, each row to use a body material class, each copied file to match its recorded source and target digest, and no row to carry a body inside a result record. Exact-copy rows must preserve byte equality; rewrite rows must match the declared private-absolute-path rewrite recipe exactly. _build_result then computes one status: it passes only when no expected negative case is missing, the secret scan has no blocking hit, the source-module floor holds, and no unexpected finding appears.
\n
Negative cases
\n
Fixture carries five
\n
The fixture carries five named negative inputs, declared in EXPECTED_NEGATIVE_CASES. Each one must be observed with its error code or the run is blocked:
\n
mathlib_available_without_probe must raise MATHLIB_AVAILABILITY_OVERCLAIM when Mathlib is claimed available with no passing probe.
consumer_skips_readiness_gate must raise CONSUMER_SKIPS_CORPUS_READINESS_GATE when a consumer attempts Mathlib-dependent work without checking readiness.
private_corpus_source_ref must raise PRIVATE_CORPUS_SOURCE_REF_FORBIDDEN when a source ref points at a non-public paths.
proof_body_leakage must raise CORPUS_READINESS_PROOF_BODY_FORBIDDEN when a corpus row carries a proof or provider body field.
release_overclaim must raise CORPUS_READINESS_RELEASE_OVERCLAIM when the metadata tries to include launch operations or proof authority.
\n
Prior Art Grounding
\n
Grounded in Lean corpus and theorem-proving work
\n
The design draws on Lean corpus and neural theorem-proving work where library availability and benchmark splits are part of the claim, requiring probes and digests before tactic-routing language is allowed. This component sits in Lean corpus and neural theorem-proving work where library availability, premise access, and benchmark splits are part of the claim. The Lean mathematical library establishes Mathlib as a large community-maintained formal mathematics corpus, miniF2F gives a cross-system benchmark for formal Olympiad statements, and LeanDojo shows why reproducible corpus extraction and accessible-premise metadata matter for theorem-proving agents. The gate borrows the discipline: corpus rows, Mathlib probes, blocked consumer cases, and copied-source digests must be visible before retrieval or tactic-routing language is allowed. It does not claim Mathlib is present or that any theorem was proved.
\n
Validation Result record Path
\n
Reader-verifiable commands, run from the microcosm-substrate/ public root:
A pass means fixture and exported bundle agree on readiness, Mathlib absence, blocked consumers, digests, and the five negative cases; it reruns no Lean, proves nothing, and installs no Mathlib. A pass means the fixture and the exported bundle agree about corpus readiness, Mathlib absence, blocked consumers, copied source digests, metadata-only result files, and the five negative cases. It does not rerun Lean or Lake, prove any theorem, or install Mathlib.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The strongest claim is narrow: a fixture and exported bundle agree on readiness, Mathlib absence, verdicts, digests, and leakage guards, bounded to projection over copied material plus one import probe. The strongest claim the evidence supports is narrow: a public fixture and an exported bundle agree about corpus readiness, Mathlib absence, blocked consumer verdicts, copied source-module digests, metadata-only result files, and the negative leakage guards. The proof boundary is algorithmic projection over copied corpus and toolchain material plus one bounded import probe.
\n
What stays refused past the line
\n
AUTHORITY_CEILING sets every outward flag false, so the component proves no install, runs no Lake build, validates no theorem, calls no provider, and approves no launch. It refuses everything past that line. The scope limit in AUTHORITY_CEILING sets mathlib_lake_project_import_authorized, formal_proof_authority, provider_calls_authorized, and release_authorized all to false. This component does not establish Mathlib is installed, run a Lake build, validate formal-result correctness, benchmark corpus quality, call a provider, expose any proof or provider body, change source files, or approve launch.
\n",
"components": [
{
"anchor": "component-corpus_readiness_mathlib_absence_gate",
"component_id": "corpus_readiness_mathlib_absence_gate",
"public_label": "Corpus Readiness Mathlib Absence Gate"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 2,
"organ": 1,
"paper_module": 1,
"principle": 1
},
"link_count": 7,
"links": [
{
"id": "corpus_readiness_mathlib_absence_gate",
"kind": "organ",
"kind_label": "component",
"label": "Corpus Readiness Mathlib Absence Gate",
"relation": "explains",
"url": "component-corpus-readiness-mathlib-absence-gate.html"
},
{
"id": "mechanism.corpus_readiness_mathlib_absence_gate.validates_public_corpus_readiness_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public corpus readiness boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-corpus-readiness-mathlib-absence-gate-validates-public-corpus-readiness-boundary"
},
{
"id": "mechanism.corpus_readiness_mathlib_absence_gate.validates_public_mathlib_absence_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public mathlib absence boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-corpus-readiness-mathlib-absence-gate-validates-public-mathlib-absence-boundary"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depends on",
"url": "paper-module-tactic-portfolio-availability.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.formal_math_lean_proof_witness",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Lean Proof Witness",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[22:paper_module.formal_math_lean_proof_witness].depends_on",
"url": "paper-module-formal-math-lean-proof-witness.html"
},
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[40:paper_module.tactic_portfolio_availability].depends_on",
"url": "paper-module-tactic-portfolio-availability.html"
},
{
"id": "paper_module.undeclared_library_prior_classifier",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Undeclared Library Prior Classifier",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[55:paper_module.undeclared_library_prior_classifier].depends_on",
"url": "paper-module-undeclared-library-prior-classifier.html"
},
{
"id": "paper_module.verifier_lab_kernel",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Kernel",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[0:paper_module.verifier_lab_kernel].depends_on",
"url": "paper-module-verifier-lab-kernel.html"
}
],
"paper_module_id": "paper_module.corpus_readiness_mathlib_absence_gate",
"reader_packet": {
"abstract": "The corpus readiness Mathlib absence gate validates copied corpus/toolchain readiness bodies before any Mathlib-dependent proof or retrieval claim is allowed.",
"authority_ceiling": "Public algorithmic projection over copied corpus/toolchain readiness system, first-wave fixture result records, and exported bundle result records only; no Lean/Lake rerun, Mathlib availability claim, formal-result correctness, proof body export, external model access, benchmark/corpus completeness, launch-scope decision, publishing-scope decision, or whole-system proof claim.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.corpus_readiness_mathlib_absence_gate.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-corpus-readiness-mathlib-absence-gate-diagram-1",
"html_url": "docs/paper-module-corpus-readiness-mathlib-absence-gate.html#paper-module-corpus-readiness-mathlib-absence-gate-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD input[\"Fixture or exported bundle input corpus rows plus consumer gate cases\"] artifacts[\"validate_runtime_source_artifacts four source digests, parse JSON, run probe\"] probe[\"runtime_lean_import_probe Std compiles, Mathlib import rejected\"] corpus[\"validate_corpus_readiness seven corpus rows checked against the probe\"] gates[\"validate_consumer_gate_cases verdicts derived from readiness facts\"] imports[\"validate_source_module_imports copied source digests for the bundle\"] result[\"_build_result one metadata-only result and board\"] input --> artifacts artifacts --> probe probe --> corpus corpus --> gates input --> imports gates --> result imports --> result",
"source_format": "mermaid",
"source_hash": "sha256:2a184f0787b1543ccd1b8f4d2c6811a44d39c5518a1b673c81096aad732f4415",
"source_line_end": 86,
"source_line_start": 70
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depends on",
"url": "paper-module-tactic-portfolio-availability.html"
}
],
"relation": "depends on"
},
{
"count": 3,
"links": [
{
"id": "corpus_readiness_mathlib_absence_gate",
"kind": "organ",
"kind_label": "component",
"label": "Corpus Readiness Mathlib Absence Gate",
"relation": "explains",
"url": "component-corpus-readiness-mathlib-absence-gate.html"
},
{
"id": "mechanism.corpus_readiness_mathlib_absence_gate.validates_public_corpus_readiness_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public corpus readiness boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-corpus-readiness-mathlib-absence-gate-validates-public-corpus-readiness-boundary"
},
{
"id": "mechanism.corpus_readiness_mathlib_absence_gate.validates_public_mathlib_absence_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public mathlib absence boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-corpus-readiness-mathlib-absence-gate-validates-public-mathlib-absence-boundary"
}
],
"relation": "explains"
},
{
"count": 2,
"links": [
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.corpus_readiness_mathlib_absence_gate",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 238,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/corpus_readiness_mathlib_absence.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Lean mathematical library",
"section_id": "grounded-in-lean-corpus-and-theorem-proving-work",
"url": "https://arxiv.org/abs/1910.09336"
},
{
"kind": "external_reference",
"label": "miniF2F",
"section_id": "grounded-in-lean-corpus-and-theorem-proving-work",
"url": "https://arxiv.org/abs/2109.00110"
},
{
"kind": "external_reference",
"label": "LeanDojo",
"section_id": "grounded-in-lean-corpus-and-theorem-proving-work",
"url": "https://arxiv.org/abs/2306.15626"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**From a recorded Lean readiness report it answers, per corpus, whether the library is present and whether the Mathlib import lane resolves, pairing copied rows with a live probe.** corpusreadinessmathlibabsencegate reads a recorded readiness report from a Lean math toolchain run and turns one question into an inspectable answer: for each math corpus, is it actually present on this host, and is the Mathlib import lane actually available, or not? It carries copied corpus and toolchain rows from a 2026-05-11 proof-state curriculum smoke run and pairs them with a live import check.",
"level": 4,
"order": 1,
"section_id": "from-recorded-lean",
"source_line_end": 14,
"source_line_start": 3,
"title": "From recorded lean"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It emits a metadata-only readiness board and consumer verdicts: which corpora are absent, which are smoke-test only, and which tasks stay blocked until a Mathlib probe passes.** The output is a readiness board and a set of consumer verdicts. It records which corpora are absent, which are usable only for translation smoke tests, and which downstream tasks stay blocked until a passing Mathlib probe exists. Everything it emits is metadata-only: no proof text, no provider output, no non-public paths.",
"level": 4,
"order": 2,
"section_id": "emits-body-free-readiness",
"source_line_end": 24,
"source_line_start": 15,
"title": "Emits metadata-only readiness"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is not a prover or Lake build but a gate that forces a formal-math route to expose corpus readiness before any proof route counts as usable.** It is not a theorem prover, a Lake build, or a Mathlib availability claim. It is a gate that forces a formal-math route to expose corpus readiness before any proof route is treated as usable.",
"level": 4,
"order": 3,
"section_id": "is-not-prover",
"source_line_end": 32,
"source_line_start": 25,
"title": "Is not prover"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 34,
"source_line_start": 33,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Formal-math agents conflate a corpus appearing on a roster with it being usable for their proof route, so a missing library surfaces only later as a broken import.** Formal-math agents fail in one specific way. They treat \"there is a corpus\" as if it meant \"this corpus is usable for the proof route I am about to take\". A roster lists miniF2F, PutnamBench, ProofNet, LeanDojo and Mathlib, the agent assumes the libraries are present, and the failure only surfaces later as a broken import or a tactic that needs a premise the host cannot resolve.",
"level": 4,
"order": 5,
"section_id": "roster-presence-mistaken-for-usability",
"source_line_end": 45,
"source_line_start": 35,
"title": "Roster presence mistaken for usability"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The loose check trusts a mathlibavailable field or PASS status; this gate refuses that drift, flagging any row whose alias claim a live probe does not confirm.** The loose version of this check trusts the roster. It reads a mathlib_available field, or a PASS lean status, and marks the corpus ready. That is the drift the gate is built to stop. A stale alias field cannot turn the gate green on its own. It has to agree with a live probe or the row is flagged.",
"level": 4,
"order": 6,
"section_id": "alias-field-must-agree-with-a-live-probe",
"source_line_end": 55,
"source_line_start": 46,
"title": "Alias field must agree with a live probe"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 57,
"source_line_start": 56,
"title": "How it works"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "run and runprojectionbundle both call build_result, which loads two inputs, scans paths against the exclusion policy, then folds four ordered checks into one result and board.** The two entrypoints run and runprojectionbundle both call buildresult. run handles the first-wave fixture with its negative cases. runprojectionbundle handles an exported bundle without negative cases and requires a source-module manifest. buildresult loads the two input files, scans every input path against the non-public-state exclusion policy, then runs four checks in order and folds them into one result and a readiness board.",
"level": 4,
"order": 8,
"section_id": "two-entrypoints-funnel-into-build-result",
"source_line_end": 87,
"source_line_start": 58,
"title": "Two entrypoints funnel into `_build_result`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateruntimesourceartifacts digest-checks four source refs and runs a probe that expects Std to compile and Mathlib to be rejected, setting mathliblakeprojectimportavailable false when the lane fails.** validateruntimesource_artifacts anchors the run to four source refs: the corpus readiness rows, a tactic-affordance probe, the Mathlib import probe Lean file, and a tactic portfolio availability file. It checks each file's SHA-256 against a recorded digest, parses the JSON ones, and calls runtimeleanimport_probe. That probe writes two small Lean files into a temporary directory. One imports Std and is expected to compile. One imports Mathlib and is expected to be rejected with the toolcha\u2026",
"level": 4,
"order": 9,
"section_id": "the-live-import-probe-std-compiles-mathlib-rejected",
"source_line_end": 104,
"source_line_start": 88,
"title": "The live import probe: Std compiles, Mathlib rejected"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatecorpus_readiness marks a corpus usable only when it exists, carries a Lake file, and the probe agrees; contradicting alias claims are flagged and every unavailable lane is blocked.** validatecorpusreadiness normalises the seven corpus rows against that runtime evidence. A corpus is marked usable for Mathlib-dependent work only when the runtime row says it exists, carries a Lake file, and the Mathlib import lane is available. If a row claims mathliblakeprojectimportavailable or any present alias in MATHLIBPRESENTALIAS_FIELDS while the probe disagrees, the row is flagged with CORPUSREADINESSRUNTIMEPROBECONTRADICTION. Absent corpora and every unavailable Mathlib lane land in blocked_ca\u2026",
"level": 4,
"order": 10,
"section_id": "corpus-rows-normalised-against-probe-evidence",
"source_line_end": 118,
"source_line_start": 105,
"title": "Corpus rows normalised against probe evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateconsumergatecases recomputes each verdict from readiness facts, not the case's own label, blocking cases needing an unavailable Mathlib lane or an absent corpus; only the miniF2F smoke passes.** validateconsumergate_cases derives each verdict from the readiness facts instead of trusting a case's own expected-decision label. A case is blocked if it requires the Mathlib lake project import while that lane is unavailable, or if its target corpus is absent. Of the seven cases, one passes: the miniF2F Lean3 translation smoke, which does not require a Mathlib project. The other six stay blocked. A case that never set readinessgatechecked is flagged with CONSUMERREADINESSGATE_UNCHECKED.",
"level": 4,
"order": 11,
"section_id": "consumer-verdicts-derived-from-readiness-facts",
"source_line_end": 131,
"source_line_start": 119,
"title": "Consumer verdicts derived from readiness facts"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatesourcemoduleimports enforces copied-body digests and byte equality, and buildresult passes only when no negative case is missing, the secret scan is clean, and the source floor holds.** validatesourcemodule_imports adds the exported-bundle floor. It requires the manifest to declare copiednonsecretmacrobody, each row to use a body material class, each copied file to match its recorded source and target digest, and no row to carry a body inside a result record. Exact-copy rows must preserve byte equality; rewrite rows must match the declared private-absolute-path rewrite recipe exactly. buildresult then computes one status: it passes only when no expected negative case is missing, the\u2026",
"level": 4,
"order": 12,
"section_id": "exported-bundle-source-floor-and-final-status",
"source_line_end": 148,
"source_line_start": 132,
"title": "Exported-bundle source floor and final status"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 150,
"source_line_start": 149,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**The fixture carries five named negative inputs, declared in EXPECTEDNEGATIVECASES.** Each one must be observed with its error code or the run is blocked: mathlibavailablewithoutprobe must raise MATHLIBAVAILABILITY_OVERCLAIM when Mathlib is claimed available with no passing probe. consumerskipsreadiness_gate must raise CONSUMERSKIPSCORPUSREADINESSGATE when a consumer attempts Mathlib-dependent work without checking readiness. privatecorpussourceref must raise PRIVATECORPUSSOURCEREF_FORBIDDEN when a source ref points at a non-public paths. proofbodyleakage must raise CORPUSREADINESSPROOFBODYFORBIDDEN when a corpus row carries a proof or provider body field. releaseoverclaim must raise CORPU\u2026",
"level": 4,
"order": 14,
"section_id": "fixture-carries-five",
"source_line_end": 168,
"source_line_start": 151,
"title": "Fixture carries five"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 170,
"source_line_start": 169,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The design draws on Lean corpus and neural theorem-proving work where library availability and benchmark splits are part of the claim, requiring probes and digests before tactic-routing language is allowed.** This component sits in Lean corpus and neural theorem-proving work where library availability, premise access, and benchmark splits are part of the claim. The Lean mathematical library establishes Mathlib as a large community-maintained formal mathematics corpus, miniF2F gives a cross-system benchmark for formal Olympiad statements, and LeanDojo shows why reproducible corpus extraction and accessible-premise metadata matter for theorem-proving agents. The gate borrows the discipline:\u2026",
"level": 4,
"order": 16,
"section_id": "grounded-in-lean-corpus-and-theorem-proving-work",
"source_line_end": 187,
"source_line_start": 171,
"title": "Grounded in Lean corpus and theorem-proving work"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Reader-verifiable commands, run from the microcosm-substrate/ public root:",
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 202,
"source_line_start": 188,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means fixture and exported bundle agree on readiness, Mathlib absence, blocked consumers, digests, and the five negative cases; it reruns no Lean, proves nothing, and installs no Mathlib.** A pass means the fixture and the exported bundle agree about corpus readiness, Mathlib absence, blocked consumers, copied source digests, metadata-only result files, and the five negative cases. It does not rerun Lean or Lake, prove any theorem, or install Mathlib.",
"level": 4,
"order": 18,
"section_id": "what-a-pass-certifies-and-excludes",
"source_line_end": 212,
"source_line_start": 203,
"title": "What a pass certifies and excludes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "claim-ceiling",
"source_line_end": 214,
"source_line_start": 213,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is narrow: a fixture and exported bundle agree on readiness, Mathlib absence, verdicts, digests, and leakage guards, bounded to projection over copied material plus one import probe.** The strongest claim the evidence supports is narrow: a public fixture and an exported bundle agree about corpus readiness, Mathlib absence, blocked consumer verdicts, copied source-module digests, metadata-only result files, and the negative leakage guards. The proof boundary is algorithmic projection over copied corpus and toolchain material plus one bounded import probe.",
"level": 4,
"order": 20,
"section_id": "the-strongest-supported-claim",
"source_line_end": 225,
"source_line_start": 215,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "AUTHORITYCEILING sets every outward flag false, so the component proves no install, runs no Lake build, validates no theorem, calls no provider, and approves no launch.** It refuses everything past that line. The scope limit in scope_limit sets mathliblakeprojectimportauthorized, formalproofauthority, providercallsauthorized, and release_authorized all to false. This component does not establish Mathlib is installed, run a Lake build, validate formal-result correctness, benchmark corpus quality, call a provider, expose any proof or provider body, change source files, or approve launch.",
"level": 4,
"order": 21,
"section_id": "what-stays-refused-past-the-line",
"source_line_end": 237,
"source_line_start": 226,
"title": "What stays refused past the line"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/corpus_readiness_mathlib_absence.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The corpus readiness Mathlib absence gate validates copied corpus/toolchain readiness bodies before any Mathlib-dependent proof or retrieval claim is allowed."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "corpus_readiness_mathlib_absence_gate.py",
"path": "src/microcosm_core/organs/corpus_readiness_mathlib_absence_gate.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for corpus readiness fixture validation, exported bundle validation, source-module manifest checks, negative-case enforcement, command-card compression, scope limits, and metadata-only result record writing.",
"symbols": [
"run",
"run_projection_bundle",
"validate_corpus_readiness",
"validate_consumer_gate_cases",
"validate_source_module_imports",
"_build_result",
"write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/corpus_readiness_mathlib_absence_gate.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/corpus_readiness_mathlib_absence.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/corpus_readiness_mathlib_absence.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Corpus Readiness Mathlib Absence Gate is the public formal-math corpus readiness boundary for Microcosm. It checks copied PROVER smoke-run corpus readiness rows, Lean/Std toolchain probe rows, Mathlib absence status, consumer-gate decisions, absent corpus blocking, source-module manifest digests, metadata-only result record policy, and five negative cases while keeping proof bodies, model-output data, non-public source refs, benchmark-completeness claims, launch-scope decision, and Mathlib proof authority out of result records.",
"summary_line": "The corpus readiness Mathlib absence gate validates copied corpus/toolchain readiness bodies before any Mathlib-dependent proof or retrieval claim is allowed.",
"title": "Corpus Readiness Mathlib Absence Gate"
},
{
"anchor": "paper-module-pattern-binding-contract",
"authority_ceiling": "Public pattern-binding fixtures, exported system bundles, route-readiness selector bundles, source-bundle refs, and metadata-only result records only; no private pattern-ledger certification, launch-scope decision, external model access, hosted-public posture, recipient work, non-public data equivalence, source-file changes, or whole-system correctness.",
"body_html": "
Teleology
\n
Binding pattern rows to source-backed evidence
\n
pattern_binding_contract binds each pattern row to a source-available bundle, public runtime ref, authority handle, scope boundary, and secret-exclusion result record, with synthetic rows allowed only as controls.pattern_binding_contract is the public root component that binds pattern rows to source-available source bundles, public runtime refs, authority-chain handles, scope boundaries, and secret-exclusion result records. Synthetic rows are allowed only as regression controls or negative cases; they are not product evidence.
\n
Purpose
\n
Why a lone pattern row cannot self-publish
\n
A mined pattern row reads like a finished insight, so this component exists to decide whether it can be admitted to the public surface and under exactly what evidence and ceiling. A mined engineering pattern is a tempting thing to publish on its own. It reads like a self-contained insight, so it is easy to lift a single row out of a private ledger and present it as a finished public claim. This component exists to stop that. It answers one question: can a given pattern row be admitted to the public surface, and if so, under exactly what evidence and what ceiling?
\n
Binding, not display
\n
The check binds rather than displays: every row must name a source bundle, governing standard, and scope boundary, while duplicates, overclaims, and standalone-leaf rows are rejected and negative cases run alongside. The check is binding rather than display. Every pattern row must name a source bundle that points at a real public runtime ref or regression-harness ref, a governing standard, and an scope boundary. A row that lacks any of these, duplicates another row's id, or claims to be a standalone public leaf is rejected. The same validator runs deliberate negative cases alongside the positive control. The result record shows that good rows pass and that each known failure mode is still caught.
\n
Truth accounting and route-readiness
\n
Truth accounting separates runtime-metadata rows from real pattern-ledger imports so accepted-row counts are not read as system progress, and route-readiness keeps selection component-first and fixture-bound. The less obvious idea is truth accounting. When an exported bundle is validated, the component separates rows that merely describe runtime metadata from rows that represent a real pattern-ledger import, and records that a high accepted-row count is not the same as system progress. This guards against the quiet inflation where counting accepted rows starts to read like a measure of how much real work has landed. The route-readiness layer closes the matching gap on the selector side: a row can look selectable in isolation, but it is only admitted through the component that owns it, its fixture contract, and a gate that refuses to let hard no-standalone rows appear as selectable targets.
\n
Public contract
\n
The four validators and their result records
\n
validate_pattern_bindings, validate_source_capsules, validate_reference_capsules, and the authority-chain and ceiling checks run over one fixture and write command-owned result records.validate_pattern_bindings checks required binding fields, rejects duplicate pattern ids, and flags a row that overclaims source authority or public-leaf readiness. validate_source_capsules and validate_reference_capsules check that each bundle names a runtime ref and keeps its body out of the result record. validate_authority_chain_handle_resolver rejects an unsupported handle that still implies authority, and validate_authority_ceiling records whether the two overclaim negative cases were actually observed. validate runs these checks over one fixture and writes command-owned result records under receipts/first_wave/pattern_binding_contract/.
\n
Route-readiness closes the selector gap
\n
The exported bundle carries the route-readiness overlays as public bodies, and validate_substrate_bundle recomputes the four checks then validates the selector so a row cannot look selectable without opening its owning component. The exported system bundle also carries the source route-readiness selector overlays as public source-open bodies: examples/pattern_binding_contract/exported_route_readiness_bundle/. validate_substrate_bundle recomputes the same four checks over the imported pattern ledger, then calls validate_route_readiness_bundle against the route-readiness audit, row-to-component router, route cards, fixture specs, decision matrix, dependency DAG, internal routing graph, and copied source validation report. This closes the old gap where a mined pattern row could look selectable without opening the component bundle that owns it.
\n
Which command answers selector admission
\n
For selector admission rather than generic binding, cold readers run microcosm pattern-route-readiness validate-bundle, with the older validate-route-readiness-bundle action kept as a compatibility route. Cold readers should use microcosm pattern-route-readiness validate-bundle against examples/pattern_binding_contract/exported_route_readiness_bundle/ when the question is selector admission rather than generic pattern binding. The older pattern-binding validate-route-readiness-bundle action remains a compatibility route to the same validator.
\n
Shape
\nDiagram of the mechanism (13 steps).Diagram source & refs
flowchart LR subgraph Inputs["Pattern-binding inputs"] Patterns["Pattern rows id, governing standard, scope boundary, source refs, projection posture"] Bundles["Source bundles metadata-only refs to public runtime or regression harness"] Handles["Authority-chain handles resolver result records"] end Validator["pattern_binding_contract required fields, duplicate ids, bundle resolution, secret-exclusion scan"] subgraph Negative["Refusal floor"] Dup["Duplicate id rejected"] Leak["Private body leak rejected"] Overclaim["Public-leaf overclaim rejected"] Unsupported["Unsupported authority handle not upgraded"] end subgraph Bundle["Exported-bundle path"] Truth["Truth accounting runtime-metadata rows vs real pattern-ledger import"] RouteReadiness["Route-readiness selector component-first admission, fixture contract, hard no-standalone gate"] end Result records["Result records refs, digests, counts, verdicts; body text omitted"] Patterns --> Validator Bundles --> Validator Handles --> Validator Validator --> Negative Validator --> Bundle Truth --> RouteReadiness Negative --> Result records Bundle --> Result records
\n
Evidence binding
\n
Accepted component row: core/organ_registry.json::implemented_organs[pattern_binding_contract]. Accepted component row: core/organ_registry.json::implemented_organs[pattern_binding_contract]. Evidence class: core/organ_evidence_classes.json::organ_evidence_classes[pattern_binding_contract] with rank 5 semantic-validator authority. The runtime locus is src/microcosm_core/organs/pattern_binding_contract.py, with focused coverage in tests/test_pattern_binding_contract.py.
\n
Paper bundle authority: core/paper_module_capsules.json#paper_module.pattern_binding_contract. Paper bundle authority: core/paper_module_capsules.json#paper_module.pattern_binding_contract. Mechanism source: core/mechanism_sources.json#mechanism.pattern_binding_contract.validates_public_pattern_bindings.
\n
Reader evidence routing
\n
Start with paper_modules/pattern_binding_contract.json for the bundle payload, then open standards/std_microcosm_pattern_binding_contract.json to check the required fields, public/private boundary, source-open body import floor, route-readiness rules, and result record expectations. Start with paper_modules/pattern_binding_contract.json for the bundle payload, then open standards/std_microcosm_pattern_binding_contract.json to check the required fields, public/private boundary, source-open body import floor, route-readiness rules, and result record expectations.
\n
Use core/fixture_manifests/pattern_binding_contract.fixture_manifest.json before inspecting fixtures or exported bundles. The manifest and the source_module_manifest.json files name the copied source body floor; result record payloads carry source refs, digests, anchors, counts, verdicts, and omission result records rather than inlining body text. Use core/fixture_manifests/pattern_binding_contract.fixture_manifest.json before inspecting fixtures or exported bundles. The manifest and the source_module_manifest.json files name the copied source body floor; result record payloads carry source refs, digests, anchors, counts, verdicts, and omission result records rather than inlining body text.
\n
Prior Art Grounding
\n
Pattern-language lineage
\n
The component follows the software pattern-language tradition of making reusable engineering structures explicit and reviewable, with the Hillside patterns library as its direct prior-art family. This component follows the software pattern-language tradition of making reusable engineering structures explicit, named, and reviewable. The Hillside patterns library is the direct prior-art family for treating patterns as shared vocabulary rather than loose implementation notes.
\n
Provenance and supply-chain attestation
\n
The binding layer borrows provenance and supply-chain attestation shapes from W3C PROV, SLSA, and in-toto, applied to pattern rows and route-readiness selectors rather than launch certification. The binding layer also borrows from provenance and supply-chain attestation patterns. W3C PROV motivates the source/ref/evidence relation shape, while SLSA and in-toto motivate digest-bound artifact claims and step-level metadata. Microcosm applies those ideas to pattern rows and route-readiness selectors, not to launch certification.
\n
Re-entry condition: if copied source bodies, route-readiness overlays, or negative-case rules change, rerun the validate, validate-system-bundle, and validate-route-readiness-bundle commands under Validation Result record Path and update this paper module plus standards/std_microcosm_pattern_binding_contract.json from the new result record fields. Re-entry condition: if copied source bodies, route-readiness overlays, or negative-case rules change, rerun the validate, validate-system-bundle, and validate-route-readiness-bundle commands under Validation Result record Path and update this paper module plus standards/std_microcosm_pattern_binding_contract.json from the new result record fields.
\n
Validation Result record Path
\n
From microcosm-substrate/, reproduce this page's proof boundary with temporary result records:
\n
What a pass shows
\n
A pass means the fixture, system-bundle, and route-readiness result records all show status: pass with metadata-only authority handles, with Scope limit stating what it does not certify. A pass means the fixture, system-bundle, and route-readiness result records all show status: pass with metadata-only authority handles; see Scope limit for what a pass does and does not certify.
\n
Inspect result record fields, not markdown constants
\n
The current authority is the runtime result record set, and cold readers should read result record fields such as status and secret_exclusion_scan rather than markdown constants or a separate sign-off result record. The current authority is the runtime result record set under receipts/first_wave/pattern_binding_contract/; do not cite a separate pattern-specific sign-off result record unless an sign-off-lane artifact is actually present. Cold readers should inspect result record fields rather than markdown constants: status, secret_exclusion_scan, source_open_body_imports, truth_accounting, route_readiness_summary, selection_contract, and source_manifest.
\n
Scope boundary
\n
Scope limit
\n
What this module is evidence for
\n
The module covers public pattern-binding mechanics and stands as evidence for the pattern_binding_contract component and its public-binding mechanism. This module covers public pattern-binding mechanics: source-bundle validation, reference-bundle validation, authority-handle validation, route-readiness selector admission, duplicate and unknown-ref rejection, private-leakage sentinel checks, and metadata-only result record shape. It is evidence for the pattern_binding_contract component and mechanism.pattern_binding_contract.validates_public_pattern_bindings.
\n
Where the ceiling stops
\n
The ceiling stops before private ledger authority, launch or deployment posture, standalone-leaf status, and provider or source-file changes; importing the bundle keeps selection component-first. The ceiling stops before private pattern-ledger authority, hosted or public launch-scope decision, deployment posture, standalone public-leaf selector status, non-public data equivalence, external model access, recipient work, source-file changes, publishing-scope decision, or whole-system correctness. Importing the route-readiness bundle does not make any mined pattern row a standalone public leaf; selection stays component-first and fixture-bound.
\n\nSource and projection details\n
Source-open body floor
\n
The source-open body floor is the imported public bundle, not the private pattern ledger. Cold readers can open examples/pattern_binding_contract/exported_substrate_bundle/ and examples/pattern_binding_contract/exported_route_readiness_bundle/ to inspect the copied source module manifests, source bundles, reference bundles, authority-chain handles, route-readiness overlays, selector contract inputs, and copied source validation report. The source-open body floor is the imported public bundle, not the private pattern ledger. Cold readers can open examples/pattern_binding_contract/exported_substrate_bundle/ and examples/pattern_binding_contract/exported_route_readiness_bundle/ to inspect the copied source module manifests, source bundles, reference bundles, authority-chain handles, route-readiness overlays, selector contract inputs, and copied source validation report. The required body floor is named by each source_module_manifest.json plus source_capsules.json, reference_capsules.json, and authority_chain_handles.json.
\n
Result records and manifests must stay metadata-only where the standard requires it: they carry refs, digests, anchors, counts, verdicts, omission result records, and secret-exclusion results. Result records and manifests must stay metadata-only where the standard requires it: they carry refs, digests, anchors, counts, verdicts, omission result records, and secret-exclusion results. They do not inline private source bodies, raw operator payloads, model-output data, recipient data, or hidden pattern-ledger material.
\n",
"components": [
{
"anchor": "component-pattern_binding_contract",
"component_id": "pattern_binding_contract",
"public_label": "Pattern Binding Contract"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 2,
"organ": 1,
"paper_module": 3,
"principle": 9
},
"link_count": 22,
"links": [
{
"id": "pattern_binding_contract",
"kind": "organ",
"kind_label": "component",
"label": "Pattern Binding Contract",
"relation": "explains",
"url": "component-pattern-binding-contract.html"
},
{
"id": "mechanism.pattern_binding_contract.validates_public_pattern_bindings",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public pattern bindings",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-pattern-binding-contract-validates-public-pattern-bindings"
},
{
"id": "mechanism.cold_clone_probe.validates_public_source_root_bootstrap",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public source root bootstrap",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-cold-clone-probe-validates-public-source-root-bootstrap"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[2:paper_module.agent_route_observability_runtime].depends_on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.cold_clone_probe",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold Clone Probe",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[89:paper_module.cold_clone_probe].depends_on",
"url": "paper-module-cold-clone-probe.html"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[13:paper_module.cold_reader_route_map].depends_on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[19:paper_module.macro_projection_import_protocol].depends_on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[1:paper_module.navigation_hologram_route_plane].depends_on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.pattern_assimilation",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Assimilation",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[71:paper_module.pattern_assimilation].depends_on",
"url": "paper-module-pattern-assimilation.html"
},
{
"id": "paper_module.voice_to_doctrine_self_improvement_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Voice to Doctrine Self-Improvement Loop",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[56:paper_module.voice_to_doctrine_self_improvement_loop].depends_on",
"url": "paper-module-voice-to-doctrine-self-improvement-loop.html"
}
],
"paper_module_id": "paper_module.pattern_binding_contract",
"reader_packet": {
"abstract": "The public pattern-binding component validates pattern rows, source bundles, authority handles, exported system bundles, and route-readiness selector overlays while keeping mined rows component-first and fixture-bound.",
"authority_ceiling": "Public pattern-binding fixtures, exported system bundles, route-readiness selector bundles, source-bundle refs, and metadata-only result records only; no private pattern-ledger certification, launch-scope decision, external model access, hosted-public posture, recipient work, non-public data equivalence, source-file changes, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.pattern_binding_contract.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-pattern-binding-contract-diagram-1",
"html_url": "docs/paper-module-pattern-binding-contract.html#paper-module-pattern-binding-contract-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart LR subgraph Inputs[\"Pattern-binding inputs\"] Patterns[\"Pattern rows id, governing standard, scope boundary, source refs, projection posture\"] Bundles[\"Source bundles metadata-only refs to public runtime or regression harness\"] Handles[\"Authority-chain handles resolver result records\"] end Validator[\"pattern_binding_contract required fields, duplicate ids, bundle resolution, secret-exclusion scan\"] subgraph Negative[\"Refusal floor\"] Dup[\"Duplicate id rejected\"] Leak[\"Private body leak rejected\"] Overclaim[\"Public-leaf overclaim rejected\"] Unsupported[\"Unsupported authority handle not upgraded\"] end subgraph Bundle[\"Exported-bundle path\"] Truth[\"Truth accounting runtime-metadata rows vs real pattern-ledger import\"] RouteReadiness[\"Route-readiness selector component-first admission, fixture contract, hard no-standalone gate\"] end Result records[\"Result records refs, digests, counts, verdicts; body text omitted\"] Patterns --> Validator Bundles --> Validator Handles --> Validator Validator --> Negative Validator --> Bundle Truth --> RouteReadiness Negative --> Result records Bundle --> Result records",
"source_format": "mermaid",
"source_hash": "sha256:7ed1be41b0ac1acdc65e8f8646ce0f608773a1ec62c55f254e07027f6c399f80",
"source_line_end": 133,
"source_line_start": 101
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
}
],
"relation": "depends on"
},
{
"count": 3,
"links": [
{
"id": "pattern_binding_contract",
"kind": "organ",
"kind_label": "component",
"label": "Pattern Binding Contract",
"relation": "explains",
"url": "component-pattern-binding-contract.html"
},
{
"id": "mechanism.pattern_binding_contract.validates_public_pattern_bindings",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public pattern bindings",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-pattern-binding-contract-validates-public-pattern-bindings"
},
{
"id": "mechanism.cold_clone_probe.validates_public_source_root_bootstrap",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public source root bootstrap",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-cold-clone-probe-validates-public-source-root-bootstrap"
}
],
"relation": "explains"
},
{
"count": 10,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.pattern_binding_contract",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 205,
"title": "JSON Bundle Binding"
},
{
"reason": "public_projection_internal_section_denylist",
"section_id": "structured-lattice-bindings",
"source_line_start": 220,
"title": "Structured lattice bindings"
},
{
"reason": "public_projection_internal_section_denylist",
"section_id": "receipt-expectations",
"source_line_start": 292,
"title": "Result record expectations"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/pattern_binding_contract.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Hillside patterns library",
"section_id": "pattern-language-lineage",
"url": "https://hillside.net/patterns/"
},
{
"kind": "external_reference",
"label": "W3C PROV",
"section_id": "provenance-and-supply-chain-attestation",
"url": "https://www.w3.org/TR/prov-overview/"
},
{
"kind": "external_reference",
"label": "SLSA",
"section_id": "provenance-and-supply-chain-attestation",
"url": "https://slsa.dev/spec/"
},
{
"kind": "external_reference",
"label": "in-toto",
"section_id": "provenance-and-supply-chain-attestation",
"url": "https://in-toto.io/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 1,
"section_id": "teleology",
"source_line_end": 4,
"source_line_start": 3,
"title": "Teleology"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "patternbinding_contract binds each pattern row to a source-available bundle, public runtime ref, authority handle, scope boundary, and secret-exclusion result record, with synthetic rows allowed only as controls.** patternbindingcontract is the public root component that binds pattern rows to source-available source bundles, public runtime refs, authority-chain handles, scope boundaries, and secret-exclusion result records. Synthetic rows are allowed only as regression controls or negative cases; they are not product evidence.",
"level": 4,
"order": 2,
"section_id": "binding-pattern-rows-to-source-backed-evidence",
"source_line_end": 14,
"source_line_start": 5,
"title": "Binding pattern rows to source-backed evidence"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 3,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A mined pattern row reads like a finished insight, so this component exists to decide whether it can be admitted to the public surface and under exactly what evidence and ceiling.** A mined engineering pattern is a tempting thing to publish on its own. It reads like a self-contained insight, so it is easy to lift a single row out of a private ledger and present it as a finished public claim. This component exists to stop that. It answers one question: can a given pattern row be admitted to the public surface, and if so, under exactly what evidence and what ceiling?",
"level": 4,
"order": 4,
"section_id": "why-a-lone-pattern-row-cannot-self-publish",
"source_line_end": 27,
"source_line_start": 17,
"title": "Why a lone pattern row cannot self-publish"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The check binds rather than displays: every row must name a source bundle, governing standard, and scope boundary, while duplicates, overclaims, and standalone-leaf rows are rejected and negative cases run alongside.** The check is binding rather than display. Every pattern row must name a source bundle that points at a real public runtime ref or regression-harness ref, a governing standard, and an scope boundary. A row that lacks any of these, duplicates another row's id, or claims to be a standalone public leaf is rejected. The same validator runs deliberate negative cases alongside the positive control. The result record shows that good rows pass and that each known failure mode is sti\u2026",
"level": 4,
"order": 5,
"section_id": "binding-not-display",
"source_line_end": 40,
"source_line_start": 28,
"title": "Binding, not display"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Truth accounting separates runtime-metadata rows from real pattern-ledger imports so accepted-row counts are not read as system progress, and route-readiness keeps selection component-first and fixture-bound.** The less obvious idea is truth accounting. When an exported bundle is validated, the component separates rows that merely describe runtime metadata from rows that represent a real pattern-ledger import, and records that a high accepted-row count is not the same as system progress. This guards against the quiet inflation where counting accepted rows starts to read like a measure of how much real work has landed. The route-readiness layer closes the matching gap on the selector side:\u2026",
"level": 4,
"order": 6,
"section_id": "truth-accounting-and-route-readiness",
"source_line_end": 55,
"source_line_start": 41,
"title": "Truth accounting and route-readiness"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "public-contract",
"source_line_end": 57,
"source_line_start": 56,
"title": "Public contract"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatepatternbindings, validatesource_capsules, validatereferencecapsules, and the authority-chain and ceiling checks run over one fixture and write command-owned result records.** validatepatternbindings checks required binding fields, rejects duplicate pattern ids, and flags a row that overclaims source authority or public-leaf readiness. validatesourcecapsules and validatereferencecapsules check that each bundle names a runtime ref and keeps its body out of the result record. validateauthoritychainhandleresolver rejects an unsupported handle that still implies authority, and validateauthorityceiling records whether the two overclaim negative cases were actually observed. validate runs\u2026",
"level": 4,
"order": 8,
"section_id": "the-four-validators-and-their-receipts",
"source_line_end": 72,
"source_line_start": 58,
"title": "The four validators and their result records"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The exported bundle carries the route-readiness overlays as public bodies, and validatesubstratebundle recomputes the four checks then validates the selector so a row cannot look selectable without opening its owning component.** The exported system bundle also carries the source route-readiness selector overlays as public source-open bodies: examples/patternbindingcontract/exportedroutereadiness_bundle/. validatesubstratebundle recomputes the same four checks over the imported pattern ledger, then calls validateroutereadiness_bundle against the route-readiness audit, row-to-component router, route cards, fixture specs, decision matrix, dependency DAG, internal routing graph, and copied s\u2026",
"level": 4,
"order": 9,
"section_id": "route-readiness-closes-the-selector-gap",
"source_line_end": 87,
"source_line_start": 73,
"title": "Route-readiness closes the selector gap"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**For selector admission rather than generic binding, cold readers run microcosm pattern-route-readiness validate-bundle, with the older validate-route-readiness-bundle action kept as a compatibility route.** Cold readers should use microcosm pattern-route-readiness validate-bundle against examples/patternbindingcontract/exportedroutereadiness_bundle/ when the question is selector admission rather than generic pattern binding. The older pattern-binding validate-route-readiness-bundle action remains a compatibility route to the same validator.",
"level": 4,
"order": 10,
"section_id": "which-command-answers-selector-admission",
"source_line_end": 98,
"source_line_start": 88,
"title": "Which command answers selector admission"
},
{
"block_kinds": [
"diagram"
],
"excerpt": null,
"level": 2,
"order": 11,
"section_id": "shape",
"source_line_end": 134,
"source_line_start": 99,
"title": "Shape"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "source-open-body-floor",
"source_line_end": 136,
"source_line_start": 135,
"title": "Source-open body floor"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The source-open body floor is the imported public bundle, not the private pattern ledger. Cold readers can open examples/patternbindingcontract/exportedsubstratebundle/ and examples/patternbindingcontract/exportedroutereadinessbundle/ to inspect the copied source module manifests, source bundles, reference bundles, authority-chain handles, route-readiness overlays, selector contract inputs, and copied source validation report.** The source-open body floor is the imported public bundle, not the private pattern ledger. Cold readers can open examples/patternbindingcontract/exportedsubstratebundle/ and examples/patternbindingcontract/exportedroutereadiness_bundle/ to inspect the copied source m\u2026",
"level": 4,
"order": 13,
"section_id": "the-source-open-body-floor-is-the-imported-public-bundle-not-the-private-pattern-ledger",
"source_line_end": 149,
"source_line_start": 137,
"title": "The source-open body floor is the imported public bundle, not the private pattern ledger"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Result records and manifests must stay metadata-only where the standard requires it: they carry refs, digests, anchors, counts, verdicts, omission result records, and secret-exclusion results. Result records and manifests must stay metadata-only where the standard requires it: they carry refs, digests, anchors, counts, verdicts, omission result records, and secret-exclusion results. They do not inline private source bodies, raw operator payloads, model-output data, recipient data, or hidden pattern-ledger material.",
"level": 4,
"order": 14,
"section_id": "receipts-and-manifests-must-stay-body-free-where-the-standard-requires-it",
"source_line_end": 158,
"source_line_start": 150,
"title": "Result records and manifests must stay metadata-only where the standard requires it"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "claim-ceiling",
"source_line_end": 160,
"source_line_start": 159,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The module covers public pattern-binding mechanics and stands as evidence for the patternbindingcontract component and its public-binding mechanism.** This module covers public pattern-binding mechanics: source-bundle validation, reference-bundle validation, authority-handle validation, route-readiness selector admission, duplicate and unknown-ref rejection, private-leakage sentinel checks, and metadata-only result record shape. It is evidence for the patternbindingcontract component and mechanism.patternbindingcontract.validatespublicpattern_bindings.",
"level": 4,
"order": 16,
"section_id": "what-this-module-is-evidence-for",
"source_line_end": 171,
"source_line_start": 161,
"title": "What this module is evidence for"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The ceiling stops before private ledger authority, launch or deployment posture, standalone-leaf status, and provider or source-file changes; importing the bundle keeps selection component-first.** The ceiling stops before private pattern-ledger authority, hosted or public launch-scope decision, deployment posture, standalone public-leaf selector status, non-public data equivalence, external model access, recipient work, source-file changes, public sharing decision, or whole-system correctness. Importing the route-readiness bundle does not make any mined pattern row a standalone public leaf; selection stays component-first and fixture-bound.",
"level": 4,
"order": 17,
"section_id": "where-the-ceiling-stops",
"source_line_end": 183,
"source_line_start": 172,
"title": "Where the ceiling stops"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "evidence-binding",
"source_line_end": 185,
"source_line_start": 184,
"title": "Evidence binding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Accepted component row: core/organregistry.json::implementedorgans[patternbindingcontract]. Accepted component row: core/organregistry.json::implementedorgans[patternbindingcontract]. Evidence class: core/organevidenceclasses.json::organevidenceclasses[patternbindingcontract] with rank 5 semantic-validator authority. The runtime locus is src/microcosmcore/components/patternbinding_contract.py, with focused coverage in tests/testpatternbinding_contract.py.",
"level": 4,
"order": 19,
"section_id": "accepted-organ-row",
"source_line_end": 196,
"source_line_start": 186,
"title": "Accepted component row"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Paper bundle authority: core/papermodulecapsules.json#papermodule.patternbindingcontract.** Paper bundle authority: core/papermodulecapsules.json#papermodule.patternbinding_contract. Mechanism source: core/mechanismsources.json#mechanism.patternbindingcontract.validatespublicpatternbindings.",
"level": 4,
"order": 20,
"section_id": "paper-capsule-authority",
"source_line_end": 204,
"source_line_start": 197,
"title": "Paper bundle authority"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "reader-evidence-routing",
"source_line_end": 239,
"source_line_start": 238,
"title": "Reader evidence routing"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Start with papermodules/patternbindingcontract.json for the bundle payload, then open standards/stdmicrocosmpatternbindingcontract.json to check the required fields, public/private boundary, source-open body import floor, route-readiness rules, and result record expectations.** Start with papermodules/patternbinding_contract.json for the bundle payload, then open standards/stdmicrocosmpatternbindingcontract.json to check the required fields, public/private boundary, source-open body import floor, route-readiness rules, and result record expectations.",
"level": 4,
"order": 22,
"section_id": "start-with-paper-modules-pattern-binding-contract-json-for-the-capsule-payload",
"source_line_end": 247,
"source_line_start": 240,
"title": "Start with `paper_modules/pattern_binding_contract.json` for the bundle payload"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Use core/fixturemanifests/patternbindingcontract.fixturemanifest.json before inspecting fixtures or exported bundles. The manifest and the sourcemodulemanifest.json files name the copied source body floor; result record payloads carry source refs, digests, anchors, counts, verdicts, and omission result records rather than inlining body text. Use core/fixturemanifests/patternbindingcontract.fixturemanifest.json before inspecting fixtures or exported bundles. The manifest and the sourcemodulemanifest.json files name the copied source body floor; result record payloads carry source refs, digests, anchors, counts, verdicts, and omission result records rather than inlining body text.",
"level": 4,
"order": 23,
"section_id": "use-core-fixture-manifests-pattern-binding-contract-fixture-manifest-json-before-inspecting-fixtures-or-exported-bundles",
"source_line_end": 256,
"source_line_start": 248,
"title": "Use `core/fixture_manifests/pattern_binding_contract.fixture_manifest.json` before inspecting fixtures or exported bundles"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate/, reproduce this page's proof boundary with temporary result records:",
"level": 2,
"order": 24,
"section_id": "validation-receipt-path",
"source_line_end": 269,
"source_line_start": 257,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the fixture, system-bundle, and route-readiness result records all show status: pass with metadata-only authority handles, with Scope limit stating what it does not certify.** A pass means the fixture, system-bundle, and route-readiness result records all show status: pass with metadata-only authority handles; see Scope limit for what a pass does and does not certify.",
"level": 4,
"order": 25,
"section_id": "what-a-pass-shows",
"source_line_end": 278,
"source_line_start": 270,
"title": "What a pass shows"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The current authority is the runtime result record set, and cold readers should read result record fields such as status and secretexclusionscan rather than markdown constants or a separate sign-off result record.** The current authority is the runtime result record set under result records/firstwave/patternbinding_contract/; do not cite a separate pattern-specific sign-off result record unless an sign-off-lane artifact is actually present. Cold readers should inspect result record fields rather than markdown constants: status, secretexclusionscan, sourceopenbodyimports, truthaccounting, routereadinesssummary, selectioncontract, and sourcemanifest.",
"level": 4,
"order": 26,
"section_id": "inspect-receipt-fields-not-markdown-constants",
"source_line_end": 291,
"source_line_start": 279,
"title": "Inspect result record fields, not markdown constants"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 27,
"section_id": "prior-art-grounding",
"source_line_end": 310,
"source_line_start": 309,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component follows the software pattern-language tradition of making reusable engineering structures explicit and reviewable, with the Hillside patterns library as its direct prior-art family.** This component follows the software pattern-language tradition of making reusable engineering structures explicit, named, and reviewable. The Hillside patterns library is the direct prior-art family for treating patterns as shared vocabulary rather than loose implementation notes.",
"level": 4,
"order": 28,
"section_id": "pattern-language-lineage",
"source_line_end": 321,
"source_line_start": 311,
"title": "Pattern-language lineage"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The binding layer borrows provenance and supply-chain attestation shapes from W3C PROV, SLSA, and in-toto, applied to pattern rows and route-readiness selectors rather than launch certification.** The binding layer also borrows from provenance and supply-chain attestation patterns. W3C PROV motivates the source/ref/evidence relation shape, while SLSA and in-toto motivate digest-bound artifact claims and step-level metadata. Microcosm applies those ideas to pattern rows and route-readiness selectors, not to launch certification.",
"level": 4,
"order": 29,
"section_id": "provenance-and-supply-chain-attestation",
"source_line_end": 333,
"source_line_start": 322,
"title": "Provenance and supply-chain attestation"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Re-entry condition: if copied source bodies, route-readiness overlays, or negative-case rules change, rerun the validate, validate-system-bundle, and validate-route-readiness-bundle commands under Validation Result record Path and update this paper module plus standards/stdmicrocosmpatternbindingcontract.json from the new result record fields. Re-entry condition: if copied source bodies, route-readiness overlays, or negative-case rules change, rerun the validate, validate-system-bundle, and validate-route-readiness-bundle commands under Validation Result record Path and update this paper module plus standards/stdmicrocosmpatternbindingcontract.json from the new result record fields.",
"level": 4,
"order": 30,
"section_id": "re-entry-condition-if-copied-macro-bodies",
"source_line_end": 342,
"source_line_start": 334,
"title": "Re-entry condition: if copied source bodies"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/pattern_binding_contract.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "teleology",
"source": "markdown_section",
"text": null
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "pattern_binding_contract.py",
"path": "src/microcosm_core/organs/pattern_binding_contract.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for pattern-binding fixture validation, source-module import checks, exported system bundle validation, authority-handle validation, and metadata-only result records.",
"symbols": [
"validate_source_module_imports",
"validate_pattern_bindings",
"validate_source_capsules",
"validate_reference_capsules",
"validate_authority_chain_handle_resolver",
"validate_authority_ceiling",
"validate",
"validate_substrate_bundle"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/pattern_binding_contract.py"
},
{
"label": "pattern_route_readiness.py",
"path": "src/microcosm_core/macro_tools/pattern_route_readiness.py",
"resolution": "resolved",
"role": "Route-readiness selector validator consumed by the pattern-binding component.",
"symbols": [
"validate_route_readiness_bundle"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/macro_tools/pattern_route_readiness.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/pattern_binding_contract.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/pattern_binding_contract.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Pattern Binding Contract is the public system membrane for mined pattern rows. It validates binding fields, source bundles, reference bundles, authority-chain handles, secret-exclusion scans, exported system bundles, and route-readiness selector overlays, then writes bounded result records that keep private bodies out and prevent individual pattern rows from becoming standalone public leaves.",
"summary_line": "The public pattern-binding component validates pattern rows, source bundles, authority handles, exported system bundles, and route-readiness selector overlays while keeping mined rows component-first and fixture-bound.",
"title": "Pattern Binding Contract"
},
{
"anchor": "paper-module-bridge-phase-continuity-runtime",
"authority_ceiling": "Public synthetic observe/apply fixture and copied body digest evidence only; no live bridge transport, provider/UI uptime, operator HUD/browser state, live phase runtime, prompt-shelf/private-memory bodies, work landing, source-file changes, launch-scope decision, or whole-system correctness.",
"body_html": "
Component replays synthetic
\n
This component replays a synthetic record of a paused background job and checks the rules for safely resuming it, opening no live bridge and touching no private memory.bridge_phase_continuity_runtime is a public component that replays a small synthetic record of a paused background job and checks that the rules for safely resuming it hold. It reads fixture files on disk. It opens no live bridge, calls no provider, reads no operator screen state, and touches no private memory.
\n
Run produces five
\n
A run produces five plain result records for packet, heartbeat, pressure, resume, and completion, each carrying status fields, counts, and error codes but no copied body, live packet, or secret. A run produces five plain result records: one for the continuation packet, one for the heartbeat, one for resource pressure, one for the resume decision, and one for the completion transition. Each record carries status fields, counts, and error codes. None of them carries a copied body, a live packet, or a secret.
\n
Component answers whether
\n
The component answers whether a disk-first pause-and-resume handoff can be validated against fixed rules from public fixtures, where a pass means the fixture held, not that a real job is alive. The component answers one bounded question. Can a disk-first pause-and-resume handoff be represented by public fixture inputs, validated against a fixed set of rules, and kept below any live-bridge or launch-scope decision? A pass means the fixture held those rules. It does not mean a real job is alive.
\n
Purpose
\n
How loose resume stories overreach
\n
Detached work is easy to describe and hard to trust, and conflating a liveness ping with resume permission, or a resume with finished work, is how bridge systems overreach. Detached background work is easy to describe and hard to trust. A running agent says it paused, left a note, and can resume. A loose version of that story lets a "still alive" ping stand in for permission to resume, or lets a successful resume stand in for finished work. Those are different claims, and conflating them is how bridge systems overreach.
\n
Fixed continuity rules a cold reader can run
\n
It makes the difference checkable with fixed rules a reader can run against fixtures: a paused job leaves a packet, double resume is refused, and only completion claims finished work. This component makes the difference checkable. It exists so a cold reader can run the continuity rules against fixture files instead of trusting a story about hidden agents. The rules are fixed. A paused job must leave a real continuation packet. Resuming the same job twice is refused. A heartbeat is liveness evidence and never resume authority. A stale heartbeat cannot pass itself off as proof of health. Only a proper completion record may claim the work finished.
\n
How it works
\n
run reads the fixture and six synthetic inputs
\n
The run entry point resolves the session fixture and manifests, then reads six synthetic transport inputs, streaming the heartbeat so a malformed line becomes a finding, not an ingested body. The entry point is run. It takes an input directory and an output directory. It reads the session fixture observe_apply_session_fixture.json, pulls the fixture-manifest and source-module-manifest references out of it, and resolves both. From the manifest it reads six synthetic transport inputs through _read_synthetic_transport_inputs: detached_job.json, continuation_packet.json, heartbeat_rows.jsonl, resource_pressure.json, worker_skip_receipt.json, and private_state_forbidden_terms.json. The heartbeat file is JSON Lines, streamed by _read_required_jsonl, so a malformed line becomes a finding and never a reason to ingest a whole live heartbeat body.
\n
The positive job's sign-off conditions
\n
_validate_synthetic_transport_contract accepts the one positive job only when it yielded to disk with an unconsumed packet, a fresh heartbeat, and agreeing phase and continuity references across job, packet, and heartbeat. The core check is _validate_synthetic_transport_contract. It reads the job, packet, heartbeat, pressure, and worker-skip rows, then evaluates one positive job against a floor of negative perturbations. The positive job (synthetic_detached_job_001) passes only when its state is yielded_to_disk, its transport label is the accepted synthetic-transport label, its continuation packet targets the same job id, the packet's claim reference matches the job's, the packet is not yet consumed, at least one fresh heartbeat is present, and the phase id and continuity reference agree across job, packet, and heartbeat. Freshness is concrete. _heartbeat_is_fresh requires status alive and an age at or under 60 seconds. _heartbeat_is_stale treats status stale, or an age over 300 seconds, as stale.
\n
Every declared bad case must be rejected
\n
The same function turns each declared bad case into a specific error code, and a pass requires the fixture both to accept the good job and to observe the whole refusal floor. The same function checks that each declared bad case is actually rejected. A missing packet, a packet missing required fields, an already-consumed packet, a heartbeat that claims resume authority, a stale heartbeat that claims live health, a blocked resource-pressure decision, an invalid worker-skip record, and a private-class fixture that leaks bodies each turn into a specific error code. When a perturbation is present and correctly rejected, the run records the matching code. When the fixture fails to carry a perturbation, the run records a finding and blocks. A pass therefore means the fixture both accepted the one good job and observed the whole refusal floor. The declared cases live in the EXPECTED_NEGATIVE_CASES constant.
\n
_validate_fixture_contract matches ids and digests
\n
_validate_fixture_contract matches the envelope ids and recomputes each source-module SHA-256 so drift becomes BRIDGE_CONTINUITY_SOURCE_DIGEST_MISMATCH, then requires can_continue, closed_with_receipts, and a rollback that refuses writes after failure._validate_fixture_contract checks the session envelope. The fixture's organ_id and its pattern_id (observe_runtime_apply_session) must match, the manifest ids must match, and the fixture's source-module references must equal the manifest's. For each listed source module it recomputes a SHA-256 digest and compares it to the recorded one, so a drifted copied body becomes BRIDGE_CONTINUITY_SOURCE_DIGEST_MISMATCH rather than silent trust. It then requires the status packet to report can_continue, the finalizer status to be closed_with_receipts, and the rollback case to refuse writes after a validation failure.
\n
Forbidden-class scan and the tracked-write gate
\n
Between the checks, run scans every path and strips excerpts through _safe_scan to report counts without text, then gates writes so a tracked path without the env flag reports tracked_receipt_writes_blocked. Between those two checks, run scans every input path with scan_paths against the forbidden-class policy and strips matched excerpts through _safe_scan, so the scan reports counts and status without carrying the offending text. Only then does it build the five result records and write them. Writes pass through a gate: when the target is a tracked path and MICROCOSM_TRACKED_RECEIPT_WRITES=1 is absent, the run reports tracked_receipt_writes_blocked instead of refreshing tracked evidence. run returns the completion-transition record.
\n
Function
Role
run
Reads the fixture and manifests, drives the checks, writes five result records, returns the completion record.
_read_required_jsonl
Streams the heartbeat JSON Lines; a bad line is a finding, not an ingested body.
_validate_synthetic_transport_contract
Accepts the one good job and requires every declared bad case to be rejected, emitting error codes.
_validate_fixture_contract
Matches ids, compares source-module SHA-256 digests, and checks the continue, finalizer, and rollback status.
\nDiagram of the mechanism (10 steps).Diagram source & refs
Source refs
_validate_fixture_contract ids, SHA-256 digests,
continue/finalizer/rollback
Skip
tracked_receipt_writes_blocked
flowchart TD Read["run reads fixture and six synthetic inputs"] --> Transport["_validate_synthetic_transport_contract"] Transport --> Good{"One valid job? yielded to disk, packet unconsumed, fresh heartbeat, phase and continuity agree"} Good -->|"yes"| Floor["Every declared bad case rejected as an error code"] Good -->|"no"| Block["Finding: valid job invalid"] Floor --> Fixture["_validate_fixture_contract ids, SHA-256 digests, continue/finalizer/rollback"] Fixture --> Scan["scan_paths and _safe_scan counts only, no excerpts"] Scan --> Gate{"Tracked path without MICROCOSM_TRACKED_RECEIPT_WRITES=1?"} Gate -->|"yes"| Skip["tracked_receipt_writes_blocked"] Gate -->|"no"| Write["Five result records written"]
\n
Negative cases
\n
Source-declared refusal floor mapped to codes
\n
The refusal floor is source-declared in EXPECTED_NEGATIVE_CASES and validated against fixture contents, with each named case mapping to the error codes the run must observe. The refusal floor is source-declared in EXPECTED_NEGATIVE_CASES and validated against fixture contents. Each named case maps to the error codes the run must observe:
The runtime borrows fixture-level shapes from Temporal, Airflow, Kubernetes leases, and W3C PROV without running any: a resumable handoff, a liveness lease that is not permission, and an inspectable provenance record. The runtime borrows shapes from durable execution and provenance work without running any of it. Temporal keeps workflow state resumable across process failure; Apache Airflow DAGs separate task ordering and retry policy from task internals; Kubernetes lease-based leader election treats a heartbeat as liveness evidence and not as work completion; and W3C PROV records how an output was produced. This component keeps only the fixture-level shape of those ideas: a resumable handoff, a liveness lease that is not permission, and a provenance record a reader can inspect.
\n
Validation Result record Path
\n
Run focused tests
\n
Run the focused tests and the corpus check from the repository root without mutating durable records or generated projections. Run the focused tests and the corpus check from the repository root without mutating durable records or generated projections:
To exercise the component directly against its fixture:
\n
microcosm bridge-phase-continuity-runtime run --input fixtures/second_wave/bridge_phase_continuity_runtime/input --out /tmp/microcosm-bridge-continuity
\n
What a passing run confirms
\n
A pass confirms the fixture accepted the one good job, rejected every declared bad case, kept the non-public-state scan clean, and held the source-module digests. A pass means the fixture accepted the one good job, rejected every declared bad case, kept the non-public-state scan clean, and held the source-module digests.
\n
Scope boundary
\n
Scope limit
\n
The bounded proof boundary
\n
The strongest supported claim is bounded: over its public fixture a disk-first pause-and-resume handoff validates against a fixed rule set into five metadata-only records, and nothing wider. The strongest claim this component supports is a bounded one: over its public fixture, a disk-first pause-and-resume handoff can be validated by a fixed rule set, and the resulting five result records are metadata-only. Its proof boundary is that fixture and its validation result records, nothing wider.
\n
What each record does not establish
\n
It runs no live bridge, provider, or private memory; a heartbeat proves liveness not resume, a resume proves single consumption not landing, and completion proves fixture state, not a finished job. It does not run live bridge transport, call a provider, read operator screen or browser state, read live phase state, or read private-memory bodies. A fresh heartbeat proves liveness, not resume authority. A resume proves the packet was consumed once, not that the apply step landed. A completion record proves the fixture reached its declared state, not that a real provider-backed job finished. The scope limit excludes source-file changes, hosted-public posture, launch, and any whole-system correctness claim.
\n",
"components": [
{
"anchor": "component-bridge_phase_continuity_runtime",
"component_id": "bridge_phase_continuity_runtime",
"public_label": "Bridge-Continuity Sign-off Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 9
},
"link_count": 19,
"links": [
{
"id": "bridge_phase_continuity_runtime",
"kind": "organ",
"kind_label": "component",
"label": "Bridge-Continuity Sign-off Replay",
"relation": "explains",
"url": "component-bridge-phase-continuity-runtime.html"
},
{
"id": "mechanism.bridge_phase_continuity_runtime.validates_synthetic_bridge_continuity",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates synthetic bridge continuity",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-bridge-phase-continuity-runtime-validates-synthetic-bridge-continuity"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.work_landing_and_continuity_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "work landing and continuity controls as transaction-scoped coordination evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-work-landing-and-continuity-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.agent_memory_temporal_conflict_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Memory Temporal-Conflict Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[32:paper_module.agent_memory_temporal_conflict_replay].depends_on",
"url": "paper-module-agent-memory-temporal-conflict-replay.html"
},
{
"id": "paper_module.concurrency_mission_control",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Concurrency Mission Control",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[47:paper_module.concurrency_mission_control].depends_on",
"url": "paper-module-concurrency-mission-control.html"
},
{
"id": "paper_module.durable_agent_work_landing_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Durable Agent Work-Landing Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[16:paper_module.durable_agent_work_landing_replay].depends_on",
"url": "paper-module-durable-agent-work-landing-replay.html"
}
],
"paper_module_id": "paper_module.bridge_phase_continuity_runtime",
"reader_packet": {
"abstract": "The public bridge-continuity fixture validates disk-first continuation packets, heartbeat/resource-pressure boundaries, resume-once semantics, worker-skip dedupe, tracked result record-write gates, and non-public-state exclusion without live bridge transport.",
"authority_ceiling": "Public synthetic observe/apply fixture and copied body digest evidence only; no live bridge transport, provider/UI uptime, operator HUD/browser state, live phase runtime, prompt-shelf/private-memory bodies, work landing, source-file changes, launch-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.bridge_phase_continuity_runtime.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-bridge-phase-continuity-runtime-diagram-1",
"html_url": "docs/paper-module-bridge-phase-continuity-runtime.html#paper-module-bridge-phase-continuity-runtime-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Read[\"run reads fixture and six synthetic inputs\"] --> Transport[\"_validate_synthetic_transport_contract\"] Transport --> Good{\"One valid job? yielded to disk, packet unconsumed, fresh heartbeat, phase and continuity agree\"} Good -->|\"yes\"| Floor[\"Every declared bad case rejected as an error code\"] Good -->|\"no\"| Block[\"Finding: valid job invalid\"] Floor --> Fixture[\"_validate_fixture_contract ids, SHA-256 digests, continue/finalizer/rollback\"] Fixture --> Scan[\"scan_paths and _safe_scan counts only, no excerpts\"] Scan --> Gate{\"Tracked path without MICROCOSM_TRACKED_RECEIPT_WRITES=1?\"} Gate -->|\"yes\"| Skip[\"tracked_receipt_writes_blocked\"] Gate -->|\"no\"| Write[\"Five result records written\"]",
"source_format": "mermaid",
"source_hash": "sha256:5db0ff5222587c0643a63df56c61fdd1f426e7194079982c8b37a9eb8f0bdf3e",
"source_line_end": 151,
"source_line_start": 140
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "bridge_phase_continuity_runtime",
"kind": "organ",
"kind_label": "component",
"label": "Bridge-Continuity Sign-off Replay",
"relation": "explains",
"url": "component-bridge-phase-continuity-runtime.html"
},
{
"id": "mechanism.bridge_phase_continuity_runtime.validates_synthetic_bridge_continuity",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates synthetic bridge continuity",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-bridge-phase-continuity-runtime-validates-synthetic-bridge-continuity"
}
],
"relation": "explains"
},
{
"count": 10,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.work_landing_and_continuity_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "work landing and continuity controls as transaction-scoped coordination evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-work-landing-and-continuity-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.bridge_phase_continuity_runtime",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 241,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/bridge_phase_continuity_runtime.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Temporal",
"section_id": "durable-execution-shapes-without-running-them",
"url": "https://docs.temporal.io/"
},
{
"kind": "external_reference",
"label": "DAGs",
"section_id": "durable-execution-shapes-without-running-them",
"url": "https://airflow.apache.org/docs/apache-airflow/stable/core-concepts/dags.html"
},
{
"kind": "external_reference",
"label": "lease-based leader election",
"section_id": "durable-execution-shapes-without-running-them",
"url": "https://kubernetes.io/docs/concepts/cluster-administration/coordinated-leader-election/"
},
{
"kind": "external_reference",
"label": "PROV",
"section_id": "durable-execution-shapes-without-running-them",
"url": "https://www.w3.org/TR/prov-overview/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component replays a synthetic record of a paused background job and checks the rules for safely resuming it, opening no live bridge and touching no private memory. bridgephasecontinuityruntime is a public component that replays a small synthetic record of a paused background job and checks that the rules for safely resuming it hold. It reads fixture files on disk. It opens no live bridge, calls no provider, reads no operator screen state, and touches no private memory.",
"level": 4,
"order": 1,
"section_id": "component-replays-synthetic",
"source_line_end": 11,
"source_line_start": 3,
"title": "Component replays synthetic"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A run produces five plain result records for packet, heartbeat, pressure, resume, and completion, each carrying status fields, counts, and error codes but no copied body, live packet, or secret.** A run produces five plain result records: one for the continuation packet, one for the heartbeat, one for resource pressure, one for the resume decision, and one for the completion transition. Each record carries status fields, counts, and error codes. None of them carries a copied body, a live packet, or a secret.",
"level": 4,
"order": 2,
"section_id": "run-produces-five",
"source_line_end": 21,
"source_line_start": 12,
"title": "Run produces five"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component answers whether a disk-first pause-and-resume handoff can be validated against fixed rules from public fixtures, where a pass means the fixture held, not that a real job is alive.** The component answers one bounded question. Can a disk-first pause-and-resume handoff be represented by public fixture inputs, validated against a fixed set of rules, and kept below any live-bridge or launch-scope decision? A pass means the fixture held those rules. It does not mean a real job is alive.",
"level": 4,
"order": 3,
"section_id": "component-answers-whether",
"source_line_end": 31,
"source_line_start": 22,
"title": "Component answers whether"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 33,
"source_line_start": 32,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Detached work is easy to describe and hard to trust, and conflating a liveness ping with resume permission, or a resume with finished work, is how bridge systems overreach.** Detached background work is easy to describe and hard to trust. A running agent says it paused, left a note, and can resume. A loose version of that story lets a \"still alive\" ping stand in for permission to resume, or lets a successful resume stand in for finished work. Those are different claims, and conflating them is how bridge systems overreach.",
"level": 4,
"order": 5,
"section_id": "how-loose-resume-stories-overreach",
"source_line_end": 43,
"source_line_start": 34,
"title": "How loose resume stories overreach"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It makes the difference checkable with fixed rules a reader can run against fixtures: a paused job leaves a packet, double resume is refused, and only completion claims finished work.** This component makes the difference checkable. It exists so a cold reader can run the continuity rules against fixture files instead of trusting a story about hidden agents. The rules are fixed. A paused job must leave a real continuation packet. Resuming the same job twice is refused. A heartbeat is liveness evidence and never resume authority. A stale heartbeat cannot pass itself off as proof of health. Only a proper completion record may claim the work finished.",
"level": 4,
"order": 6,
"section_id": "fixed-continuity-rules-a-cold-reader-can-run",
"source_line_end": 55,
"source_line_start": 44,
"title": "Fixed continuity rules a cold reader can run"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 57,
"source_line_start": 56,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The run entry point resolves the session fixture and manifests, then reads six synthetic transport inputs, streaming the heartbeat so a malformed line becomes a finding, not an ingested body.** The entry point is run. It takes an input directory and an output directory. It reads the session fixture observeapplysession_fixture.json, pulls the fixture-manifest and source-module-manifest references out of it, and resolves both. From the manifest it reads six synthetic transport inputs through readsynthetictransportinputs: detached_job.json, continuationpacket.json, heartbeatrows.jsonl, resource_pressure.json, workerskipreceipt.json, and privatestateforbidden_terms.json. The heartbeat file is\u2026",
"level": 4,
"order": 8,
"section_id": "run-reads-the-fixture-and-six-synthetic-inputs",
"source_line_end": 72,
"source_line_start": 58,
"title": "`run` reads the fixture and six synthetic inputs"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatesynthetictransportcontract accepts the one positive job only when it yielded to disk with an unconsumed packet, a fresh heartbeat, and agreeing phase and continuity references across job, packet, and heartbeat.** The core check is validatesynthetictransportcontract. It reads the job, packet, heartbeat, pressure, and worker-skip rows, then evaluates one positive job against a floor of negative perturbations. The positive job (syntheticdetachedjob001) passes only when its state is yieldedto_disk, its transport label is the accepted synthetic-transport label, its continuation packet targets the same job id, the packet's claim reference matches the job's, the packet is not yet consumed,\u2026",
"level": 4,
"order": 9,
"section_id": "the-positive-job-s-acceptance-conditions",
"source_line_end": 89,
"source_line_start": 73,
"title": "The positive job's sign-off conditions"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The same function turns each declared bad case into a specific error code, and a pass requires the fixture both to accept the good job and to observe the whole refusal floor.** The same function checks that each declared bad case is actually rejected. A missing packet, a packet missing required fields, an already-consumed packet, a heartbeat that claims resume authority, a stale heartbeat that claims live health, a blocked resource-pressure decision, an invalid worker-skip record, and a private-class fixture that leaks bodies each turn into a specific error code. When a perturbation is present and correctly rejected, the run records the matching code. When the fixture fails to carry a per\u2026",
"level": 4,
"order": 10,
"section_id": "every-declared-bad-case-must-be-rejected",
"source_line_end": 104,
"source_line_start": 90,
"title": "Every declared bad case must be rejected"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatefixture_contract matches the envelope ids and recomputes each source-module SHA-256 so drift becomes BRIDGECONTINUITYSOURCEDIGESTMISMATCH, then requires cancontinue, closedwith_receipts, and a rollback that refuses writes after failure. validatefixturecontract checks the session envelope. The fixture's organid and its patternid (observeruntimeapply_session) must match, the manifest ids must match, and the fixture's source-module references must equal the manifest's. For each listed source module it recomputes a SHA-256 digest and compares it to the recorded one, so a drifted copied body becomes BRIDGECONTINUITYSOURCEDIGESTMISMATCH rather than silent trust. It then requires the statu\u2026",
"level": 4,
"order": 11,
"section_id": "validate-fixture-contract-matches-ids-and-digests",
"source_line_end": 119,
"source_line_start": 105,
"title": "`_validate_fixture_contract` matches ids and digests"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "**Between the checks, run scans every path and strips excerpts through safescan to report counts without text, then gates writes so a tracked path without the env flag reports trackedreceiptwritesblocked. Between those two checks, run scans every input path with scan_paths against the forbidden-class policy and strips matched excerpts through safescan, so the scan reports counts and status without carrying the offending text. Only then does it build the five result records and write them. Writes pass through a gate: when the target is a tracked path and MICROCOSMTRACKEDRECEIPT_WRITES=1 is absent, the run reports trackedreceiptwrites_blocked instead of refreshing tracked evidence. run return\u2026",
"level": 4,
"order": 12,
"section_id": "forbidden-class-scan-and-the-tracked-write-gate",
"source_line_end": 152,
"source_line_start": 120,
"title": "Forbidden-class scan and the tracked-write gate"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 154,
"source_line_start": 153,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The refusal floor is source-declared in EXPECTEDNEGATIVE_CASES and validated against fixture contents, with each named case mapping to the error codes the run must observe. The refusal floor is source-declared in EXPECTEDNEGATIVE_CASES and validated against fixture contents. Each named case maps to the error codes the run must observe: missingpacketduplicateresumeandresourceblock: MISSINGCONTINUATIONPACKET, RESOURCEPRESSUREDISPATCHBLOCKED, CONTINUATIONPACKETALREADYCONSUMED continuationpacketmissingrequiredfields: MISSINGCONTINUATIONPACKET_FIELDS heartbeatclaimsresumeauthority: HEARTBEATNOTRESUMEAUTHORITY bridgepacketprivatehudbody: BRIDGEPACKETPRIVATEHUDBODY staleheartbeatoverclaimsliveness\u2026",
"level": 4,
"order": 14,
"section_id": "source-declared-refusal-floor-mapped-to-codes",
"source_line_end": 170,
"source_line_start": 155,
"title": "Source-declared refusal floor mapped to codes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 172,
"source_line_start": 171,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The runtime borrows fixture-level shapes from Temporal, Airflow, Kubernetes leases, and W3C PROV without running any: a resumable handoff, a liveness lease that is not permission, and an inspectable provenance record.** The runtime borrows shapes from durable execution and provenance work without running any of it. Temporal keeps workflow state resumable across process failure; Apache Airflow DAGs separate task ordering and retry policy from task internals; Kubernetes lease-based leader election treats a heartbeat as liveness evidence and not as work completion; and W3C PROV records how an output was produced. This component keeps only the fixture-level shape of those ideas: a resumable h\u2026",
"level": 4,
"order": 16,
"section_id": "durable-execution-shapes-without-running-them",
"source_line_end": 189,
"source_line_start": 173,
"title": "Durable-execution shapes without running them"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 191,
"source_line_start": 190,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "**Run the focused tests and the corpus check from the repository root without mutating durable records or generated projections.** Run the focused tests and the corpus check from the repository root without mutating durable records or generated projections: To exercise the component directly against its fixture:",
"level": 4,
"order": 18,
"section_id": "run-focused-tests",
"source_line_end": 208,
"source_line_start": 192,
"title": "Run focused tests"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass confirms the fixture accepted the one good job, rejected every declared bad case, kept the non-public-state scan clean, and held the source-module digests.** A pass means the fixture accepted the one good job, rejected every declared bad case, kept the non-public-state scan clean, and held the source-module digests.",
"level": 4,
"order": 19,
"section_id": "what-a-passing-run-confirms",
"source_line_end": 216,
"source_line_start": 209,
"title": "What a passing run confirms"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 218,
"source_line_start": 217,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest supported claim is bounded: over its public fixture a disk-first pause-and-resume handoff validates against a fixed rule set into five metadata-only records, and nothing wider.** The strongest claim this component supports is a bounded one: over its public fixture, a disk-first pause-and-resume handoff can be validated by a fixed rule set, and the resulting five result records are metadata-only. Its proof boundary is that fixture and its validation result records, nothing wider.",
"level": 4,
"order": 21,
"section_id": "the-bounded-proof-boundary",
"source_line_end": 228,
"source_line_start": 219,
"title": "The bounded proof boundary"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It runs no live bridge, provider, or private memory; a heartbeat proves liveness not resume, a resume proves single consumption not landing, and completion proves fixture state, not a finished job.** It does not run live bridge transport, call a provider, read operator screen or browser state, read live phase state, or read private-memory bodies. A fresh heartbeat proves liveness, not resume authority. A resume proves the packet was consumed once, not that the apply step landed. A completion record proves the fixture reached its declared state, not that a real provider-backed job finished. The scope limit excludes source-file changes, hosted-public posture, launch, and any whole-system co\u2026",
"level": 4,
"order": 22,
"section_id": "what-each-record-does-not-prove",
"source_line_end": 240,
"source_line_start": 229,
"title": "What each record does not establish"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/bridge_phase_continuity_runtime.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public bridge-continuity fixture validates disk-first continuation packets, heartbeat/resource-pressure boundaries, resume-once semantics, worker-skip dedupe, tracked result record-write gates, and non-public-state exclusion without live bridge transport."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "bridge_phase_continuity_runtime.py",
"path": "src/microcosm_core/organs/bridge_phase_continuity_runtime.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public synthetic bridge continuity validation, copied source-module digest checks, result record-write gating, negative cases, and scope limits.",
"symbols": [
"run",
"_validate_synthetic_transport_contract",
"_validate_fixture_contract",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/bridge_phase_continuity_runtime.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/bridge_phase_continuity_runtime.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/bridge_phase_continuity_runtime.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Bridge Phase Continuity Runtime is the public observe/apply continuity membrane for detached work. It consumes synthetic transport fixtures, validates copied observe-runtime body digests, writes five metadata-only result records, checks seven negative-case classes, and keeps heartbeat/resume evidence below work-landing, provider, UI, source-file changes, and launch-scope decision.",
"summary_line": "The public bridge-continuity fixture validates disk-first continuation packets, heartbeat/resource-pressure boundaries, resume-once semantics, worker-skip dedupe, tracked result record-write gates, and non-public-state exclusion without live bridge transport.",
"title": "Bridge Phase Continuity Runtime"
},
{
"anchor": "paper-module-cognitive-operator-registry",
"authority_ceiling": "Public registry-contract fixture and copied source body evidence only; no live operator execution, registry mutation, source-file changes, external model access, launch-scope decision, non-public data equivalence, operator correctness proof, or whole-system correctness.",
"body_html": "
Registry checks public
\n
The registry checks a public catalogue of named thinking moves and asks which operators have actually changed a live decision, not just described one tidily.cognitive_operator_registry is a checker for the system's catalogue of named thinking moves. Each entry describes one reusable operator, for example "reduce competing pressure to one bounded action" or "compile a handoff packet when validated work cannot be committed". The component reads the public catalogue and answers one question: which of these operators has actually changed a live decision, and which is only a tidy description of one.
\n
Validates every operator
\n
It validates every operator row against a fixed shape, makes each active operator cite a dogfood record with concrete evidence, and rejects voice or authority overclaims. It runs over a public fixture. It reads three input files, validates every operator row against a fixed shape, requires each operator marked active to point at a dogfood result record carrying concrete evidence, and rejects any row that claims the owner's voice or that grants itself launch, provider, or mutation authority. It writes result records that name refs, hashes, counts, and verdicts, never the operator text itself.
\n
Purpose
\n
active requires resolved cognition-delta evidence
\n
An entry may call itself active only when it cites a dogfood record whose cognition_delta_evidence names a decision that changed, with evidence surfaces resolved against the public system. A team that writes its reusable thinking moves down as a registry tends to accumulate entries faster than it can prove any of them helped. A loose registry passes on prose alone: an entry says it was useful, and the reader has to take its word. This component refuses that. An entry may call itself active only if it points to a dogfood record, and that record must carry cognition_delta_evidence describing a concrete decision that came out differently because the operator was applied. Where the record cites evidence surfaces or command paths, the checker resolves each one against the public system rather than trusting the prose.
\n
How it works
\n
Two entrypoints over _build_result
\n
Two entrypoints, run over the first-wave fixture and run_registry_bundle over the exported bundle, both call _build_result to run the ordered pipeline. The runtime lives in src/microcosm_core/organs/cognitive_operator_registry.py. There are two entrypoints. run validates the first-wave fixture and includes the negative cases; run_registry_bundle validates the exported bundle and makes the copied source-body manifest mandatory. Both call _build_result, which runs the pipeline in this order.
\n
The five-step pipeline and pass rule
\n
The pipeline loads inputs, checks copied bodies, runs the secret scan, validates operator rows, and exercises negatives; status is pass only when all four conditions hold. First, _load_payloads reads the three public inputs: operator_registry.json, operator_standard.json, and dogfood_index.json (plus the eight negative inputs in fixture mode). Second, _source_module_manifest_result checks the copied source bodies carried by the exported bundle. Third, scan_paths runs the secret-exclusion pass over every input. Fourth, _positive_findings validates the operator rows. Fifth, _negative_findings exercises the declared rejection cases. The status is pass only when there are no positive findings, no expected negative case went unobserved, no blocking secret hit was found, and the source-module check is pass or not_present.
\n
_positive_findings carries the core contract
\n
_positive_findings checks each row's shape and status, resolves every active operator's dogfood result record and evidence commands, and flags any policy that sets an OVERCLAIM_KEYS flag true._positive_findings carries the core contract. For each operator row it requires all eleven fields in OPERATOR_REQUIRED_FIELDS and a status drawn from candidate, active, or retired. For each active operator it looks up a dogfood row keyed by operator_id, requires the nine fields in DOGFOOD_REQUIRED_RECEIPT_FIELDS, and calls _dogfood_receipt_ref_resolves to confirm the result record ref points at a real dogfood row. It then calls _record_dogfood_evidence_resolution_findings, which parses each evidence surface as a command and resolves its path and handle tokens against the public system. A row whose prose says it was dogfooded but whose evidence does not resolve is recorded as a failure. Finally it checks the registry policy against OVERCLAIM_KEYS: if the policy sets any of release_authorized, provider_calls_authorized, source_mutation_authorized, or the other listed flags to true, that is an overclaim.
\n
The bundle body-floor manifest check
\n
_source_module_manifest_result proves the bundle's copied bodies by class, module_count, sha256 across three digest fields, and declared anchors, recording ids and hashes but never the body text._source_module_manifest_result is the body-floor check used by the bundle entrypoint. The exported bundle at examples/cognitive_operator_registry/exported_cognitive_operator_registry_bundle copies the source registry, standard, and projection-tool bodies. The manifest must classify each as copied_non_secret_macro_body, keep body_in_receipt false, match its declared module_count, and for every row: use a material class, copy the body into source_modules, hash to the recorded sha256 across all three digest fields, and carry the declared anchors in the copied text. The result summary records module ids, classes, line counts, and hashes, never the body text.
\n
Function
Role
run
First-wave fixture entrypoint; writes the result, board, validation, and fixture sign-off records
run_registry_bundle
Exported-bundle entrypoint; requires the source-body manifest
_build_result
Runs the five-step pipeline and computes the pass or blocked status
_positive_findings
Shape, status, active-operator dogfood, and policy-overclaim checks
_dogfood_receipt_ref_resolves
Confirms a dogfood result record ref points at a real row
_record_dogfood_evidence_resolution_findings
Resolves each evidence surface command against the public system
_negative_findings
Exercises the eight declared rejection cases
_source_module_manifest_result
Proves copied source bodies by sha256 and anchors, body text excluded
\nDiagram of the mechanism (6 steps).Diagram source & refs
flowchart TD Inputs["operator_registry.json operator_standard.json dogfood_index.json"] Positive["_positive_findings shape, status, dogfood, policy overclaim"] Negative["_negative_findings eight rejection cases"] Manifest["_source_module_manifest_result copied bodies by sha256"] Secret["scan_paths secret exclusion"] Result["result record refs, hashes, counts, verdicts; body text omitted"] Inputs --> Positive Inputs --> Negative Inputs --> Manifest Inputs --> Secret Positive --> Result Negative --> Result Manifest --> Result Secret --> Result
\n
Negative cases
\n
Eight perturbed inputs and their error codes
\n
The fixture ships eight perturbed inputs, each of which _negative_findings must observe and map to a specific error code, from missing fields to unresolvable dogfood evidence. The fixture ships eight perturbed inputs, and a pass requires every one to be observed and recorded. _negative_findings reads each and emits the matching error code. Missing operator fields give COGOP_MISSING_REQUIRED_FIELD. An active operator with no backing dogfood row gives COGOP_ACTIVE_WITHOUT_DOGFOOD. A dogfood row with no cognition-delta evidence gives COGOP_DOGFOOD_MISSING_COGNITION_DELTA. Two operators sharing a slug or a near-identical claim with no recorded accretion decision give COGOP_OPERATOR_SPRAWL. A policy that grants itself launch, provider, or mutation authority gives COGOP_AUTHORITY_OVERCLAIM. A row that claims operator-voice authority gives COGOP_OPERATOR_VOICE_FORBIDDEN. A private source body or model-output data in a public input gives COGOP_PRIVATE_SOURCE_FORBIDDEN. A dogfood result record whose evidence surfaces do not resolve gives COGOP_COGNITION_DELTA_EVIDENCE_UNRESOLVABLE.
\n
Prior Art Grounding
\n
Cognitive work analysis, PROV, schema, policy
\n
The design draws on cognitive work analysis, W3C PROV provenance, JSON Schema shape checks, and Open Policy Agent's data/policy separation, applied as a public contract check. This component sits in the lineage of cognitive work analysis, provenance tracking, schema validation, and policy-gated registries. Cognitive Work Analysis, summarized in this information-systems design overview, studies reusable moves in complex sociotechnical work. W3C PROV connects a claim to the activities, agents, and evidence used to judge it. JSON Schema is the required-shape pattern behind the operator-row check. Open Policy Agent is a precedent for keeping policy evaluation distinct from the data being evaluated. This component applies those patterns as a public contract check over already-public inputs.
\n
Validation Result record Path
\n
Run the first-wave fixture into disposable output from the Microcosm root:
\n
Run the exported bundle through the same component:
A pass means the public operator contract replayed from public inputs with every shape check holding and every rejection observed, but does not certify the operators as correct. A pass means the public operator contract replayed from public inputs: every shape check held, every expected rejection was observed, and no body text reached the result records. It does not certify the operators as generally correct.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim is that the operator contract is replayable from public inputs, with the checker rejecting bad shape, unresolved evidence, sprawl, voice, overclaim, and leakage. The strongest claim the evidence supports is that the public operator contract is replayable from public inputs. The checker rejects missing shape, missing or unresolved dogfood evidence, near-duplicate sprawl, operator-voice authority, authority overclaim, and private-source leakage, and it proves the copied source bodies by hash and anchor while keeping their text out of the result records.
\n
Where the proof boundary stops
\n
The component is source-linked only, mutates no operators, proves no operator correct beyond its dogfood record, and excludes voice export, external model access, mutation, or launch. The proof boundary stops there. This component does not become source authority for the operator registry, does not mutate operators, and does not establish that any listed operator is correct beyond its own dogfood record. Its scope limit excludes private operator bodies, operator-voice export, external model access, source-file changes, and launch or public sharing. If the focused tests report a copied-body mismatch, that is a source-copy repair, not a change to this reader page.
\n",
"components": [
{
"anchor": "component-cognitive_operator_registry",
"component_id": "cognitive_operator_registry",
"public_label": "Cognitive Operator Registry"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 9
},
"link_count": 19,
"links": [
{
"id": "cognitive_operator_registry",
"kind": "organ",
"kind_label": "component",
"label": "Cognitive Operator Registry",
"relation": "explains",
"url": "component-cognitive-operator-registry.html"
},
{
"id": "mechanism.cognitive_operator_registry.validates_public_operator_contract",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public operator contract",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-cognitive-operator-registry-validates-public-operator-contract"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.cognitive_operator_registry",
"reader_packet": {
"abstract": "The public cognitive-operator registry fixture validates operator-shape rows, active-operator dogfood result records, anti-sprawl decisions, copied source registry/standard/tool bodies, and scope limits without becoming operator source authority.",
"authority_ceiling": "Public registry-contract fixture and copied source body evidence only; no live operator execution, registry mutation, source-file changes, external model access, launch-scope decision, non-public data equivalence, operator correctness proof, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.cognitive_operator_registry.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-cognitive-operator-registry-diagram-1",
"html_url": "docs/paper-module-cognitive-operator-registry.html#paper-module-cognitive-operator-registry-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Inputs[\"operator_registry.json operator_standard.json dogfood_index.json\"] Positive[\"_positive_findings shape, status, dogfood, policy overclaim\"] Negative[\"_negative_findings eight rejection cases\"] Manifest[\"_source_module_manifest_result copied bodies by sha256\"] Secret[\"scan_paths secret exclusion\"] Result[\"result record refs, hashes, counts, verdicts; body text omitted\"] Inputs --> Positive Inputs --> Negative Inputs --> Manifest Inputs --> Secret Positive --> Result Negative --> Result Manifest --> Result Secret --> Result",
"source_format": "mermaid",
"source_hash": "sha256:864c67ad80874dd75a3c58347edbd064437fee7f242fce439387b39860aecd51",
"source_line_end": 134,
"source_line_start": 118
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "cognitive_operator_registry",
"kind": "organ",
"kind_label": "component",
"label": "Cognitive Operator Registry",
"relation": "explains",
"url": "component-cognitive-operator-registry.html"
},
{
"id": "mechanism.cognitive_operator_registry.validates_public_operator_contract",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public operator contract",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-cognitive-operator-registry-validates-public-operator-contract"
}
],
"relation": "explains"
},
{
"count": 10,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.cognitive_operator_registry",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 235,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/cognitive_operator_registry.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "information-systems design overview",
"section_id": "cognitive-work-analysis-prov-schema-policy",
"url": "https://files.eric.ed.gov/fulltext/EJ1082064.pdf"
},
{
"kind": "external_reference",
"label": "PROV",
"section_id": "cognitive-work-analysis-prov-schema-policy",
"url": "https://www.w3.org/TR/prov-overview/"
},
{
"kind": "external_reference",
"label": "JSON Schema",
"section_id": "cognitive-work-analysis-prov-schema-policy",
"url": "https://json-schema.org/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The registry checks a public catalogue of named thinking moves and asks which operators have actually changed a live decision, not just described one tidily.** cognitiveoperatorregistry is a checker for the system's catalogue of named thinking moves. Each entry describes one reusable operator, for example \"reduce competing pressure to one bounded action\" or \"compile a handoff packet when validated work cannot be committed\". The component reads the public catalogue and answers one question: which of these operators has actually changed a live decision, and which is only a tidy description of one.",
"level": 4,
"order": 1,
"section_id": "registry-checks-public",
"source_line_end": 13,
"source_line_start": 3,
"title": "Registry checks public"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It validates every operator row against a fixed shape, makes each active operator cite a dogfood record with concrete evidence, and rejects voice or authority overclaims.** It runs over a public fixture. It reads three input files, validates every operator row against a fixed shape, requires each operator marked active to point at a dogfood result record carrying concrete evidence, and rejects any row that claims the owner's voice or that grants itself launch, provider, or mutation authority. It writes result records that name refs, hashes, counts, and verdicts, never the operator text itself.",
"level": 4,
"order": 2,
"section_id": "validates-every-operator",
"source_line_end": 25,
"source_line_start": 14,
"title": "Validates every operator"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 3,
"section_id": "purpose",
"source_line_end": 27,
"source_line_start": 26,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**An entry may call itself active only when it cites a dogfood record whose cognitiondeltaevidence names a decision that changed, with evidence surfaces resolved against the public system.** A team that writes its reusable thinking moves down as a registry tends to accumulate entries faster than it can prove any of them helped. A loose registry passes on prose alone: an entry says it was useful, and the reader has to take its word. This component refuses that. An entry may call itself active only if it points to a dogfood record, and that record must carry cognitiondeltaevidence describing a concrete decision that came out differently because the operator was applied. Where the record cites\u2026",
"level": 4,
"order": 4,
"section_id": "active-requires-resolved-cognition-delta-evidence",
"source_line_end": 42,
"source_line_start": 28,
"title": "`active` requires resolved cognition-delta evidence"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 5,
"section_id": "how-it-works",
"source_line_end": 44,
"source_line_start": 43,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Two entrypoints, run over the first-wave fixture and runregistry_bundle over the exported bundle, both call buildresult to run the ordered pipeline.** The runtime lives in src/microcosmcore/components/cognitiveoperator_registry.py. There are two entrypoints. run validates the first-wave fixture and includes the negative cases; runregistrybundle validates the exported bundle and makes the copied source-body manifest mandatory. Both call buildresult, which runs the pipeline in this order.",
"level": 4,
"order": 6,
"section_id": "two-entrypoints-over-build-result",
"source_line_end": 56,
"source_line_start": 45,
"title": "Two entrypoints over `_build_result`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The pipeline loads inputs, checks copied bodies, runs the secret scan, validates operator rows, and exercises negatives; status is pass only when all four conditions hold.** First, loadpayloads reads the three public inputs: operator_registry.json, operatorstandard.json, and dogfoodindex.json (plus the eight negative inputs in fixture mode). Second, sourcemodulemanifestresult checks the copied source bodies carried by the exported bundle. Third, scan_paths runs the secret-exclusion pass over every input. Fourth, positivefindings validates the operator rows. Fifth, negativefindings exercises the declared rejection cases. The status is pass only when there are no positive findings, no expec\u2026",
"level": 4,
"order": 7,
"section_id": "the-five-step-pipeline-and-pass-rule",
"source_line_end": 71,
"source_line_start": 57,
"title": "The five-step pipeline and pass rule"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "positivefindings checks each row's shape and status, resolves every active operator's dogfood result record and evidence commands, and flags any policy that sets an OVERCLAIMKEYS flag true. positivefindings carries the core contract. For each operator row it requires all eleven fields in OPERATORREQUIREDFIELDS and a status drawn from candidate, active, or retired. For each active operator it looks up a dogfood row keyed by operator_id, requires the nine fields in DOGFOODREQUIREDRECEIPTFIELDS, and calls dogfoodreceiptref_resolves to confirm the result record ref points at a real dogfood row. It then calls recorddogfoodevidenceresolution_findings, which parses each evidence surface as a comma\u2026",
"level": 4,
"order": 8,
"section_id": "positive-findings-carries-the-core-contract",
"source_line_end": 90,
"source_line_start": 72,
"title": "`_positive_findings` carries the core contract"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "sourcemodulemanifestresult proves the bundle's copied bodies by class, module_count, sha256 across three digest fields, and declared anchors, recording ids and hashes but never the body text.** sourcemodulemanifestresult is the body-floor check used by the bundle entrypoint. The exported bundle at examples/cognitiveoperatorregistry/exportedcognitiveoperatorregistrybundle copies the source registry, standard, and projection-tool bodies. The manifest must classify each as copiednonsecretmacrobody, keep bodyinreceipt false, match its declared module_count, and for every row: use a material class, copy the body into source_modules, hash to the recorded sha256 across all three digest fields, and\u2026",
"level": 4,
"order": 9,
"section_id": "the-bundle-body-floor-manifest-check",
"source_line_end": 135,
"source_line_start": 91,
"title": "The bundle body-floor manifest check"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 10,
"section_id": "negative-cases",
"source_line_end": 137,
"source_line_start": 136,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The fixture ships eight perturbed inputs, each of which negative_findings must observe and map to a specific error code, from missing fields to unresolvable dogfood evidence.** The fixture ships eight perturbed inputs, and a pass requires every one to be observed and recorded. negativefindings reads each and emits the matching error code. Missing operator fields give COGOPMISSINGREQUIRED_FIELD. An active operator with no backing dogfood row gives COGOPACTIVEWITHOUT_DOGFOOD. A dogfood row with no cognition-delta evidence gives COGOPDOGFOODMISSINGCOGNITIONDELTA. Two operators sharing a slug or a near-identical claim with no recorded accretion decision give COGOPOPERATORSPRAWL. A policy that g\u2026",
"level": 4,
"order": 11,
"section_id": "eight-perturbed-inputs-and-their-error-codes",
"source_line_end": 156,
"source_line_start": 138,
"title": "Eight perturbed inputs and their error codes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "prior-art-grounding",
"source_line_end": 158,
"source_line_start": 157,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The design draws on cognitive work analysis, W3C PROV provenance, JSON Schema shape checks, and Open Policy Agent's data/policy separation, applied as a public contract check.** This component sits in the lineage of cognitive work analysis, provenance tracking, schema validation, and policy-gated registries. Cognitive Work Analysis, summarized in this information-systems design overview, studies reusable moves in complex sociotechnical work. W3C PROV connects a claim to the activities, agents, and evidence used to judge it. JSON Schema is the required-shape pattern behind the operator-row check. [Open Policy Agent](https://www.openpolicyagent.org/docs/latest) is a precedent for keeping po\u2026",
"level": 4,
"order": 13,
"section_id": "cognitive-work-analysis-prov-schema-policy",
"source_line_end": 175,
"source_line_start": 159,
"title": "Cognitive work analysis, PROV, schema, policy"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run the first-wave fixture into disposable output from the Microcosm root: Run the exported bundle through the same component: Then run the focused tests:",
"level": 2,
"order": 14,
"section_id": "validation-receipt-path",
"source_line_end": 198,
"source_line_start": 176,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the public operator contract replayed from public inputs with every shape check holding and every rejection observed, but does not certify the operators as correct.** A pass means the public operator contract replayed from public inputs: every shape check held, every expected rejection was observed, and no body text reached the result records. It does not certify the operators as generally correct.",
"level": 4,
"order": 15,
"section_id": "what-a-pass-means-here",
"source_line_end": 208,
"source_line_start": 199,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "claim-ceiling",
"source_line_end": 210,
"source_line_start": 209,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest supported claim is that the operator contract is replayable from public inputs, with the checker rejecting bad shape, unresolved evidence, sprawl, voice, overclaim, and leakage.** The strongest claim the evidence supports is that the public operator contract is replayable from public inputs. The checker rejects missing shape, missing or unresolved dogfood evidence, near-duplicate sprawl, operator-voice authority, authority overclaim, and private-source leakage, and it proves the copied source bodies by hash and anchor while keeping their text out of the result records.",
"level": 4,
"order": 17,
"section_id": "the-strongest-supported-claim",
"source_line_end": 221,
"source_line_start": 211,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component is source-linked only, mutates no operators, proves no operator correct beyond its dogfood record, and excludes voice export, external model access, mutation, or launch.** The proof boundary stops there. This component does not become source authority for the operator registry, does not mutate operators, and does not establish that any listed operator is correct beyond its own dogfood record. Its scope limit excludes private operator bodies, operator-voice export, external model access, source-file changes, and launch or public sharing. If the focused tests report a copied-body mismatch, that is a source-copy repair, not a change to this reader page.",
"level": 4,
"order": 18,
"section_id": "where-the-proof-boundary-stops",
"source_line_end": 234,
"source_line_start": 222,
"title": "Where the proof boundary stops"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/cognitive_operator_registry.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public cognitive-operator registry fixture validates operator-shape rows, active-operator dogfood result records, anti-sprawl decisions, copied source registry/standard/tool bodies, and scope limits without becoming operator source authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "cognitive_operator_registry.py",
"path": "src/microcosm_core/organs/cognitive_operator_registry.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public cognitive-operator registry validation, exported source-module manifest checks, negative-case enforcement, and scope limits.",
"symbols": [
"run",
"run_registry_bundle",
"_source_module_manifest_result",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/cognitive_operator_registry.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/cognitive_operator_registry.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/cognitive_operator_registry.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Cognitive Operator Registry is the public evidence membrane for reusable cognition as typed system. It validates public operator rows, dogfood result records with cognition-delta evidence, negative cases for missing fields, missing dogfood, sprawl, operator-voice claims, authority overclaims, and private-source leakage, then checks copied source registry, standard, and validator bodies by digest while keeping bodies out of result records.",
"summary_line": "The public cognitive-operator registry fixture validates operator-shape rows, active-operator dogfood result records, anti-sprawl decisions, copied source registry/standard/tool bodies, and scope limits without becoming operator source authority.",
"title": "Cognitive Operator Registry"
},
{
"anchor": "paper-module-agent-closeout-faithfulness-audit",
"authority_ceiling": "Public fixture and exported-bundle result records only; no arbitrary live commit proof, work log mutation or closure, live Git mutation, external model access, launch-scope decision, broad completion certification, or pytest-pass claim without explicit exit-zero status.",
"body_html": "
Completion language needs observable evidence
\n
The finish-line sentence an agent writes is admitted only claim by claim, once observable git and pytest evidence backs each part of it.agent_closeout_faithfulness_audit checks the sentence an agent writes when it finishes a task: "I committed the change, closed the ledger item, and the test passed." It runs the supplied public fixture through real git and pytest subprocesses and admits each claim only when an evidence object it can observe actually backs it.
\n
A fixture repository is copied into a throwaway directory, committed with real git, and its HEAD is read back. A declared test span is run, and the exit code alone decides whether the span passed.
\n
Four metadata-only records carry the result
\n
It emits only metadata-only records: a result file, a board, a validation record, and an exported bundle validation record. It emits metadata-only result records: a result file, a board, a validation record, and an exported bundle validation record.
\n
Purpose
\n
A done report is only text
\n
A finished-work report is just text: the commit, the ledger row, and the "passed" may or may not be real, and an audit that trusts the wording certifies nothing. When an agent reports that work is done, the report is text. The commit may or may not exist. The ledger row may or may not be there. "The test passed" can mean the test ran, or it can mean nothing was checked. A loose audit that trusts the wording certifies all of those as done and learns nothing.
\n
Ran and passed are separate facts
\n
It asks one question per fixture: is each completion claim backed by an evidence object that exists, and is "passed" backed by an explicit exit-zero check rather than by the phrasing of the claim? The component answers one question over a fixed fixture: is each completion claim backed by an evidence object that genuinely exists, and is a "passed" claim backed by an explicit exit-zero check rather than by the phrasing of the claim? The distinction it defends is narrow. "The span ran" and "the span passed" are separate facts. A completion sentence that conflates them is the exact failure mode here.
\n
How it works
\n
Two entrypoints forward to evaluate
\n
run and run_agent_closeout_bundle are the two entrypoints, and both forward to evaluate, which carries the whole mechanism. The runtime is src/microcosm_core/organs/agent_closeout_faithfulness_audit.py. run and run_agent_closeout_bundle are the two entrypoints; both forward to evaluate, which carries the mechanism.
\n
evaluate runs in this order:
\n
Load closeout_claims.json and fixture_ledger.json. If public_fixture_repo is not a directory, emit CLOSEOUT_PUBLIC_FIXTURE_REPO_MISSING and stop.
_prepare_public_fixture_repo copies the fixture repo into a temporary directory and runs git init, two git config calls, git add ., and git commit, then git rev-parse HEAD. Six subprocesses in total. The observed HEAD is returned. Each subprocess row is a return code and two sha256 digests from _sha256_text, never raw output. _run_subprocess builds the five setup rows in that metadata-only shape.
_select_pytest_python picks the first interpreter that answers pytest --version with exit zero, so the span runs under an interpreter that actually has pytest installed.
Walk the claim rows. A commit claim verifies only when its commit_ref is HEAD or the observed head, else CLOSEOUT_FAKE_COMMIT_CLAIM. A task_ledger_cap claim verifies only when its cap_id is present in fixture_ledger.json, else CLOSEOUT_FAKE_CAP_CLAIM. A pytest_span claim runs python -m pytest <nodeid> -q inside the temporary repo: span_ran is true when the process exits 0 or 1, and passed is true only on exit 0. A span that did not run emits CLOSEOUT_FAKE_TEST_CLAIM. A claim whose expected_pass is true but whose pass_status_checked is false, or whose span did not exit zero, emits CLOSEOUT_PYTEST_PASS_STATUS_NOT_CHECKED.
\n
Function
Role
evaluate
_prepare_public_fixture_repo
Copies and commits the fixture repo, reads HEAD, records six metadata-only subprocess rows
_select_pytest_python
Chooses an interpreter that has pytest before running the span
evaluate_negative_case
Mutates one claim row to force a named rejection
_semantic_closeout_contract_findings
Re-checks claims against the fixture test and arithmetic source on the negative path
run / run_agent_closeout_bundle
First-wave and exported-bundle entrypoints over the shared evaluator
\n
A pass counts pass status separately
\n
A pass records five counters, where verified means the evidence object exists or the span ran, and pass status is tallied separately only when the exit-zero check held. A pass records verified_claim_count, git_subprocess_count, pytest_span_ran_count, pytest_pass_status_checked_count, and head_verified_by_subprocess. Here verified means the evidence object exists or the span ran. Pass status is counted separately, only when the exit-zero check held. That keeps a referenced test distinct from a passed test in the record itself.
\nDiagram of the mechanism (8 steps).Diagram source & refs
evaluate_negative_case mutates one claim row at a time and confirms the matching rejection fires. The four cases named in EXPECTED_NEGATIVE_CASES: evaluate_negative_case mutates one claim row at a time and confirms the matching rejection fires.
\n
fake_commit_claim: rewrites the commit ref to all zeros, expects CLOSEOUT_FAKE_COMMIT_CLAIM.
fake_cap_claim: points the cap at an id missing from the fixture ledger, expects CLOSEOUT_FAKE_CAP_CLAIM.
fake_test_claim: names a test node that does not exist, expects CLOSEOUT_FAKE_TEST_CLAIM.
unchecked_pass_claim: sets pass_status_checked to false, expects CLOSEOUT_PYTEST_PASS_STATUS_NOT_CHECKED.
\n
Losing a case loses the proof
\n
If any of these rejections stops appearing in the observed cases, the component no longer proves it refuses fabricated commit, cap, test-node, or unchecked-pass claims. If these stop appearing in the observed cases, the component no longer proves that it refuses fabricated commit, cap, test-node, or unchecked-pass claims.
\n
Prior Art Grounding
\n
Claim, artifact, and evidence stay separate
\n
The posture is claim-then-evidence verification drawn from FEVER, TruthfulQA, and the artifact-evaluation tradition: commit, ledger, and span refs must line up with observed objects before completion language is admitted. The posture is claim-then-evidence verification rather than trust in fluent summary text. FEVER framed fact extraction and verification as a task over cited evidence. TruthfulQA showed that fluent answers can be untrue without an explicit truthfulness check. The wider artifact-evaluation tradition keeps a claim, its artifact, and its validation evidence as three separate things. This component applies that separation to agent completion: commit refs, ledger refs, and test spans must line up with observed objects before the completion language is admitted.
A focused-test pass means completion language was checked against fixture evidence across seven named witnesses and all four negative classes; the corpus check only proves the bundle and instance agree. A pass on the focused test means completion language was checked against fixture evidence: three verified claims, at least five git subprocess witnesses, one pytest witness, one span that ran, one checked pass status, head_verified_by_subprocess true, and observation of all four negative classes. The corpus check only proves that the source record and the generated instance agree.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
On the supplied fixture repo, ledger, and completion claims, commit, cap, and span references are checked against observed git and pytest evidence, and a pass is admitted only after an explicit exit-zero check. The strongest claim: over the supplied fixture repo, fixture ledger, and completion claims, this component proves that commit, cap, and span references are checked against observed git and pytest evidence, and that a pass claim is admitted only after an explicit exit-zero check. The proof boundary is that fixture and its records, nothing wider.
\n
What stays refused past that line
\n
Past that line it proves nothing: no live commit landed, no ledger mutation, no git, provider, or source change, no launch, and no turning a span that merely ran into one that passed. It refuses everything past that line. It does not establish any live commit landed, close or mutate ledger work, mutate git, use external model services, change source files, include launch operations, certify completion prose in general, or turn a span that merely ran into a span that passed. The scope limit is public fixture evidence existence and pytest span witnessing only. AUTHORITY_CEILING records those refusals in machine-readable form.
\n",
"components": [
{
"anchor": "component-agent_closeout_faithfulness_audit",
"component_id": "agent_closeout_faithfulness_audit",
"public_label": "Agent Completion Faithfulness Audit"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 2
},
"link_count": 7,
"links": [
{
"id": "agent_closeout_faithfulness_audit",
"kind": "organ",
"kind_label": "component",
"label": "Agent Completion Faithfulness Audit",
"relation": "explains",
"url": "component-agent-closeout-faithfulness-audit.html"
},
{
"id": "mechanism.agent_closeout_faithfulness_audit.validates_closeout_evidence_claims",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates completion evidence claims",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-closeout-faithfulness-audit-validates-closeout-evidence-claims"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.durable_agent_work_landing_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Durable Agent Work-Landing Replay",
"relation": "depends on",
"url": "paper-module-durable-agent-work-landing-replay.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.agent_closeout_faithfulness_audit",
"reader_packet": {
"abstract": "The public completion-faithfulness fixture checks commit, ledger-cap, and pytest-span claims with real git/pytest subprocess witnesses while refusing unchecked pass overclaims.",
"authority_ceiling": "Public fixture and exported-bundle result records only; no arbitrary live commit proof, work log mutation or closure, live Git mutation, external model access, launch-scope decision, broad completion certification, or pytest-pass claim without explicit exit-zero status.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.agent_closeout_faithfulness_audit.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-agent-closeout-faithfulness-audit-diagram-1",
"html_url": "docs/paper-module-agent-closeout-faithfulness-audit.html#paper-module-agent-closeout-faithfulness-audit-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Claims[\"completion_claims.json commit, cap, span\"] --> Eval[\"evaluate()\"] Ledger[\"fixture_ledger.json cap rows\"] --> Eval Repo[\"public_fixture_repo\"] --> Prep[\"_prepare_public_fixture_repo git init..commit, rev-parse HEAD\"] Prep --> Eval Eval --> Pytest[\"python -m pytest nodeid -q\"] Pytest --> Eval Eval --> Pass[\"pass: verified counts, metadata-only record\"] Eval --> Reject[\"reject: fake commit, cap, test, unchecked pass\"]",
"source_format": "mermaid",
"source_hash": "sha256:f8e9d34d0293b715f3bef43ae984d4dcc2cc7247e9616d29773b485e5e9ebb03",
"source_line_end": 117,
"source_line_start": 107
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.durable_agent_work_landing_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Durable Agent Work-Landing Replay",
"relation": "depends on",
"url": "paper-module-durable-agent-work-landing-replay.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "agent_closeout_faithfulness_audit",
"kind": "organ",
"kind_label": "component",
"label": "Agent Completion Faithfulness Audit",
"relation": "explains",
"url": "component-agent-closeout-faithfulness-audit.html"
},
{
"id": "mechanism.agent_closeout_faithfulness_audit.validates_closeout_evidence_claims",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates completion evidence claims",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-closeout-faithfulness-audit-validates-closeout-evidence-claims"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.agent_closeout_faithfulness_audit",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 207,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_closeout_faithfulness_audit.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "FEVER",
"section_id": "claim-artifact-and-evidence-stay-separate",
"url": "https://arxiv.org/abs/1803.05355"
},
{
"kind": "external_reference",
"label": "TruthfulQA",
"section_id": "claim-artifact-and-evidence-stay-separate",
"url": "https://arxiv.org/abs/2109.07958"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The finish-line sentence an agent writes is admitted only claim by claim, once observable git and pytest evidence backs each part of it.** agentcompletionfaithfulness_audit checks the sentence an agent writes when it finishes a task: \"I committed the change, closed the ledger item, and the test passed.\" It runs the supplied public fixture through real git and pytest subprocesses and admits each claim only when an evidence object it can observe actually backs it.",
"level": 4,
"order": 1,
"section_id": "closeout-language-needs-observable-evidence",
"source_line_end": 12,
"source_line_start": 3,
"title": "Completion language needs observable evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evidence: a real git commit, a read-back HEAD, and a test span whose exit code alone decides whether it passed.** The component does not read the completion text and score it against a rubric. A fixture repository is copied into a throwaway directory, committed with real git, and its HEAD is read back. A declared test span is run, and the exit code alone decides whether the span passed.",
"level": 4,
"order": 2,
"section_id": "the-audit-rebuilds-git-and-test-evidence",
"source_line_end": 22,
"source_line_start": 13,
"title": "The audit rebuilds Git and test evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It emits only metadata-only records: a result file, a board, a validation record, and an exported bundle validation record.** It emits metadata-only result records: a result file, a board, a validation record, and an exported bundle validation record.",
"level": 4,
"order": 3,
"section_id": "four-body-free-records-carry-the-result",
"source_line_end": 28,
"source_line_start": 23,
"title": "Four metadata-only records carry the result"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 30,
"source_line_start": 29,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A finished-work report is just text: the commit, the ledger row, and the \"passed\" may or may not be real, and an audit that trusts the wording certifies nothing.** When an agent reports that work is done, the report is text. The commit may or may not exist. The ledger row may or may not be there. \"The test passed\" can mean the test ran, or it can mean nothing was checked. A loose audit that trusts the wording certifies all of those as done and learns nothing.",
"level": 4,
"order": 5,
"section_id": "a-done-report-is-only-text",
"source_line_end": 39,
"source_line_start": 31,
"title": "A done report is only text"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It asks one question per fixture: is each completion claim backed by an evidence object that exists, and is \"passed\" backed by an explicit exit-zero check rather than by the phrasing of the claim?** The component answers one question over a fixed fixture: is each completion claim backed by an evidence object that genuinely exists, and is a \"passed\" claim backed by an explicit exit-zero check rather than by the phrasing of the claim? The distinction it defends is narrow. \"The span ran\" and \"the span passed\" are separate facts. A completion sentence that conflates them is the exact failure mode here.",
"level": 4,
"order": 6,
"section_id": "ran-and-passed-are-separate-facts",
"source_line_end": 51,
"source_line_start": 40,
"title": "Ran and passed are separate facts"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 53,
"source_line_start": 52,
"title": "How it works"
},
{
"block_kinds": [
"list",
"paragraph",
"table"
],
"excerpt": "run and runagentcompletionbundle are the two entrypoints, and both forward to evaluate, which carries the whole mechanism.** The runtime is src/microcosmcore/components/agentcompletionfaithfulnessaudit.py. run and runagentcompletion_bundle are the two entrypoints; both forward to evaluate, which carries the mechanism. evaluate runs in this order: Load completionclaims.json and fixtureledger.json. If publicfixturerepo is not a directory, emit CompletionPUBLICFIXTUREREPOMISSING and stop. preparepublicfixturerepo copies the fixture repo into a temporary directory and runs git init, two git config calls, git add ., and git commit, then git rev-parse HEAD. Six subprocesses in total. The observed\u2026",
"level": 4,
"order": 8,
"section_id": "two-entrypoints-forward-to-evaluate",
"source_line_end": 94,
"source_line_start": 54,
"title": "Two entrypoints forward to `evaluate`"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "**A pass records five counters, where verified means the evidence object exists or the span ran, and pass status is tallied separately only when the exit-zero check held. A pass records verifiedclaim_count, gitsubprocesscount, pytestspanrancount, pytestpassstatuschecked_count, and headverifiedby_subprocess. Here verified means the evidence object exists or the span ran. Pass status is counted separately, only when the exit-zero check held. That keeps a referenced test distinct from a passed test in the record itself.",
"level": 4,
"order": 9,
"section_id": "a-pass-counts-pass-status-separately",
"source_line_end": 118,
"source_line_start": 95,
"title": "A pass counts pass status separately"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 10,
"section_id": "negative-cases",
"source_line_end": 120,
"source_line_start": 119,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "evaluatenegative_case mutates one claim row at a time and confirms the matching rejection fires. The four cases named in EXPECTEDNEGATIVE_CASES: evaluatenegativecase mutates one claim row at a time and confirms the matching rejection fires. fakecommitclaim: rewrites the commit ref to all zeros, expects CompletionFAKECOMMIT_CLAIM. fakecapclaim: points the cap at an id missing from the fixture ledger, expects CompletionFAKECAP_CLAIM. faketestclaim: names a test node that does not exist, expects CompletionFAKETEST_CLAIM. uncheckedpassclaim: sets passstatuschecked to false, expects CompletionPYTESTPASSSTATUSNOT_CHECKED.",
"level": 4,
"order": 11,
"section_id": "one-mutated-row-per-rejection",
"source_line_end": 135,
"source_line_start": 121,
"title": "One mutated row per rejection"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**If any of these rejections stops appearing in the observed cases, the component no longer proves it refuses fabricated commit, cap, test-node, or unchecked-pass claims.** If these stop appearing in the observed cases, the component no longer proves that it refuses fabricated commit, cap, test-node, or unchecked-pass claims.",
"level": 4,
"order": 12,
"section_id": "losing-a-case-loses-the-proof",
"source_line_end": 142,
"source_line_start": 136,
"title": "Losing a case loses the proof"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "prior-art-grounding",
"source_line_end": 144,
"source_line_start": 143,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The posture is claim-then-evidence verification drawn from FEVER, TruthfulQA, and the artifact-evaluation tradition: commit, ledger, and span refs must line up with observed objects before completion language is admitted.** The posture is claim-then-evidence verification rather than trust in fluent summary text. FEVER framed fact extraction and verification as a task over cited evidence. TruthfulQA showed that fluent answers can be untrue without an explicit truthfulness check. The wider artifact-evaluation tradition keeps a claim, its artifact, and its validation evidence as three separate things. This component applies that separation to agent completion: commit refs, ledger refs, and t\u2026",
"level": 4,
"order": 14,
"section_id": "claim-artifact-and-evidence-stay-separate",
"source_line_end": 159,
"source_line_start": 145,
"title": "Claim, artifact, and evidence stay separate"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "surfaces:",
"level": 2,
"order": 15,
"section_id": "validation-receipt-path",
"source_line_end": 169,
"source_line_start": 160,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A focused-test pass means completion language was checked against fixture evidence across seven named witnesses and all four negative classes; the corpus check only proves the bundle and instance agree.** A pass on the focused test means completion language was checked against fixture evidence: three verified claims, at least five git subprocess witnesses, one pytest witness, one span that ran, one checked pass status, headverifiedby_subprocess true, and observation of all four negative classes. The corpus check only proves that the source record and the generated instance agree.",
"level": 4,
"order": 16,
"section_id": "what-a-focused-test-pass-certifies",
"source_line_end": 181,
"source_line_start": 170,
"title": "What a focused-test pass certifies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "claim-ceiling",
"source_line_end": 183,
"source_line_start": 182,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**On the supplied fixture repo, ledger, and completion claims, commit, cap, and span references are checked against observed git and pytest evidence, and a pass is admitted only after an explicit exit-zero check.** The strongest claim: over the supplied fixture repo, fixture ledger, and completion claims, this component proves that commit, cap, and span references are checked against observed git and pytest evidence, and that a pass claim is admitted only after an explicit exit-zero check. The proof boundary is that fixture and its records, nothing wider.",
"level": 4,
"order": 18,
"section_id": "the-strongest-supported-claim",
"source_line_end": 194,
"source_line_start": 184,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Past that line it proves nothing: no live commit landed, no ledger mutation, no git, provider, or source change, no launch, and no turning a span that merely ran into one that passed.** It refuses everything past that line. It does not establish any live commit landed, close or mutate ledger work, mutate git, use external model services, change source files, include launch operations, certify completion prose in general, or turn a span that merely ran into a span that passed. The scope limit is public fixture evidence existence and pytest span witnessing only. scope_limit records those refusals in machine-readable form.",
"level": 4,
"order": 19,
"section_id": "what-stays-refused-past-that-line",
"source_line_end": 206,
"source_line_start": 195,
"title": "What stays refused past that line"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/agent_closeout_faithfulness_audit.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public completion-faithfulness fixture checks commit, ledger-cap, and pytest-span claims with real git/pytest subprocess witnesses while refusing unchecked pass overclaims."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "agent_closeout_faithfulness_audit.py",
"path": "src/microcosm_core/organs/agent_closeout_faithfulness_audit.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public completion evidence validation, external subprocess witnesses, negative cases, and source-module manifest checks.",
"symbols": [
"run",
"run_agent_closeout_bundle",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/agent_closeout_faithfulness_audit.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_closeout_faithfulness_audit.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_closeout_faithfulness_audit.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Agent Completion Faithfulness Audit is the public fixture witness for completion evidence language. It builds a fixture repo, verifies commit and HEAD evidence with git subprocesses, runs pytest for the declared span, checks cap claims against a fixture ledger, validates a copied source-module manifest for agent experience diagnostics, and writes bounded result records that keep source bodies out.",
"summary_line": "The public completion-faithfulness fixture checks commit, ledger-cap, and pytest-span claims with real git/pytest subprocess witnesses while refusing unchecked pass overclaims.",
"title": "Agent Completion Faithfulness Audit"
},
{
"anchor": "paper-module-cold-reader-route-map",
"authority_ceiling": "Public route-map fixture, exported-bundle result records, and copied source cold-entry body evidence only; no route registry control, external model access, source-file changes, launch-scope decision, non-public data equivalence, trading or financial decisions, or whole-system correctness.",
"body_html": "
Route map checks
\n
The route map checks a newcomer's first ten minutes, confirming each first-run step's command, docs reference, and evidence record resolves against real source, not prose.cold_reader_route_map checks the guided path a newcomer follows in their first ten minutes with Microcosm. It reads a public route map whose rows bind each first-run step to a runnable command, a public documentation reference, and an evidence record, then it confirms that every one of those bindings resolves against real source instead of prose.
\n
Takes three fixture
\n
It takes three fixture files plus an optional copied-source manifest and emits metadata-only findings, a route board, validation records, and a card carrying ids, refs, and digests only. The component takes three fixture files (route_map.json, route_receipts.json, route_policy.json) plus an optional copied-source manifest, and emits deterministic findings, a route board, validation records, and a compact card. Every emitted record is metadata-only: it carries route ids, commands, refs, digests, and body_in_receipt: false markers, never inlined source or live state.
\n
Source replay is what lifts it above field-presence checking: an unrunnable command, an unresolved docs heading, or a non-passing evidence ref each blocks the route. The discipline that makes it more than a field-presence checker is source replay. A route that promises a command the system does not actually run is blocked, as is a docs reference that resolves to no heading and an evidence ref that does not open a passing record. That is the failure a hand-written quick-start guide slides into the moment the commands change underneath it.
\n
Purpose
\n
What to run first, with evidence
\n
A cold reader should not reverse-engineer the path from a README, so the route map answers what to run first and what evidence shows that path is wired. A cold technical reader should not have to reverse-engineer the product path from a long README or a raw tree of result files. The route map answers one question: what do I run first, and what evidence shows that path is wired?
\n
Each row as a claim-boundary check
\n
A prose route map lists commands nothing re-checks; this one treats each row as a claim-boundary check so the trusted sequence stays tied to re-runnable source. A loose route map answers that question with prose. It lists impressive commands, and nothing checks that they still exist or that their supporting records still pass. The route map here treats each row as a claim-boundary check, so the sequence a newcomer trusts stays tied to source that a test can re-run.
\n
How it works
\n
Fixture and exported-bundle entrypoints
\n
run builds the fixture with negative cases and writes the record set, while run_route_map_bundle runs the same builder with negatives off and the copied-source manifest required.run is the fixture entrypoint. It calls _build_result with negative cases included, then writes the result, board, and validation records, plus a cold_reader_route_map_fixture_acceptance_v1 record when the caller passes acceptance_out. run_route_map_bundle is the exported-bundle entrypoint. It runs the same builder with negative cases off and the copied-source manifest required, so the bundle proves replay shape from public inputs alone.
\n
_build_result joins four checks and a scan
\n
_build_result runs four independent checks plus a secret-exclusion scan, passing only when no check files a blocking finding and the scan reports no blocking hit._build_result assembles four independent checks plus a secret-exclusion scan over the inputs, and passes only when none of the four checks files a blocking finding and the scan reports no blocking hit.
\n
Required routes and front-door ordering
\n
_positive_findings requires each route to name a command, docs refs, and a result record ref, and the front-door sequence to match live first-screen strings in ordinal order._positive_findings walks the routes named in route_policy. Each required route must exist, name a command, carry docs_refs, and point at a result record ref. The first-run sequence must start with the three front-door routes in FRONT_DOOR_ROUTE_COMMANDS (tour_project, status_card, proof_lab), those front-door commands must match the live first-screen strings, and the whole sequence must be present and ordinal sorted. A gap, a missing field, or a drifted front-door command becomes a finding.
\n
Replaying commands, refs, and signals against source
\n
_route_source_replay_result builds a source corpus and blocks any command token, docs ref, result record ref, or claimed signal that does not resolve against real public source._route_source_replay_result is the replay pass. It builds one text corpus from the copied source modules, a fixed set of public source refs, every route's docs refs, and every result record ref, then it tests each route against that corpus. _command_support splits a command into material tokens and blocks it if a token appears nowhere in the corpus. _document_ref_result requires each docs ref to resolve to a real public file and, when the ref names a heading anchor, to a real heading. _receipt_ref_result opens each result record ref and blocks it unless the file is JSON with status equal to pass. _mechanism_signal_support takes the human-readable signals a row claims to show and blocks the row if their terms are absent from the corpus.
\n
The copied-body floor for the bundle
\n
_source_module_import_result recomputes each copied body's digest against the manifest sha256, checks provenance anchors survive, and requires a copied_non_secret_macro_body exact-copy declaration._source_module_import_result governs the copied-body floor used by the exported bundle. For each row in source_module_manifest.json it recomputes the file digest, compares it to the manifest sha256, checks that required provenance anchors survive in the copied text, and confirms the row declares copied_non_secret_macro_body with a material class and an exact_copy relation. A digest mismatch, a missing anchor, a body marked for a record, or an unsupported class each files a finding.
\n
Negative findings evaluates
\n
_negative_findings evaluates the five perturbed inputs and records the error code each one must raise, listed under Negative cases below._negative_findings evaluates the five perturbed inputs and records the error code each one must raise, listed under Negative cases below.
\nDiagram of the mechanism (7 steps).Diagram source & refs
flowchart TD Inputs["route_map, route_receipts, route_policy, manifest"] --> Positive["_positive_findings required routes, commands, docs refs, result record refs, order"] Inputs --> Replay["_route_source_replay_result commands and refs matched against the source corpus"] Inputs --> Modules["_source_module_import_result copied-body digests and anchors"] Inputs --> Negative["_negative_findings five rejection cases"] Positive --> Result["_build_result pass only when no blocking finding"] Replay --> Result Modules --> Result Negative --> Result Result --> Records["run and run_route_map_bundle write metadata-only records"]
\n
Function to role:
\n
Function
Role
run
Fixture entrypoint; builds with negative cases, writes the record set.
run_route_map_bundle
Exported-bundle entrypoint; manifest required, no negatives.
_build_result
Joins the four checks; sets pass only with no blocking finding.
_positive_findings
Required routes name a command, docs refs, result record ref, ordinal order.
_route_source_replay_result
Replays commands, docs refs, result record refs, signals against source.
_source_module_import_result
Verifies copied-body digests, anchors, and import class.
_negative_findings
Confirms the five perturbed inputs raise their error codes.
\n
Negative cases
\n
Expected negative cases
\n
EXPECTED_NEGATIVE_CASES names the five inputs the fixture must reject and the code each one raises.EXPECTED_NEGATIVE_CASES names the five inputs the fixture must reject and the code each one raises:
\n
missing_command_ref raises COLD_ROUTE_MISSING_COMMAND_REF: a route with no command cannot guide anyone.
missing_receipt_ref raises COLD_ROUTE_MISSING_RECEIPT_REF: a route with no evidence ref is not backed.
route_sequence_gap raises COLD_ROUTE_SEQUENCE_GAP: a sequence that does not start from tour_project is malformed.
release_overclaim raises COLD_ROUTE_AUTHORITY_OVERCLAIM: a row that sets a key in OVERCLAIM_KEYS, such as release_authorized or whole_system_correctness_claim, is refused.
private_source_leakage raises COLD_ROUTE_PRIVATE_SOURCE_FORBIDDEN: a row carrying a key in FORBIDDEN_PRIVATE_KEYS, such as private_source_body, proof_body, or provider_payload_body, is refused so a route card cannot become a disclosure channel.
\n
If fixture run
\n
If the fixture run does not observe one of these five, _build_result marks the run blocked for the missing case. If the fixture run does not observe one of these five, _build_result marks the run blocked for the missing case.
\n
Reader-Specific Evidence Routing
\n
Command order and evidence order kept together
\n
The route map keeps a command order and an evidence order, so a reader sees the compact sequence first yet keeps each step's result record-ref and scope-limit context in drilldowns. The route map keeps two orders: the command order and the evidence order. A reader sees the compact command sequence first, then can open full drilldowns without losing the result record-ref and scope-limit context attached to each step.
\n
The accepted path is:
\n
plectis hello <project>
plectis tour --card <project>
plectis status --card <project>
plectis authority --card
plectis workingness --card
plectis legibility-scorecard
Read selected_route_id from the compact tour or status card.
Full drilldowns stay available after the compact path is visible:
\n
plectis tour <project> for the full first-screen route tree.
plectis compile <project> for project-state materialization detail.
plectis proof-lab --out /tmp/microcosm-proof-lab for proof-lab component records.
plectis serve <project> --host 127.0.0.1 --port 8765 for the local observatory.
\n
Compact board must match the terminal row
\n
A browser-first reader follows the same order, and the compact board must carry the same command ref, evidence ref, class, and scope limit as the terminal row, or it is a preview. A browser-first reader follows the same order in compressed form, and the compact board must carry the same command ref, evidence ref, evidence class, and scope limit as the terminal row, or it is a preview only.
\n
Prior Art Grounding
\n
Diataxis and literate programming lineage
\n
The design follows Diataxis's task-shaped documentation and Knuth's literate programming, borrowing the ordered-reader pattern without turning documentation sequence into source authority. The component follows documentation systems that treat reader state and task shape as first-class. Diataxis separates tutorials, how-to guides, reference, and explanation so a reader is not pushed through one undifferentiated pile. Knuth's literate programming is the older anchor for writing an executable system for human comprehension as well as machine execution. The route map borrows the ordered-reader pattern of first command, evidence ref, evidence class, scope limit, and next drilldown, without turning documentation sequence into source authority.
\n
Validation Result record Path
\n
From microcosm-substrate/, reproduce this page's proof boundary with temporary records:
\n
What a pass means here
\n
A pass means the eleven-route fixture, the exported bundle, and the copied source-module digests all validated with no blocking finding and stay consistent with the source record and generated views. A pass means the eleven-route fixture, the exported bundle, and the copied source-module digests all validated with no blocking finding, and that this page stays consistent with its source record, generated diagram, and Atlas card.
\n
Scope boundary
\n
Scope limit
\n
The narrow strongest claim
\n
The strongest claim is narrow: a first-run route map whose commands, docs refs, and evidence refs resolve against public source, with an ordinal front-door sequence and five named rejections. The strongest claim this component supports is narrow: a public first-run route map whose commands, docs refs, and evidence refs each resolve against real public source, whose front-door sequence is ordinal sorted, and whose five negative cases are rejected with named codes. Its AUTHORITY_CEILING sets route_registry_authority, source_mutation_authorized, release_authorized, provider_calls_authorized, private_data_equivalence_claim, and whole_system_correctness_claim all to false.
\n
What the route map cannot become
\n
It is not route registry control, mutates no source, calls no providers, excludes launch, and cannot promote the documentation sequence into proof beyond its public fixtures and records. It refuses to become route registry control, to change source files projects, to use external model services, or to include launch operations or public sharing. It does not establish the whole system correct, and it does not claim equivalence to any non-public data. The route map can tell a cold reader what to run first and which record bounds that run. It cannot promote the documentation sequence into proof beyond those public fixtures and records.
\n",
"components": [
{
"anchor": "component-cold_reader_route_map",
"component_id": "cold_reader_route_map",
"public_label": "Cold Reader Route Map"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 9
},
"link_count": 21,
"links": [
{
"id": "cold_reader_route_map",
"kind": "organ",
"kind_label": "component",
"label": "Cold Reader Route Map",
"relation": "explains",
"url": "component-cold-reader-route-map.html"
},
{
"id": "mechanism.cold_reader_route_map.validates_public_first_run_route_map",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public first run route map",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-cold-reader-route-map-validates-public-first-run-route-map"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.entry_and_reveal_route_readiness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "entry and reveal routes as bounded cold-reader readiness evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-entry-and-reveal-route-readiness-bundle"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[2:paper_module.agent_route_observability_runtime].depends_on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.doctrine_fact_claim_audit",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Doctrine Fact Claim Audit",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[48:paper_module.doctrine_fact_claim_audit].depends_on",
"url": "paper-module-doctrine-fact-claim-audit.html"
},
{
"id": "paper_module.first_screen_composition_root",
"kind": "paper_module",
"kind_label": "paper module",
"label": "First-Screen Composition Root",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[90:paper_module.first_screen_composition_root].depends_on",
"url": "paper-module-first-screen-composition-root.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[1:paper_module.navigation_hologram_route_plane].depends_on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.pattern_assimilation",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Assimilation",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[71:paper_module.pattern_assimilation].depends_on",
"url": "paper-module-pattern-assimilation.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[9:paper_module.pattern_binding_contract].depends_on",
"url": "paper-module-pattern-binding-contract.html"
},
{
"id": "paper_module.routing_anti_patterns_registry",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Routing Anti-Patterns Registry",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[57:paper_module.routing_anti_patterns_registry].depends_on",
"url": "paper-module-routing-anti-patterns-registry.html"
},
{
"id": "paper_module.self_ignorance_coverage_ledger",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Self-Ignorance Coverage Ledger",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[49:paper_module.self_ignorance_coverage_ledger].depends_on",
"url": "paper-module-self-ignorance-coverage-ledger.html"
},
{
"id": "paper_module.voice_to_doctrine_self_improvement_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Voice to Doctrine Self-Improvement Loop",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[56:paper_module.voice_to_doctrine_self_improvement_loop].depends_on",
"url": "paper-module-voice-to-doctrine-self-improvement-loop.html"
}
],
"paper_module_id": "paper_module.cold_reader_route_map",
"reader_packet": {
"abstract": "The public cold-reader route-map fixture validates first-run command order, docs refs, result record refs, scope limits, copied cold-entry source-module digests, and non-public-state exclusion without becoming route registry control.",
"authority_ceiling": "Public route-map fixture, exported-bundle result records, and copied source cold-entry body evidence only; no route registry control, external model access, source-file changes, launch-scope decision, non-public data equivalence, trading or financial decisions, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.cold_reader_route_map.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-cold-reader-route-map-diagram-1",
"html_url": "docs/paper-module-cold-reader-route-map.html#paper-module-cold-reader-route-map-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Inputs[\"route_map, route_receipts, route_policy, manifest\"] --> Positive[\"_positive_findings required routes, commands, docs refs, result record refs, order\"] Inputs --> Replay[\"_route_source_replay_result commands and refs matched against the source corpus\"] Inputs --> Modules[\"_source_module_import_result copied-body digests and anchors\"] Inputs --> Negative[\"_negative_findings five rejection cases\"] Positive --> Result[\"_build_result pass only when no blocking finding\"] Replay --> Result Modules --> Result Negative --> Result Result --> Records[\"run and run_route_map_bundle write metadata-only records\"]",
"source_format": "mermaid",
"source_hash": "sha256:dde9e9db9b2a2d896351cb403370a99a84ca243b3e24290f27a0bd2eb51da43d",
"source_line_end": 141,
"source_line_start": 130
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "cold_reader_route_map",
"kind": "organ",
"kind_label": "component",
"label": "Cold Reader Route Map",
"relation": "explains",
"url": "component-cold-reader-route-map.html"
},
{
"id": "mechanism.cold_reader_route_map.validates_public_first_run_route_map",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public first run route map",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-cold-reader-route-map-validates-public-first-run-route-map"
}
],
"relation": "explains"
},
{
"count": 10,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.entry_and_reveal_route_readiness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "entry and reveal routes as bounded cold-reader readiness evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-entry-and-reveal-route-readiness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.cold_reader_route_map",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 291,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/cold_reader_route_map.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Diataxis",
"section_id": "diataxis-and-literate-programming-lineage",
"url": "https://diataxis.fr/"
},
{
"kind": "external_reference",
"label": "literate programming",
"section_id": "diataxis-and-literate-programming-lineage",
"url": "https://cs.stanford.edu/~knuth/lp.html"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The route map checks a newcomer's first ten minutes, confirming each first-run step's command, docs reference, and evidence record resolves against real source, not prose.** coldreaderroute_map checks the guided path a newcomer follows in their first ten minutes with Microcosm. It reads a public route map whose rows bind each first-run step to a runnable command, a public documentation reference, and an evidence record, then it confirms that every one of those bindings resolves against real source instead of prose.",
"level": 4,
"order": 1,
"section_id": "route-map-checks",
"source_line_end": 13,
"source_line_start": 3,
"title": "Route map checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It takes three fixture files plus an optional copied-source manifest and emits metadata-only findings, a route board, validation records, and a card carrying ids, refs, and digests only.** The component takes three fixture files (routemap.json, routereceipts.json, route_policy.json) plus an optional copied-source manifest, and emits deterministic findings, a route board, validation records, and a compact card. Every emitted record is metadata-only: it carries route ids, commands, refs, digests, and bodyinreceipt: false markers, never inlined source or live state.",
"level": 4,
"order": 2,
"section_id": "takes-three-fixture",
"source_line_end": 25,
"source_line_start": 14,
"title": "Takes three fixture"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Source replay is what lifts it above field-presence checking: an unrunnable command, an unresolved docs heading, or a non-passing evidence ref each blocks the route.** The discipline that makes it more than a field-presence checker is source replay. A route that promises a command the system does not actually run is blocked, as is a docs reference that resolves to no heading and an evidence ref that does not open a passing record. That is the failure a hand-written quick-start guide slides into the moment the commands change underneath it.",
"level": 4,
"order": 3,
"section_id": "source-replay-is",
"source_line_end": 36,
"source_line_start": 26,
"title": "Source replay is"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 38,
"source_line_start": 37,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A cold reader should not reverse-engineer the path from a README, so the route map answers what to run first and what evidence shows that path is wired.** A cold technical reader should not have to reverse-engineer the product path from a long README or a raw tree of result files. The route map answers one question: what do I run first, and what evidence shows that path is wired?",
"level": 4,
"order": 5,
"section_id": "what-to-run-first-with-evidence",
"source_line_end": 46,
"source_line_start": 39,
"title": "What to run first, with evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A prose route map lists commands nothing re-checks; this one treats each row as a claim-boundary check so the trusted sequence stays tied to re-runnable source.** A loose route map answers that question with prose. It lists impressive commands, and nothing checks that they still exist or that their supporting records still pass. The route map here treats each row as a claim-boundary check, so the sequence a newcomer trusts stays tied to source that a test can re-run.",
"level": 4,
"order": 6,
"section_id": "each-row-as-a-claim-boundary-check",
"source_line_end": 57,
"source_line_start": 47,
"title": "Each row as a claim-boundary check"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 59,
"source_line_start": 58,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**run builds the fixture with negative cases and writes the record set, while runroutemap_bundle runs the same builder with negatives off and the copied-source manifest required.** run is the fixture entrypoint. It calls buildresult with negative cases included, then writes the result, board, and validation records, plus a coldreaderroutemapfixtureacceptancev1 record when the caller passes acceptanceout. runroutemapbundle is the exported-bundle entrypoint. It runs the same builder with negative cases off and the copied-source manifest required, so the bundle proves replay shape from public inputs alone.",
"level": 4,
"order": 8,
"section_id": "fixture-and-exported-bundle-entrypoints",
"source_line_end": 71,
"source_line_start": 60,
"title": "Fixture and exported-bundle entrypoints"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult runs four independent checks plus a secret-exclusion scan, passing only when no check files a blocking finding and the scan reports no blocking hit.** buildresult assembles four independent checks plus a secret-exclusion scan over the inputs, and passes only when none of the four checks files a blocking finding and the scan reports no blocking hit.",
"level": 4,
"order": 9,
"section_id": "build-result-joins-four-checks-and-a-scan",
"source_line_end": 80,
"source_line_start": 72,
"title": "`_build_result` joins four checks and a scan"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "positivefindings requires each route to name a command, docs refs, and a result record ref, and the front-door sequence to match live first-screen strings in ordinal order.** positivefindings walks the routes named in route_policy. Each required route must exist, name a command, carry docs_refs, and point at a result record ref. The first-run sequence must start with the three front-door routes in FRONTDOORROUTECOMMANDS (tourproject, statuscard, prooflab), those front-door commands must match the live first-screen strings, and the whole sequence must be present and ordinal sorted. A gap, a missing field, or a drifted front-door command becomes a finding.",
"level": 4,
"order": 10,
"section_id": "required-routes-and-front-door-ordering",
"source_line_end": 93,
"source_line_start": 81,
"title": "Required routes and front-door ordering"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "routesourcereplayresult builds a source corpus and blocks any command token, docs ref, result record ref, or claimed signal that does not resolve against real public source.** routesourcereplayresult is the replay pass. It builds one text corpus from the copied source modules, a fixed set of public source refs, every route's docs refs, and every result record ref, then it tests each route against that corpus. commandsupport splits a command into material tokens and blocks it if a token appears nowhere in the corpus. documentref_result requires each docs ref to resolve to a real public file and, when the ref names a heading anchor, to a real heading. receiptref_result opens each result recor\u2026",
"level": 4,
"order": 11,
"section_id": "replaying-commands-refs-and-signals-against-source",
"source_line_end": 109,
"source_line_start": 94,
"title": "Replaying commands, refs, and signals against source"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourcemoduleimportresult recomputes each copied body's digest against the manifest sha256, checks provenance anchors survive, and requires a copiednonsecretmacrobody exact-copy declaration.** sourcemoduleimportresult governs the copied-body floor used by the exported bundle. For each row in sourcemodulemanifest.json it recomputes the file digest, compares it to the manifest sha256, checks that required provenance anchors survive in the copied text, and confirms the row declares copiednonsecretmacrobody with a material class and an exact_copy relation. A digest mismatch, a missing anchor, a body marked for a record, or an unsupported class each files a finding.",
"level": 4,
"order": 12,
"section_id": "the-copied-body-floor-for-the-bundle",
"source_line_end": 122,
"source_line_start": 110,
"title": "The copied-body floor for the bundle"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "negativefindings evaluates the five perturbed inputs and records the error code each one must raise, listed under Negative cases below. negative_findings evaluates the five perturbed inputs and records the error code each one must raise, listed under Negative cases below. Function to role:",
"level": 4,
"order": 13,
"section_id": "negative-findings-evaluates",
"source_line_end": 154,
"source_line_start": 123,
"title": "Negative findings evaluates"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 156,
"source_line_start": 155,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVE_CASES names the five inputs the fixture must reject and the code each one raises. EXPECTEDNEGATIVE_CASES names the five inputs the fixture must reject and the code each one raises: missingcommandref raises COLDROUTEMISSINGCOMMANDREF: a route with no command cannot guide anyone. missingreceiptref raises COLDROUTEMISSINGRECEIPTREF: a route with no evidence ref is not backed. routesequencegap raises COLDROUTESEQUENCE_GAP: a sequence that does not start from tour_project is malformed. releaseoverclaim raises COLDROUTEAUTHORITYOVERCLAIM: a row that sets a key in OVERCLAIMKEYS, such as releaseauthorized or wholesystemcorrectness_claim, is refused. privatesourceleakage raises COLD\u2026",
"level": 4,
"order": 15,
"section_id": "expected-negative-cases",
"source_line_end": 176,
"source_line_start": 157,
"title": "Expected negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "If the fixture run does not observe one of these five, build_result marks the run blocked for the missing case.** If the fixture run does not observe one of these five, buildresult marks the run blocked for the missing case.",
"level": 4,
"order": 16,
"section_id": "if-fixture-run",
"source_line_end": 182,
"source_line_start": 177,
"title": "If fixture run"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "reader-specific-evidence-routing",
"source_line_end": 184,
"source_line_start": 183,
"title": "Reader-Specific Evidence Routing"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**The route map keeps a command order and an evidence order, so a reader sees the compact sequence first yet keeps each step's result record-ref and scope-limit context in drilldowns.** The route map keeps two orders: the command order and the evidence order. A reader sees the compact command sequence first, then can open full drilldowns without losing the result record-ref and scope-limit context attached to each step. The accepted path is: plectis hello plectis tour --card plectis status --card plectis authority --card plectis workingness --card plectis legibility-scorecard Read selectedrouteid from the compact tour or status card. plectis explain <\u2026",
"level": 4,
"order": 18,
"section_id": "command-order-and-evidence-order-kept-together",
"source_line_end": 216,
"source_line_start": 185,
"title": "Command order and evidence order kept together"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A browser-first reader follows the same order, and the compact board must carry the same command ref, evidence ref, class, and scope limit as the terminal row, or it is a preview.** A browser-first reader follows the same order in compressed form, and the compact board must carry the same command ref, evidence ref, evidence class, and scope limit as the terminal row, or it is a preview only.",
"level": 4,
"order": 19,
"section_id": "compact-board-must-match-the-terminal-row",
"source_line_end": 225,
"source_line_start": 217,
"title": "Compact board must match the terminal row"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "prior-art-grounding",
"source_line_end": 227,
"source_line_start": 226,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The design follows Diataxis's task-shaped documentation and Knuth's literate programming, borrowing the ordered-reader pattern without turning documentation sequence into source authority.** The component follows documentation systems that treat reader state and task shape as first-class. Diataxis separates tutorials, how-to guides, reference, and explanation so a reader is not pushed through one undifferentiated pile. Knuth's literate programming is the older anchor for writing an executable system for human comprehension as well as machine execution. The route map borrows the ordered-reader pattern of first command, evidence ref, evidence class, scope limit, and next drilldown, without\u2026",
"level": 4,
"order": 21,
"section_id": "diataxis-and-literate-programming-lineage",
"source_line_end": 242,
"source_line_start": 228,
"title": "Diataxis and literate programming lineage"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate/, reproduce this page's proof boundary with temporary records:",
"level": 2,
"order": 22,
"section_id": "validation-receipt-path",
"source_line_end": 254,
"source_line_start": 243,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the eleven-route fixture, the exported bundle, and the copied source-module digests all validated with no blocking finding and stay consistent with the source record and generated views.** A pass means the eleven-route fixture, the exported bundle, and the copied source-module digests all validated with no blocking finding, and that this page stays consistent with its source record, generated diagram, and Atlas card.",
"level": 4,
"order": 23,
"section_id": "what-a-pass-means-here",
"source_line_end": 263,
"source_line_start": 255,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 24,
"section_id": "claim-ceiling",
"source_line_end": 265,
"source_line_start": 264,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is narrow: a first-run route map whose commands, docs refs, and evidence refs resolve against public source, with an ordinal front-door sequence and five named rejections.** The strongest claim this component supports is narrow: a public first-run route map whose commands, docs refs, and evidence refs each resolve against real public source, whose front-door sequence is ordinal sorted, and whose five negative cases are rejected with named codes. Its scope_limit sets routeregistryauthority, sourcemutationauthorized, release_authorized, providercallsauthorized, privatedataequivalence_claim, and wholesystemcorrectness_claim all to false.",
"level": 4,
"order": 25,
"section_id": "the-narrow-strongest-claim",
"source_line_end": 278,
"source_line_start": 266,
"title": "The narrow strongest claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is not route registry control, mutates no source, calls no providers, excludes launch, and cannot promote the documentation sequence into proof beyond its public fixtures and records.** It refuses to become route registry control, to change source files projects, to use external model services, or to include launch operations or public sharing. It does not establish the whole system correct, and it does not claim equivalence to any non-public data. The route map can tell a cold reader what to run first and which record bounds that run. It cannot promote the documentation sequence into proof beyond those public fixtures and records.",
"level": 4,
"order": 26,
"section_id": "what-the-route-map-cannot-become",
"source_line_end": 290,
"source_line_start": 279,
"title": "What the route map cannot become"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/cold_reader_route_map.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public cold-reader route-map fixture validates first-run command order, docs refs, result record refs, scope limits, copied cold-entry source-module digests, and non-public-state exclusion without becoming route registry control."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "cold_reader_route_map.py",
"path": "src/microcosm_core/organs/cold_reader_route_map.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public cold-reader route-map validation, exported source-module manifest checks, negative-case enforcement, and scope limits.",
"symbols": [
"run",
"run_route_map_bundle",
"_route_source_replay_result",
"_source_module_import_result",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/cold_reader_route_map.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/cold_reader_route_map.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/cold_reader_route_map.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Cold-Reader Route Map is the public first-run route membrane. It validates route rows, route-to-result record bindings, ordinal first-run sequencing, command/docs/result record refs, launch/provider/private-source overclaim rejection, exact copied source cold-entry source modules, and metadata-only result records so a cold agent can see what to run first and what proof bounds each row.",
"summary_line": "The public cold-reader route-map fixture validates first-run command order, docs refs, result record refs, scope limits, copied cold-entry source-module digests, and non-public-state exclusion without becoming route registry control.",
"title": "Cold-Reader Route Map"
},
{
"anchor": "paper-module-proof-diagnostic-evidence-spine",
"authority_ceiling": "Public diagnostic result record refs, copied Ring2 runtime artifacts, and copied public component source body only; no Lean/Lake execution, formal proof authority, formal-result correctness, external model access authority, runtime correctness, launch-scope decision, publishing-scope decision, whole-system equivalence, or whole-system correctness.",
"body_html": "
proof_diagnostic_evidence_spine sits before formal proof work and recomputes which earlier diagnostic records may count as evidence, writing a board of accepted and rejected checks with reasons.proof_diagnostic_evidence_spine is the component that sits one step before formal proof work and decides which diagnostic records are allowed to count as evidence. It reads the records left by earlier formal-math evaluation and premise-retrieval steps, recomputes whether each one actually holds, and writes a board that lists which checks were accepted, which were rejected, and why.
\n
Rather than trust a row's self-declared pass, it resolves and re-hashes the named files and opens the cited result records, so a pass means evidence is present and coupled, not correct mathematics. It exists because a diagnostic record can lie by omission. A check row can name a real failure-taxonomy report and a real result record, declare itself passing, and be believed on its own word. This component does not trust that word. It resolves the files the row names, re-hashes them, opens the result records the row cites, and only then decides. A pass here means the named evidence is present and coupled. It does not mean the underlying mathematics is correct.
\n
metadata-only diagnostic evidence records
\n
The output is a metadata-only record set carrying no proof or provider-output body, because the component judges and arranges existing records rather than running a proof checker itself. The result is a metadata-only set of records: proof-evidence checks, a model-output data policy result, a diagnostic board, and a fixture sign-off record. None of them carry a proof body or a provider output body. The component arranges and judges existing records. It never runs a proof checker itself.
\n
Purpose
\n
Evidence inflation is cheap to write: trusting a row's own expected_result label would let a drifted digest, an unrelated result record, or a mislabelled check pass silently. The problem is evidence inflation. A row that says "this check passed" is cheap to write and expensive to trust. If the spine accepted the row's own expected_result label, a drifted digest, a result record that never mentions the ref it claims, or a check labelled as failure-taxonomy evidence while pointing at an unrelated file would all pass silently.
\n
Rederiving the verdict, not reading labels
\n
The component asks per row whether the check recomputes against the system it names, ignoring the fixture's pass label because the label is the thing under suspicion. The component answers one question per row: does this check recompute against the system it names, or is it asserting more than its refs support. A loose gate that reads labels fails because the label is the thing under suspicion. The spine ignores the fixture's pass label and rederives the verdict.
\n
How it works
\n
Two entry points and the shared pipeline
\n
run drives the base fixture and run_evidence_bundle the exported bundle with added artifact checks, both scanning inputs, running the validators in order, and merging observed negative cases before writing records. The runtime is src/microcosm_core/organs/proof_diagnostic_evidence_spine.py. Two entry points drive it. run handles the base fixture. run_evidence_bundle handles the exported evidence bundle and adds copied-artifact and source-body checks. Both load their inputs, run a secret-exclusion scan over them, call the validators below in order, merge the negative cases each validator observed, then write the records.
\n
The per-check verdict in _classify_evidence_check
\n
_classify_evidence_check resolves and re-hashes each source ref, confirms every cited result record truly references it, applies a semantic floor, and accepts only when all of that holds with no negative code declared. The core of a per-check verdict is _classify_evidence_check. For each source_ref on a check row it resolves a real public file with _resolve_public_ref, hashes the bytes with _sha256_file, and compares that hash to the expected digest the module holds in SOURCE_DIGESTS. It opens each result record anchor and confirms the result record payload actually contains the source ref and its digest, so a check is result record-backed only when the result record it cites genuinely references it. On top of that it applies a semantic floor through _check_semantic_floor: a check whose id mentions failure_taxonomy must resolve to a source carrying a failure-taxonomy report and a result record carrying a failure-mode ledger; a graph_update check needs graph-update candidates with ids and a matching anchor. The row is accepted only when every source resolves, every digest matches, every cited result record backs the ref, the semantic floor is satisfied, and no expected-negative error code is declared.
\n
Running the classifier over every check
\n
validate_evidence_receipts runs the classifier over every check, stamps any self-declared pass label as non-authoritative, and reports its own verdict as accepted ids or a recompute-failed finding.validate_evidence_receipts runs that classifier over every check. It records legacy_expected_result_label_present and stamps any such label as ignored_non_authoritative_fixture_label, then reports the verdict it computed itself. Accepted checks land in accepted_check_ids, rejected ones become an EVIDENCE_RECEIPT_ANCHOR_RECOMPUTE_FAILED finding with its rejection reasons.
\n
Advisory metadata kept, authority denied
\n
validate_provider_payload_policy rejects any provider or oracle row carrying a forbidden body key or claiming oracle authority, while a clean row keeps its advisory metadata under a one-sided retain-metadata, deny-authority rule.validate_provider_payload_policy handles provider and oracle advisory rows. It walks each payload with _forbidden_body_key_paths looking for any key in FORBIDDEN_BODY_KEYS, a fixed list that includes proof_body, ground_truth_proof, provider_output_body, oracle_payload_body, and prompt. A row with a forbidden key, or one that claims oracle authority, is rejected and kept as a regression fixture. A clean row keeps its advisory metadata and is marked advisory_metadata_preserved. The rule is one-sided: metadata is retained, authority is denied.
\n
Rejecting authority upgrades, keeping stale coupling
\n
validate_authority_ceiling rejects a board's source-authority and a check's runtime-correctness upgrades, while validate_stale_source_coupling keeps a drifted fingerprint as evidence and marks it stale.validate_authority_ceiling rejects two upgrade attempts. A diagnostic board that sets claims_source_authority produces DIAGNOSTIC_BOARD_AUTHORITY_UPGRADE. A passing check that sets claims_runtime_correctness produces EVIDENCE_PASS_OVERCLAIMS_RUNTIME_CORRECTNESS. validate_stale_source_coupling keeps a drifted fingerprint as evidence rather than dropping it, emitting PROOF_RECEIPT_SOURCE_COUPLING_STALE and setting source_fingerprint_status to stale.
\n
Digest-checking the copied Ring2 artifacts
\n
On the bundle path, validate_copied_macro_body_artifacts re-hashes each copied Ring2 artifact and passes only when all thirteen targets are present and digest-matched, blocking on any mismatch. On the bundle path, validate_copied_macro_body_artifacts re-hashes each copied Ring2 runtime artifact against PUBLIC_RING2_ARTIFACT_IMPORTS. It passes only when all thirteen expected targets are declared, present, and digest-matched under either the exact-copy or the source-faithful public-light-edit policy. Any missing file or digest mismatch blocks the status.
\n
Function
Role
run
Base fixture orchestrator: load, scan, validate, write records
run_evidence_bundle
Exported-bundle path with freshness reuse and copied-artifact checks
validate_evidence_receipts
Recompute every check row, ignore its self-declared pass label
Reject board source-authority and runtime-correctness upgrades
validate_copied_macro_body_artifacts
Digest-check the copied Ring2 runtime artifacts
\nDiagram of the mechanism (7 steps).Diagram source & refs
flowchart TD Check["Check row source_refs, receipt_anchor_refs, source_digest_refs"] Resolve["Resolve each source ref re-hash file (sha256)"] Result record["Open result record anchor does it contain this ref and digest?"] Floor["Semantic floor failure-taxonomy / graph-update source and result record match"] Accept["Accepted verdict recomputed"] Reject["Rejected, kept as diagnostic evidence"] Board["diagnostic board evidence accounting only"] Check --> Resolve --> Result record --> Floor Floor -->|all agree| Accept Floor -->|any mismatch| Reject Accept --> Board Reject --> Board
\n
Negative cases
\n
Six negative inputs the run must observe
\n
The fixture declares six negative inputs the run must observe, each mapped to its error codes, and the run blocks if any is missing. The fixture declares the negative cases the run must observe, listed in EXPECTED_NEGATIVE_CASES. A run is blocked if any of them is missing:
\n
provider_proof_body_payload_rejected: a model-output data carrying a forbidden proof body, giving FORBIDDEN_PROOF_BODY and PROVIDER_PAYLOAD_NOT_AUTHORITY.
oracle_payload_claims_authority_rejected: an oracle payload claiming authority, giving ORACLE_PAYLOAD_BODY_FORBIDDEN and ORACLE_PAYLOAD_NOT_AUTHORITY.
evidence_receipt_missing_required_fields: a result record lacking a validator id or scope boundary, giving MISSING_VALIDATOR_ID and MISSING_RECEIPT_ANTI_CLAIM.
The component follows assurance-case practice, tying evidence to a claim and its limits, and borrows result record discipline from GSN, W3C PROV, and observability, treating a diagnostic as an evidence cell rather than a proof. The component follows assurance-case practice: evidence is tied to a claim, its assumptions, and its limits before it is treated as support. NASA's Goal Structuring Notation example for spacecraft assurance frames assurance as model-structured evidence rather than document-level persuasion, NTRS 20160005295. The result record discipline also borrows from W3C PROV and from observability practice, where a diagnostic artifact is an evidence cell with provenance and not a proof of correctness.
\n
Validation Result record Path
\n
Fixture and bundle validation commands
\n
Run these from the repository root. Neither command mutates a durable record or a generated projection: Run these from the repository root. Neither command mutates a durable record or a generated projection:
Clearing this validation shows the fixture's expected refusals all fired, but proves nothing about the underlying mathematics or any provider's authority. A pass means the observed negative cases match EXPECTED_NEGATIVE_CASES. It does not establish formal-result correctness or provider authority.
\n
Scope boundary
\n
Scope limit
\n
Bounded evidence accounting
\n
The strongest claim is bounded evidence accounting: over the fixture inputs, the component recomputes which refs, digests, result record anchors, and negative cases hold and refuses to let a row inflate past that. The strongest claim the evidence supports is bounded evidence accounting: for the fixture inputs, this component recomputes which public refs, digests, result record anchors, and negative cases hold, and refuses to let a row inflate itself past that.
\n
What it refuses to establish
\n
It runs no Lean or Lake, calls no providers, exposes no proof or provider-output body, and proves no runtime correctness, with PROOF_AUTHORITY_CEILING and FORBIDDEN_BODY_KEYS encoding that boundary in code. It refuses the rest. It does not run Lean or Lake, it does not use external model services, it does not expose proof bodies or provider output bodies, and it does not establish that the imported runtime is correct. Its constants PROOF_AUTHORITY_CEILING and FORBIDDEN_BODY_KEYS encode that boundary in code. A row that adds claims_runtime_correctness is rejected as an overclaim.
\n",
"components": [
{
"anchor": "component-proof_diagnostic_evidence_spine",
"component_id": "proof_diagnostic_evidence_spine",
"public_label": "Proof Diagnostic Evidence Spine"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 6
},
"link_count": 16,
"links": [
{
"id": "proof_diagnostic_evidence_spine",
"kind": "organ",
"kind_label": "component",
"label": "Proof Diagnostic Evidence Spine",
"relation": "explains",
"url": "component-proof-diagnostic-evidence-spine.html"
},
{
"id": "mechanism.proof_diagnostic_evidence_spine.validates_ring2_diagnostic_evidence_membrane",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates ring2 diagnostic evidence membrane",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-proof-diagnostic-evidence-spine-validates-ring2-diagnostic-evidence-membrane"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depends on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
},
{
"id": "paper_module.formal_evidence_cell_anchor_resolver",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Evidence Cell Anchor Resolver",
"relation": "depends on",
"url": "paper-module-formal-evidence-cell-anchor-resolver.html"
},
{
"id": "paper_module.formal_math_lean_proof_witness",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Lean Proof Witness",
"relation": "depends on",
"url": "paper-module-formal-math-lean-proof-witness.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.ring2_premise_precision_recall",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Ring-2 Premise Precision Recall",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[42:paper_module.ring2_premise_precision_recall].depends_on",
"url": "paper-module-ring2-premise-precision-recall.html"
},
{
"id": "paper_module.target_shape_tactic_routing",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Target Shape Tactic Routing",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[41:paper_module.target_shape_tactic_routing].depends_on",
"url": "paper-module-target-shape-tactic-routing.html"
},
{
"id": "paper_module.verifier_lab_kernel",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Kernel",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[0:paper_module.verifier_lab_kernel].depends_on",
"url": "paper-module-verifier-lab-kernel.html"
}
],
"paper_module_id": "paper_module.proof_diagnostic_evidence_spine",
"reader_packet": {
"abstract": "The public proof-diagnostic evidence spine fixture validates Ring2 diagnostic result record refs, copied runtime artifact digests, provider/proof-body exclusions, stale-coupling visibility, and scope limits without becoming formal proof authority.",
"authority_ceiling": "Public diagnostic result record refs, copied Ring2 runtime artifacts, and copied public component source body only; no Lean/Lake execution, formal proof authority, formal-result correctness, external model access authority, runtime correctness, launch-scope decision, publishing-scope decision, whole-system equivalence, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.proof_diagnostic_evidence_spine.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-proof-diagnostic-evidence-spine-diagram-1",
"html_url": "docs/paper-module-proof-diagnostic-evidence-spine.html#paper-module-proof-diagnostic-evidence-spine-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Check[\"Check row source_refs, receipt_anchor_refs, source_digest_refs\"] Resolve[\"Resolve each source ref re-hash file (sha256)\"] Result record[\"Open result record anchor does it contain this ref and digest?\"] Floor[\"Semantic floor failure-taxonomy / graph-update source and result record match\"] Accept[\"Accepted verdict recomputed\"] Reject[\"Rejected, kept as diagnostic evidence\"] Board[\"diagnostic board evidence accounting only\"] Check --> Resolve --> Result record --> Floor Floor -->|all agree| Accept Floor -->|any mismatch| Reject Accept --> Board Reject --> Board",
"source_format": "mermaid",
"source_hash": "sha256:27b5316dd63b14ccd07dbfab487b7bd6d89402c57af5de3c15ccbaf1c0e31b91",
"source_line_end": 75,
"source_line_start": 60
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depends on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
},
{
"id": "paper_module.formal_evidence_cell_anchor_resolver",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Evidence Cell Anchor Resolver",
"relation": "depends on",
"url": "paper-module-formal-evidence-cell-anchor-resolver.html"
},
{
"id": "paper_module.formal_math_lean_proof_witness",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Lean Proof Witness",
"relation": "depends on",
"url": "paper-module-formal-math-lean-proof-witness.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "proof_diagnostic_evidence_spine",
"kind": "organ",
"kind_label": "component",
"label": "Proof Diagnostic Evidence Spine",
"relation": "explains",
"url": "component-proof-diagnostic-evidence-spine.html"
},
{
"id": "mechanism.proof_diagnostic_evidence_spine.validates_ring2_diagnostic_evidence_membrane",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates ring2 diagnostic evidence membrane",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-proof-diagnostic-evidence-spine-validates-ring2-diagnostic-evidence-membrane"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.proof_diagnostic_evidence_spine",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 122,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/proof_diagnostic_evidence_spine.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "NTRS 20160005295",
"section_id": "assurance-cases-and-provenance-discipline",
"url": "https://ntrs.nasa.gov/citations/20160005295"
},
{
"kind": "external_reference",
"label": "W3C PROV",
"section_id": "assurance-cases-and-provenance-discipline",
"url": "https://www.w3.org/TR/prov-overview/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "proofdiagnosticevidencespine sits before formal proof work and recomputes which earlier diagnostic records may count as evidence, writing a board of accepted and rejected checks with reasons. proofdiagnosticevidencespine is the component that sits one step before formal proof work and decides which diagnostic records are allowed to count as evidence. It reads the records left by earlier formal-math evaluation and premise-retrieval steps, recomputes whether each one actually holds, and writes a board that lists which checks were accepted, which were rejected, and why.",
"level": 4,
"order": 1,
"section_id": "proof-diagnostic-evidence-spine-sits-before-formal-proof-work",
"source_line_end": 6,
"source_line_start": 3,
"title": "`proof_diagnostic_evidence_spine` sits before formal proof work"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Rather than trust a row's self-declared pass, it resolves and re-hashes the named files and opens the cited result records, so a pass means evidence is present and coupled, not correct mathematics. It exists because a diagnostic record can lie by omission. A check row can name a real failure-taxonomy report and a real result record, declare itself passing, and be believed on its own word. This component does not trust that word. It resolves the files the row names, re-hashes them, opens the result records the row cites, and only then decides. A pass here means the named evidence is present and coupled. It does not mean the underlying mathematics is correct.",
"level": 4,
"order": 2,
"section_id": "rather-than-trust-a-row-s-self-declared-pass",
"source_line_end": 10,
"source_line_start": 7,
"title": "Rather than trust a row's self-declared pass"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The output is a metadata-only record set carrying no proof or provider-output body, because the component judges and arranges existing records rather than running a proof checker itself. The result is a metadata-only set of records: proof-evidence checks, a model-output data policy result, a diagnostic board, and a fixture sign-off record. None of them carry a proof body or a provider output body. The component arranges and judges existing records. It never runs a proof checker itself.",
"level": 4,
"order": 3,
"section_id": "body-free-diagnostic-evidence-records",
"source_line_end": 14,
"source_line_start": 11,
"title": "metadata-only diagnostic evidence records"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Evidence inflation is cheap to write: trusting a row's own expectedresult label would let a drifted digest, an unrelated result record, or a mislabelled check pass silently. The problem is evidence inflation. A row that says \"this check passed\" is cheap to write and expensive to trust. If the spine accepted the row's own expectedresult label, a drifted digest, a result record that never mentions the ref it claims, or a check labelled as failure-taxonomy evidence while pointing at an unrelated file would all pass silently.",
"level": 4,
"order": 5,
"section_id": "evidence-inflation",
"source_line_end": 20,
"source_line_start": 17,
"title": "Evidence inflation"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component asks per row whether the check recomputes against the system it names, ignoring the fixture's pass label because the label is the thing under suspicion. The component answers one question per row: does this check recompute against the system it names, or is it asserting more than its refs support. A loose gate that reads labels fails because the label is the thing under suspicion. The spine ignores the fixture's pass label and rederives the verdict.",
"level": 4,
"order": 6,
"section_id": "rederiving-the-verdict-not-reading-labels",
"source_line_end": 24,
"source_line_start": 21,
"title": "Rederiving the verdict, not reading labels"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run drives the base fixture and runevidencebundle the exported bundle with added artifact checks, both scanning inputs, running the validators in order, and merging observed negative cases before writing records. The runtime is src/microcosmcore/components/proofdiagnosticevidencespine.py. Two entry points drive it. run handles the base fixture. runevidencebundle handles the exported evidence bundle and adds copied-artifact and source-body checks. Both load their inputs, run a secret-exclusion scan over them, call the validators below in order, merge the negative cases each validator observed, then write the records.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-and-the-shared-pipeline",
"source_line_end": 30,
"source_line_start": 27,
"title": "Two entry points and the shared pipeline"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "classifyevidencecheck resolves and re-hashes each source ref, confirms every cited result record truly references it, applies a semantic floor, and accepts only when all of that holds with no negative code declared. The core of a per-check verdict is classifyevidencecheck. For each sourceref on a check row it resolves a real public file with resolvepublicref, hashes the bytes with sha256file, and compares that hash to the expected digest the module holds in SOURCEDIGESTS. It opens each result record anchor and confirms the result record payload actually contains the source ref and its digest, so a check is result record-backed only when the result record it cites genuinely references it. On\u2026",
"level": 4,
"order": 9,
"section_id": "the-per-check-verdict-in-classify-evidence-check",
"source_line_end": 34,
"source_line_start": 31,
"title": "The per-check verdict in `_classify_evidence_check`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateevidencereceipts runs the classifier over every check, stamps any self-declared pass label as non-authoritative, and reports its own verdict as accepted ids or a recompute-failed finding. validateevidencereceipts runs that classifier over every check. It records legacyexpectedresultlabelpresent and stamps any such label as ignorednonauthoritativefixturelabel, then reports the verdict it computed itself. Accepted checks land in acceptedcheckids, rejected ones become an EVIDENCERECEIPTANCHORRECOMPUTEFAILED finding with its rejection reasons.",
"level": 4,
"order": 10,
"section_id": "running-the-classifier-over-every-check",
"source_line_end": 38,
"source_line_start": 35,
"title": "Running the classifier over every check"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateproviderpayloadpolicy rejects any provider or oracle row carrying a forbidden body key or claiming oracle authority, while a clean row keeps its advisory metadata under a one-sided retain-metadata, deny-authority rule. validateproviderpayloadpolicy handles provider and oracle advisory rows. It walks each payload with forbiddenbodykeypaths looking for any key in FORBIDDENBODYKEYS, a fixed list that includes proofbody, groundtruthproof, provideroutputbody, oraclepayloadbody, and prompt. A row with a forbidden key, or one that claims oracle authority, is rejected and kept as a regression fixture. A clean row keeps its advisory metadata and is marked advisorymetadata_preserved. The rule\u2026",
"level": 4,
"order": 11,
"section_id": "advisory-metadata-kept-authority-denied",
"source_line_end": 42,
"source_line_start": 39,
"title": "Advisory metadata kept, authority denied"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateauthorityceiling rejects a board's source-authority and a check's runtime-correctness upgrades, while validatestalesourcecoupling keeps a drifted fingerprint as evidence and marks it stale. validateauthorityceiling rejects two upgrade attempts. A diagnostic board that sets claimssourceauthority produces DIAGNOSTICBOARDAUTHORITYUPGRADE. A passing check that sets claimsruntimecorrectness produces EVIDENCEPASSOVERCLAIMSRUNTIMECORRECTNESS. validatestalesourcecoupling keeps a drifted fingerprint as evidence rather than dropping it, emitting PROOFRECEIPTSOURCECOUPLINGSTALE and setting sourcefingerprint_status to stale.",
"level": 4,
"order": 12,
"section_id": "rejecting-authority-upgrades-keeping-stale-coupling",
"source_line_end": 46,
"source_line_start": 43,
"title": "Rejecting authority upgrades, keeping stale coupling"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "On the bundle path, validatecopiedmacrobodyartifacts re-hashes each copied Ring2 artifact and passes only when all thirteen targets are present and digest-matched, blocking on any mismatch. On the bundle path, validatecopiedmacrobodyartifacts re-hashes each copied Ring2 runtime artifact against PUBLICRING2ARTIFACT_IMPORTS. It passes only when all thirteen expected targets are declared, present, and digest-matched under either the exact-copy or the source-faithful public-light-edit policy. Any missing file or digest mismatch blocks the status.",
"level": 4,
"order": 13,
"section_id": "digest-checking-the-copied-ring2-artifacts",
"source_line_end": 76,
"source_line_start": 47,
"title": "Digest-checking the copied Ring2 artifacts"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 78,
"source_line_start": 77,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The fixture declares six negative inputs the run must observe, each mapped to its error codes, and the run blocks if any is missing. The fixture declares the negative cases the run must observe, listed in EXPECTEDNEGATIVECASES. A run is blocked if any of them is missing: providerproofbodypayloadrejected: a model-output data carrying a forbidden proof body, giving FORBIDDENPROOFBODY and PROVIDERPAYLOADNOT_AUTHORITY. oraclepayloadclaimsauthorityrejected: an oracle payload claiming authority, giving ORACLEPAYLOADBODYFORBIDDEN and ORACLEPAYLOADNOTAUTHORITY. evidencereceiptmissingrequiredfields: a result record lacking a validator id or scope boundary, giving MISSINGVALIDATORID and MISSINGRECEIP\u2026",
"level": 4,
"order": 15,
"section_id": "six-negative-inputs-the-run-must-observe",
"source_line_end": 89,
"source_line_start": 79,
"title": "Six negative inputs the run must observe"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 91,
"source_line_start": 90,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component follows assurance-case practice, tying evidence to a claim and its limits, and borrows result record discipline from GSN, W3C PROV, and observability, treating a diagnostic as an evidence cell rather than a proof. The component follows assurance-case practice: evidence is tied to a claim, its assumptions, and its limits before it is treated as support. NASA's Goal Structuring Notation example for spacecraft assurance frames assurance as model-structured evidence rather than document-level persuasion, NTRS 20160005295. The result record discipline also borrows from W3C PROV and from observability practice, where a diagnostic artifact is an evidence cell with provenance and not\u2026",
"level": 4,
"order": 17,
"section_id": "assurance-cases-and-provenance-discipline",
"source_line_end": 95,
"source_line_start": 92,
"title": "Assurance cases and provenance discipline"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 97,
"source_line_start": 96,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run these from the repository root. Neither command mutates a durable record or a generated projection: Run these from the repository root. Neither command mutates a durable record or a generated projection:",
"level": 4,
"order": 19,
"section_id": "fixture-and-bundle-validation-commands",
"source_line_end": 107,
"source_line_start": 98,
"title": "Fixture and bundle validation commands"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Clearing this validation shows the fixture's expected refusals all fired, but proves nothing about the underlying mathematics or any provider's authority. A pass means the observed negative cases match EXPECTEDNEGATIVECASES. It does not establish formal-result correctness or provider authority.",
"level": 4,
"order": 20,
"section_id": "what-a-pass-means",
"source_line_end": 111,
"source_line_start": 108,
"title": "What a pass means"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 113,
"source_line_start": 112,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim is bounded evidence accounting: over the fixture inputs, the component recomputes which refs, digests, result record anchors, and negative cases hold and refuses to let a row inflate past that. The strongest claim the evidence supports is bounded evidence accounting: for the fixture inputs, this component recomputes which public refs, digests, result record anchors, and negative cases hold, and refuses to let a row inflate itself past that.",
"level": 4,
"order": 22,
"section_id": "bounded-evidence-accounting",
"source_line_end": 117,
"source_line_start": 114,
"title": "Bounded evidence accounting"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It runs no Lean or Lake, calls no providers, exposes no proof or provider-output body, and proves no runtime correctness, with PROOFAUTHORITYCEILING and FORBIDDENBODYKEYS encoding that boundary in code. It refuses the rest. It does not run Lean or Lake, it does not use external model services, it does not expose proof bodies or provider output bodies, and it does not establish that the imported runtime is correct. Its constants PROOFAUTHORITYCEILING and FORBIDDENBODYKEYS encode that boundary in code. A row that adds claimsruntimecorrectness is rejected as an overclaim.",
"level": 4,
"order": 23,
"section_id": "what-it-refuses-to-establish",
"source_line_end": 121,
"source_line_start": 118,
"title": "What it refuses to establish"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/proof_diagnostic_evidence_spine.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public proof-diagnostic evidence spine fixture validates Ring2 diagnostic result record refs, copied runtime artifact digests, provider/proof-body exclusions, stale-coupling visibility, and scope limits without becoming formal proof authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "proof_diagnostic_evidence_spine.py",
"path": "src/microcosm_core/organs/proof_diagnostic_evidence_spine.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for proof-diagnostic evidence validation, exported evidence-bundle digest checks, provider/proof-body exclusion, negative cases, and scope limits.",
"symbols": [
"run",
"run_evidence_bundle",
"validate_copied_macro_body_artifacts",
"validate_evidence_receipts",
"validate_provider_payload_policy",
"validate_authority_ceiling",
"EXPECTED_NEGATIVE_CASES",
"PROOF_AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/proof_diagnostic_evidence_spine.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/proof_diagnostic_evidence_spine.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/proof_diagnostic_evidence_spine.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Proof Diagnostic Evidence Spine is the public evidence membrane before formal proof authority. It validates Ring2 failure-taxonomy and graph-update artifacts, verifier-trace repair and evidence-cell result record refs, copied runtime artifact digests, model-output data policy rows, negative cases, and a copied public component source-body floor while keeping proof bodies and provider output bodies out of public result records.",
"summary_line": "The public proof-diagnostic evidence spine fixture validates Ring2 diagnostic result record refs, copied runtime artifact digests, provider/proof-body exclusions, stale-coupling visibility, and scope limits without becoming formal proof authority.",
"title": "Proof Diagnostic Evidence Spine"
},
{
"anchor": "paper-module-proof-derived-governed-mutation-authorization",
"authority_ceiling": "Public synthetic governed-mutation result records and copied source body digest evidence only; no standing account secret authority, live cloud/account mutation, source-file changes, external model access, hidden vote, policy-after-execution authority, benchmark security claim, launch-scope decision, public sharing, hosting, or whole-system correctness.",
"body_html": "
The component replays whether a proposed change should be allowed to run, checking synthetic proposals against their evidence and writing a metadata-only record of what was admitted and why. This component replays a make-believe question that an agent system has to answer safely: should this proposed change be allowed to run? It takes a small bundle of synthetic proposals and their supporting evidence, checks each proposal against that evidence, and writes a metadata-only result record that shows which proposals were admitted and why.
\n
Three proposals under one authorization replay
\n
Three proposals, an inspection, a scoped write, and its undo, are admitted only when an independent evidence chain resolves, so wording, identity, hidden votes, and bare success claims all fail. The bundle holds three proposals: a look-only inspection, a small scoped config write, and an undo of that write. A proposal is admitted only when an independent evidence chain resolves for it. Impressive wording, an admin-looking identity, a hidden vote, or a final answer that just says it worked all fail on their own.
\n
Entry points and emitted authorization evidence
\n
The component runs via run on the fixture or run_authorization_bundle on the exported example, emitting counts, refs, hashes, and verdicts but never proof bodies, votes, or account secrets. The component is at src/microcosm_core/organs/proof_derived_governed_mutation_authorization.py. Run it with run for the fixture, or run_authorization_bundle for the exported example. The result record carries counts, refs, hashes, and verdicts, never proof bodies, policy votes, or account secrets.
\n
Purpose
\n
The action that grants itself permission
\n
The danger is a step that authorizes itself by claiming a standing account secret, recording a hidden governance-vote, or reporting success after the fact, all of which a trusting check would admit. The danger this guards against is an action that grants itself permission. In an agent system a step can claim a standing account secret, record a governance-vote nobody can see, or report success after the fact. A loose check that trusts the proposal's own summary would admit all three.
\n
Authorization derived from resolved evidence files
\n
Each admitted proposal must resolve against separate evidence files, and a recomputed evidence-chain hash must match the declared one, so a hand-written hash constant cannot survive. Authorization here is derived, not asserted. Each admitted proposal must resolve against separate evidence files: redacted proof cells that name validator result refs, at least two visible policy verdicts evaluated before any execution identity, a logged side-effect diff for the two changing actions, a matching undo record, and a cold-replay pass. The component recomputes an evidence-chain hash from those resolved rows and rejects the proposal when the declared hash does not match. A hand-written hash constant cannot survive that step.
\n
The anti-bake gate binds to a real commit
\n
The anti-bake gate requires each proposal to bind to a real git commit that touched this component's source or test, with its refs re-derived so a fixture cannot pre-write its own answer. The harder part is the anti-bake gate. Passing the synthetic chain is not enough. Every admitted proposal must also bind to a real repository record: a concrete git commit that the component resolves with a git subprocess and checks touched this component's own source or its focused test. The declared proof, policy, and undo refs on that record are re-derived from the evidence files and compared, so a fixture cannot pre-write its own answer.
\n
How it works
\n
The two entry points share one validator order
\n
run validates the fixture with negative cases and run_authorization_bundle validates the exported example without them, both calling the same validators in the same order.run loads the fixture input directory with negative cases included and writes the durable result set. run_authorization_bundle runs the exported example without the negative cases and writes one bundle result file. Both call the same validators in the same order.
\n
Per-proposal checks in _validate_proposal_row
\n
_validate_proposal_row checks the fourteen required fields against the other files, recomputes the evidence-chain hash, and rejects an undo-less write, thin verdicts, or an action outside the three fixed classes. Each proposal must carry the fourteen fields in REQUIRED_PROPOSAL_FIELDS, including proof_cell_refs, policy_verdict_refs, policy_evaluated_before_execution, side_effect_class, evidence_chain_hash, and cold_replay_ref. _validate_proposal_row checks each field against the other files rather than against the proposal's own claim. It recomputes the hash through _proposal_evidence_chain_hash over the resolved proof digests, policy digests, side-effect ref, undo ref, and cold-replay ref. A write or undo proposal with no undo ref is rejected as an irreversible change. Fewer than two visible allow or warn verdicts from distinct evaluators blocks the proposal. The three action classes admitted are fixed in REQUIRED_ACTION_CLASSES: read_only_inspection, scoped_config_write, and rollback.
\n
A gate for each supporting evidence file
\n
Each supporting file has its own gate over proof cells, policy verdicts, side-effect rows, undo records, and cold-replay passes, with a minimum count and a body-redacted requirement on each. The supporting files each have their own gate. validate_proof_evidence_cells needs three or more cells, each naming its proposal, carrying evidence refs and validator result refs, and marked body-redacted with no exported proof body. validate_policy_verdicts needs six or more verdicts and at least two visible allow or warn verdicts per proposal, each resolving back to a proof cell. validate_side_effect_ledger needs exactly two logged, reversible, synthetic side-effect rows. validate_rollback_receipts needs at least two passing undo records that do not claim irreversible authority. validate_cold_replay needs at least three redacted cold-replay passes.
\n
The real-record gate in validate_governed_mutation_records
\n
validate_governed_mutation_records resolves each commit with _git_commit_evidence, sets scope_verified only when a changed file is this component's source or test, and downgrades any proposal without an accepted record.validate_governed_mutation_records runs the real-record gate. A record must declare a repo record class, name a proposal that exists, and carry a commit ref that _git_commit_evidence resolves. That helper runs git rev-parse and git show --name-only, then sets scope_verified only when a changed file ends with this component's source or test path. The record's declared proof, policy, and undo refs are re-derived from the evidence indices and must match. A proposal whose id is not in the accepted real-record set is downgraded from admitted to blocked.
\n
Digest check on copied internal control bodies
\n
_source_module_manifest_result verifies each copied internal control body by sha256 against its declared digest, and the summary holds every body-in-result record flag false so only refs and hashes are reported._source_module_manifest_result checks the copied internal control bodies shipped with the exported example. It reads source_module_manifest.json, confirms each row is classed as copied body material, verifies each target file exists under source_modules/, and recomputes its sha256 against the declared digest and required anchors. _source_open_body_import_summary projects that result and holds body_in_receipt, body_text_in_receipt, and body_text_exported_in_receipts at false, so refs and hashes are reported but body text is not.
\n
Function
Role
run, run_authorization_bundle
Fixture and exported-example entry points that write the result set
Digest check on copied internal control bodies, body text kept out
EXPECTED_NEGATIVE_CASES, AUTHORITY_CEILING
The falsification set and the scope limit
\nDiagram of the mechanism (7 steps).Diagram source & refs
Source refs
read-only, scoped write, undo
mutation_proposals.json
real commit ref
governed_mutation_records.json
flowchart TD Proposals["mutation_proposals.json read-only, scoped write, undo"] Evidence["proof cells, policy verdicts, side effects, undo records, cold replay"] Hash{"declared hash == derived hash?"} Records["governed_mutation_records.json real commit ref"] AntiBake{"git commit touched source/test? re-derived refs match?"} Result["metadata-only result record counts, refs, hashes, verdicts"] Ceiling["scope_limit no account secrets, no live change"] Proposals --> Evidence Evidence --> Hash Hash -->|match| AntiBake Hash -->|mismatch| Result Records --> AntiBake AntiBake -->|bound| Result AntiBake -->|unbound| Result Result --> Ceiling
\n
Negative cases
\n
Eight rejections mapped to GOV_MUT_* codes
\n
EXPECTED_NEGATIVE_CASES lists eight rejections, each mapped to a GOV_MUT_* code, from standing account secrets and hidden votes to unlogged side effects and final-answer-only success.EXPECTED_NEGATIVE_CASES lists eight rejections, each mapped to a GOV_MUT_* error code. A proposal that claims a standing account secret fails GOV MUT STANDING account secret AUTHORITY. A proposal evaluated by policy after it ran fails GOV_MUT_POLICY_AFTER_EXECUTION. A hidden vote fails GOV_MUT_HIDDEN_POLICY_VOTE. A live cloud account secret fails GOV MUT LIVE CLOUD account secret. A change with no undo ref fails GOV_MUT_IRREVERSIBLE_MUTATION. An unlogged side effect fails GOV_MUT_UNLOGGED_SIDE_EFFECT. Consensus with no resolving evidence fails GOV_MUT_CONSENSUS_WITHOUT_EVIDENCE. A proposal graded only by a final-answer success fails GOV_MUT_FINAL_ANSWER_ONLY_SUCCESS.
\n
Each rejection turns on evidence, not on labels
\n
The eight cases are the security argument: each isolates one overclaim and fails because the public evidence tables do not resolve, never because the proposal looked suspicious. These eight are the security argument. Each isolates one overclaim: a standing account secret, an after-the-fact policy check, a hidden vote, a live cloud account secret, a write with no undo ref, an unlogged side effect, a missing evidence chain, or a bare success claim. None of them depends on a suspicious label. Each fails because the public evidence tables the predicate checks do not resolve, not because the proposal looked wrong.
\n
Prior Art Grounding
\n
Admission control and policy-as-code lineage
\n
The shape follows admission-control and policy-as-code practice from Open Policy Agent and Kubernetes admission controllers, with undo and side-effect parts adjacent to controlled-rollout feature-flag patterns, kept synthetic and replay-only. The shape follows admission-control and policy-as-code practice: a proposed state change is evaluated before it mutates the system, and the decision is separate from the actor's own assertion. The closest public anchors are Open Policy Agent, which separates policy decision-making from enforcement over structured input, and Kubernetes admission controllers, which validate or mutate API requests before they persist. The undo and side-effect parts are adjacent to controlled-rollout practice, including the feature-flag patterns described by Martin Fowler. This component keeps the pattern synthetic and replay-only.
A pass means the focused tests hold and this Markdown still satisfies the shared coverage contract, not that any real change was authorized. A pass means the focused tests hold and this Markdown still satisfies the shared paper-module coverage contract. It does not mean any real change was authorized.
\n
Scope boundary
\n
Scope limit
\n
The bounded claim the evidence supports
\n
On a synthetic bundle, three proposals were admitted only after a full evidence chain, matching hash, and a scoped real commit, while the same predicate rejects eight named overclaims. The strongest claim the evidence supports is a bounded one: for a synthetic bundle, three proposals were admitted only after proof cells, two visible resolving policy verdicts, logged side effects, undo records, cold-replay passes, an evidence-chain hash that matched its recomputation, and a real commit that touched this component's own source or test. The negative cases show the same predicate rejecting eight named overclaims. Copied internal control bodies are checked by digest with body text kept out of the result record.
\n
What stays unauthorized
\n
The boundary grants no live mutation, account secrets, cloud reach, source-file changes, external model access, or exported body, with AUTHORITY_CEILING recording each as not authorized and a green result staying fixture evidence. That is the whole proof boundary. This component does not grant live mutation authority, hold standing account secrets, reach any cloud or account, change source files, call a provider, export a proof body or governance-vote, or claim benchmark safety. AUTHORITY_CEILING records every one of those as not authorized. A green result is validator evidence over a fixture, not launch-scope decision, hosting approval, or whole-system correctness.
\n",
"components": [
{
"anchor": "component-proof_derived_governed_mutation_authorization",
"component_id": "proof_derived_governed_mutation_authorization",
"public_label": "Proof Derived Governed Mutation Authorization"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 2
},
"link_count": 7,
"links": [
{
"id": "proof_derived_governed_mutation_authorization",
"kind": "organ",
"kind_label": "component",
"label": "Proof Derived Governed Mutation Authorization",
"relation": "explains",
"url": "component-proof-derived-governed-mutation-authorization.html"
},
{
"id": "mechanism.proof_derived_governed_mutation_authorization.validates_synthetic_governed_mutation_authorization",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates synthetic governed mutation authorization",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-proof-derived-governed-mutation-authorization-validates-synthetic-governed-mutation-authorization"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "P-4",
"kind": "principle",
"kind_label": "principle",
"label": "Possession is not permission",
"relation": "governed by",
"url": "doctrine.html#dcard-p-4"
},
{
"id": "AX-3",
"kind": "axiom",
"kind_label": "axiom",
"label": "Authority by derivation, not possession",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-3"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch12_release_claim_language_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 12 launch claim-Language Gate",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[66:paper_module.batch12_release_claim_language_gate].depends_on",
"url": "paper-module-batch12-release-claim-language-gate.html"
},
{
"id": "paper_module.batch4_proof_authority_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 4 Proof, Authority, and Runtime Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[74:paper_module.batch4_proof_authority_runtime].depends_on",
"url": "paper-module-batch4-proof-authority-runtime.html"
},
{
"id": "paper_module.batch8_policy_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Policy Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[60:paper_module.batch8_policy_engines_capsule].depends_on",
"url": "paper-module-batch8-policy-engines-capsule.html"
}
],
"paper_module_id": "paper_module.proof_derived_governed_mutation_authorization",
"reader_packet": {
"abstract": "The public proof-derived governed-mutation fixture validates synthetic mutation proposals through proof cells, visible pre-execution policy verdicts, logged side effects, rollback result records, cold replay, negative cases, and copied source internal control bodies without granting live mutation authority.",
"authority_ceiling": "Public synthetic governed-mutation result records and copied source body digest evidence only; no standing account secret authority, live cloud/account mutation, source-file changes, external model access, hidden vote, policy-after-execution authority, benchmark security claim, launch-scope decision, public sharing, hosting, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.proof_derived_governed_mutation_authorization.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-proof-derived-governed-mutation-authorization-diagram-1",
"html_url": "docs/paper-module-proof-derived-governed-mutation-authorization.html#paper-module-proof-derived-governed-mutation-authorization-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Proposals[\"mutation_proposals.json read-only, scoped write, undo\"] Evidence[\"proof cells, policy verdicts, side effects, undo records, cold replay\"] Hash{\"declared hash == derived hash?\"} Records[\"governed_mutation_records.json real commit ref\"] AntiBake{\"git commit touched source/test? re-derived refs match?\"} Result[\"metadata-only result record counts, refs, hashes, verdicts\"] Ceiling[\"scope_limit no account secrets, no live change\"] Proposals --> Evidence Evidence --> Hash Hash -->|match| AntiBake Hash -->|mismatch| Result Records --> AntiBake AntiBake -->|bound| Result AntiBake -->|unbound| Result Result --> Ceiling",
"source_format": "mermaid",
"source_hash": "sha256:6b606992628a913364d5b1b30cee269979f31718804510e07bfdeef252148fbd",
"source_line_end": 82,
"source_line_start": 65
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-3",
"kind": "axiom",
"kind_label": "axiom",
"label": "Authority by derivation, not possession",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-3"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "proof_derived_governed_mutation_authorization",
"kind": "organ",
"kind_label": "component",
"label": "Proof Derived Governed Mutation Authorization",
"relation": "explains",
"url": "component-proof-derived-governed-mutation-authorization.html"
},
{
"id": "mechanism.proof_derived_governed_mutation_authorization.validates_synthetic_governed_mutation_authorization",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates synthetic governed mutation authorization",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-proof-derived-governed-mutation-authorization-validates-synthetic-governed-mutation-authorization"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "P-4",
"kind": "principle",
"kind_label": "principle",
"label": "Possession is not permission",
"relation": "governed by",
"url": "doctrine.html#dcard-p-4"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.proof_derived_governed_mutation_authorization",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 126,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/proof_derived_governed_mutation_authorization.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Open Policy Agent",
"section_id": "admission-control-and-policy-as-code-lineage",
"url": "https://www.openpolicyagent.org/docs"
},
{
"kind": "external_reference",
"label": "admission controllers",
"section_id": "admission-control-and-policy-as-code-lineage",
"url": "https://kubernetes.io/docs/reference/access-authn-authz/admission-controllers/"
},
{
"kind": "external_reference",
"label": "Martin Fowler",
"section_id": "admission-control-and-policy-as-code-lineage",
"url": "https://martinfowler.com/articles/feature-toggles.html"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component replays whether a proposed change should be allowed to run, checking synthetic proposals against their evidence and writing a metadata-only record of what was admitted and why. This component replays a make-believe question that an agent system has to answer safely: should this proposed change be allowed to run? It takes a small bundle of synthetic proposals and their supporting evidence, checks each proposal against that evidence, and writes a metadata-only result record that shows which proposals were admitted and why.",
"level": 4,
"order": 1,
"section_id": "the-component-replays-whether-a-proposed-change-should-be-allowed-to-run",
"source_line_end": 6,
"source_line_start": 3,
"title": "The component replays whether a proposed change should be allowed to run"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Three proposals, an inspection, a scoped write, and its undo, are admitted only when an independent evidence chain resolves, so wording, identity, hidden votes, and bare success claims all fail. The bundle holds three proposals: a look-only inspection, a small scoped config write, and an undo of that write. A proposal is admitted only when an independent evidence chain resolves for it. Impressive wording, an admin-looking identity, a hidden vote, or a final answer that just says it worked all fail on their own.",
"level": 4,
"order": 2,
"section_id": "three-proposals-under-one-authorization-replay",
"source_line_end": 10,
"source_line_start": 7,
"title": "Three proposals under one authorization replay"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component runs via run on the fixture or runauthorizationbundle on the exported example, emitting counts, refs, hashes, and verdicts but never proof bodies, votes, or account secrets. The component is at src/microcosmcore/components/proofderivedgovernedmutationauthorization.py. Run it with run for the fixture, or runauthorization_bundle for the exported example. The result record carries counts, refs, hashes, and verdicts, never proof bodies, policy votes, or account secrets.",
"level": 4,
"order": 3,
"section_id": "entry-points-and-emitted-authorization-evidence",
"source_line_end": 14,
"source_line_start": 11,
"title": "Entry points and emitted authorization evidence"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The danger is a step that authorizes itself by claiming a standing account secret, recording a hidden governance-vote, or reporting success after the fact, all of which a trusting check would admit. The danger this guards against is an action that grants itself permission. In an agent system a step can claim a standing account secret, record a governance-vote nobody can see, or report success after the fact. A loose check that trusts the proposal's own summary would admit all three.",
"level": 4,
"order": 5,
"section_id": "the-action-that-grants-itself-permission",
"source_line_end": 20,
"source_line_start": 17,
"title": "The action that grants itself permission"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Each admitted proposal must resolve against separate evidence files, and a recomputed evidence-chain hash must match the declared one, so a hand-written hash constant cannot survive. Authorization here is derived, not asserted. Each admitted proposal must resolve against separate evidence files: redacted proof cells that name validator result refs, at least two visible policy verdicts evaluated before any execution identity, a logged side-effect diff for the two changing actions, a matching undo record, and a cold-replay pass. The component recomputes an evidence-chain hash from those resolved rows and rejects the proposal when the declared hash does not match. A hand-written hash constant\u2026",
"level": 4,
"order": 6,
"section_id": "authorization-derived-from-resolved-evidence-files",
"source_line_end": 24,
"source_line_start": 21,
"title": "Authorization derived from resolved evidence files"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The anti-bake gate requires each proposal to bind to a real git commit that touched this component's source or test, with its refs re-derived so a fixture cannot pre-write its own answer. The harder part is the anti-bake gate. Passing the synthetic chain is not enough. Every admitted proposal must also bind to a real repository record: a concrete git commit that the component resolves with a git subprocess and checks touched this component's own source or its focused test. The declared proof, policy, and undo refs on that record are re-derived from the evidence files and compared, so a fixture cannot pre-write its own answer.",
"level": 4,
"order": 7,
"section_id": "the-anti-bake-gate-binds-to-a-real-commit",
"source_line_end": 28,
"source_line_start": 25,
"title": "The anti-bake gate binds to a real commit"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 30,
"source_line_start": 29,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run validates the fixture with negative cases and runauthorizationbundle validates the exported example without them, both calling the same validators in the same order. run loads the fixture input directory with negative cases included and writes the durable result set. runauthorizationbundle runs the exported example without the negative cases and writes one bundle result file. Both call the same validators in the same order.",
"level": 4,
"order": 9,
"section_id": "the-two-entry-points-share-one-validator-order",
"source_line_end": 34,
"source_line_start": 31,
"title": "The two entry points share one validator order"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateproposalrow checks the fourteen required fields against the other files, recomputes the evidence-chain hash, and rejects an undo-less write, thin verdicts, or an action outside the three fixed classes. Each proposal must carry the fourteen fields in REQUIREDPROPOSALFIELDS, including proofcellrefs, policyverdictrefs, policyevaluatedbeforeexecution, sideeffectclass, evidencechainhash, and coldreplayref. validateproposalrow checks each field against the other files rather than against the proposal's own claim. It recomputes the hash through proposalevidencechainhash over the resolved proof digests, policy digests, side-effect ref, undo ref, and cold-replay ref. A write or undo proposal\u2026",
"level": 4,
"order": 10,
"section_id": "per-proposal-checks-in-validate-proposal-row",
"source_line_end": 38,
"source_line_start": 35,
"title": "Per-proposal checks in `_validate_proposal_row`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Each supporting file has its own gate over proof cells, policy verdicts, side-effect rows, undo records, and cold-replay passes, with a minimum count and a body-redacted requirement on each. The supporting files each have their own gate. validateproofevidencecells needs three or more cells, each naming its proposal, carrying evidence refs and validator result refs, and marked body-redacted with no exported proof body. validatepolicyverdicts needs six or more verdicts and at least two visible allow or warn verdicts per proposal, each resolving back to a proof cell. validatesideeffectledger needs exactly two logged, reversible, synthetic side-effect rows. validaterollbackreceipts needs at lea\u2026",
"level": 4,
"order": 11,
"section_id": "a-gate-for-each-supporting-evidence-file",
"source_line_end": 42,
"source_line_start": 39,
"title": "A gate for each supporting evidence file"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validategovernedmutationrecords resolves each commit with gitcommitevidence, sets scopeverified only when a changed file is this component's source or test, and downgrades any proposal without an accepted record. validategovernedmutationrecords runs the real-record gate. A record must declare a repo record class, name a proposal that exists, and carry a commit ref that gitcommitevidence resolves. That helper runs git rev-parse and git show --name-only, then sets scopeverified only when a changed file ends with this component's source or test path. The record's declared proof, policy, and undo refs are re-derived from the evidence indices and must match. A proposal whose id is not in the acc\u2026",
"level": 4,
"order": 12,
"section_id": "the-real-record-gate-in-validate-governed-mutation-records",
"source_line_end": 46,
"source_line_start": 43,
"title": "The real-record gate in `validate_governed_mutation_records`"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "sourcemodulemanifestresult verifies each copied internal control body by sha256 against its declared digest, and the summary holds every body-in-result record flag false so only refs and hashes are reported. sourcemodulemanifestresult checks the copied internal control bodies shipped with the exported example. It reads sourcemodulemanifest.json, confirms each row is classed as copied body material, verifies each target file exists under sourcemodules/, and recomputes its sha256 against the declared digest and required anchors. sourceopenbodyimportsummary projects that result and holds bodyinreceipt, bodytextinreceipt, and bodytextexportedinreceipts at false, so refs and hashes are reported\u2026",
"level": 4,
"order": 13,
"section_id": "digest-check-on-copied-control-plane-bodies",
"source_line_end": 83,
"source_line_start": 47,
"title": "Digest check on copied internal control bodies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 85,
"source_line_start": 84,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVECASES lists eight rejections, each mapped to a GOVMUT code, from standing account secrets and hidden votes to unlogged side effects and final-answer-only success. EXPECTEDNEGATIVECASES lists eight rejections, each mapped to a GOVMUT error code. A proposal that claims a standing account secret fails GOVMUTSTANDINGCREDENTIALAUTHORITY. A proposal evaluated by policy after it ran fails GOVMUTPOLICYAFTEREXECUTION. A hidden vote fails GOVMUTHIDDENPOLICYVOTE. A live cloud account secret fails GOVMUTLIVECLOUDCREDENTIAL. A change with no undo ref fails GOVMUTIRREVERSIBLEMUTATION. An unlogged side effect fails GOVMUTUNLOGGEDSIDEEFFECT. Consensus with no resolving evidence fails GOVMUT\u2026",
"level": 4,
"order": 15,
"section_id": "eight-rejections-mapped-to-gov-mut-codes",
"source_line_end": 89,
"source_line_start": 86,
"title": "Eight rejections mapped to `GOV_MUT_*` codes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The eight cases are the security argument: each isolates one overclaim and fails because the public evidence tables do not resolve, never because the proposal looked suspicious. These eight are the security argument. Each isolates one overclaim: a standing account secret, an after-the-fact policy check, a hidden vote, a live cloud account secret, a write with no undo ref, an unlogged side effect, a missing evidence chain, or a bare success claim. None of them depends on a suspicious label. Each fails because the public evidence tables the predicate checks do not resolve, not because the proposal looked wrong.",
"level": 4,
"order": 16,
"section_id": "each-rejection-turns-on-evidence-not-on-labels",
"source_line_end": 93,
"source_line_start": 90,
"title": "Each rejection turns on evidence, not on labels"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 95,
"source_line_start": 94,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The shape follows admission-control and policy-as-code practice from Open Policy Agent and Kubernetes admission controllers, with undo and side-effect parts adjacent to controlled-rollout feature-flag patterns, kept synthetic and replay-only. The shape follows admission-control and policy-as-code practice: a proposed state change is evaluated before it mutates the system, and the decision is separate from the actor's own assertion. The closest public anchors are Open Policy Agent, which separates policy decision-making from enforcement over structured input, and Kubernetes admission controllers, which validate or mutate API requests before they persist. The undo and side-effect parts are ad\u2026",
"level": 4,
"order": 18,
"section_id": "admission-control-and-policy-as-code-lineage",
"source_line_end": 99,
"source_line_start": 96,
"title": "Admission control and policy-as-code lineage"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 101,
"source_line_start": 100,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": null,
"level": 4,
"order": 20,
"section_id": "reader-validation-without-durable-writes",
"source_line_end": 111,
"source_line_start": 102,
"title": "Reader validation without durable writes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means the focused tests hold and this Markdown still satisfies the shared coverage contract, not that any real change was authorized. A pass means the focused tests hold and this Markdown still satisfies the shared paper-module coverage contract. It does not mean any real change was authorized.",
"level": 4,
"order": 21,
"section_id": "what-a-pass-means",
"source_line_end": 115,
"source_line_start": 112,
"title": "What a pass means"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 117,
"source_line_start": 116,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "On a synthetic bundle, three proposals were admitted only after a full evidence chain, matching hash, and a scoped real commit, while the same predicate rejects eight named overclaims. The strongest claim the evidence supports is a bounded one: for a synthetic bundle, three proposals were admitted only after proof cells, two visible resolving policy verdicts, logged side effects, undo records, cold-replay passes, an evidence-chain hash that matched its recomputation, and a real commit that touched this component's own source or test. The negative cases show the same predicate rejecting eight named overclaims. Copied internal control bodies are checked by digest with body text kept out of th\u2026",
"level": 4,
"order": 23,
"section_id": "the-bounded-claim-the-evidence-supports",
"source_line_end": 121,
"source_line_start": 118,
"title": "The bounded claim the evidence supports"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The boundary grants no live mutation, account secrets, cloud reach, source-file changes, external model access, or exported body, with AUTHORITYCEILING recording each as not authorized and a green result staying fixture evidence. That is the whole proof boundary. This component does not grant live mutation authority, hold standing account secrets, reach any cloud or account, change source files, call a provider, export a proof body or governance-vote, or claim benchmark safety. AUTHORITYCEILING records every one of those as not authorized. A green result is validator evidence over a fixture, not launch-scope decision, hosting approval, or whole-system correctness.",
"level": 4,
"order": 24,
"section_id": "what-stays-unauthorized",
"source_line_end": 125,
"source_line_start": 122,
"title": "What stays unauthorized"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/proof_derived_governed_mutation_authorization.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public proof-derived governed-mutation fixture validates synthetic mutation proposals through proof cells, visible pre-execution policy verdicts, logged side effects, rollback result records, cold replay, negative cases, and copied source internal control bodies without granting live mutation authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "proof_derived_governed_mutation_authorization.py",
"path": "src/microcosm_core/organs/proof_derived_governed_mutation_authorization.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for governed-mutation authorization fixture validation, exported bundle validation, copied source-module digest checks, negative-case enforcement, cards, and scope limits.",
"symbols": [
"run",
"run_authorization_bundle",
"validate_mutation_proposals",
"validate_proof_evidence_cells",
"validate_policy_verdicts",
"validate_side_effect_ledger",
"validate_rollback_receipts",
"validate_cold_replay"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/proof_derived_governed_mutation_authorization.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/proof_derived_governed_mutation_authorization.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/proof_derived_governed_mutation_authorization.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Proof-Derived Governed Mutation Authorization is the public mutation-authority replay contract. It checks three synthetic proposals, proof-cell validator refs, visible policy verdicts, side-effect logs, rollback result records, cold replay, eight negative cases, and six exact copied source pattern/result record/internal control bodies while keeping account secrets, proof bodies, model-output data, account refs, source-file changes, and launch-scope decision out of result records.",
"summary_line": "The public proof-derived governed-mutation fixture validates synthetic mutation proposals through proof cells, visible pre-execution policy verdicts, logged side effects, rollback result records, cold replay, negative cases, and copied source internal control bodies without granting live mutation authority.",
"title": "Proof-Derived Governed Mutation Authorization"
},
{
"anchor": "paper-module-durable-agent-work-landing-replay",
"authority_ceiling": "Public synthetic replay result records and copied source work-landing/internal control source bodies only; no live Git mutation, unrelated staging, broad checkpoint authority, arbitrary commit proof, external model access, non-public body export, public sharing, hosting, launch-scope decision, or whole-system correctness.",
"body_html": "
Landing language must follow recorded evidence
\n
An agent's "I committed the fix" is a claim, and this component reads recorded landing examples and decides whether each one's evidence supports the words used to describe it. An agent that says "I committed the fix" is making a claim, and the claim is the part that is usually wrong. This component reads recorded examples of an agent finishing a piece of work and decides whether the evidence in each example supports the words used to describe it.
\n
Four inputs produce four metadata-only records
\n
It reads a fixture of landing attempts, a policy, a protocol, and broken cases, plus copied internal control bodies in bundle mode, and emits a metadata-only board, result, validation, and negative-coverage record. It takes a fixture of recorded landing attempts, a landing policy, a projection protocol, and a set of deliberately broken cases. In bundle mode it also reads the copied bodies of the internal control source files the model was tested against. It emits a metadata-only result set: a board, a result record, a validation record, and a fourth record confirming every expected negative case was caught.
\n
Recorded claims only, with no live mutation
\n
It runs no Git, stages nothing, moves no ledger, and excludes launch: it reads recorded claims and rejects the ones the evidence does not back. It does not run Git, stage a file, move a ledger, or authorize a launch. It reads recorded claims and rejects the ones the evidence does not back.
\n
Purpose
\n
Does the evidence support the landing language
\n
The narrow question is whether a recorded landing's evidence supports its language, since a loose check passes any row that merely describes itself well: a confident row with nothing behind it. The question is narrow. Given a recorded landing attempt, does the recorded evidence support the landing language? A loose check passes any row that describes itself well, which is exactly the failure mode: a confident row with nothing behind it.
\n
HEAD movement and validation order are separate floors
\n
A landed-commit claim needs recorded HEAD movement, a commit-path claim needs validation before the attempt, and every row still needs blocker and ledger evidence. The component turns two ordinary rules into rejections. A row that uses landed-commit language is rejected unless the recorded Git HEAD moved between before and after, so "I landed it" cannot stand on a HEAD that never advanced. A row on a commit path is rejected unless validation is recorded as preceding the commit attempt, so "it passed" cannot be back-filled after the fact. Blocker capture for metadata-blocked rows and work log completion for every row are the other two floors.
\n
Copied internal control bodies ground the check
\n
Six digest-provenanced internal control files let readers inspect the mechanics behind the landing rules instead of trusting a paraphrase of them. The check is also source-backed. The mechanics it tests rows against are not paraphrased. The six internal control source files are copied into the bundle with digest provenance, so a reader can see which code the model was checked against instead of trusting a summary.
\n
How it works
\n
Two entrypoints compose the same checks
\n
Fixture and bundle entrypoints both call _build_result, differing only in negative-case inclusion, manifest requirements, and fresh-record reuse. Two entry points drive the same builder. run reads the first-wave fixture with include_negative=True and writes four durable records. run_work_landing_bundle reads an exported bundle with include_negative=False, adds the source-module manifest check, and can reuse a fresh prior record when the input digest is unchanged. Both call _build_result, which composes four independent checks and one secret scan into a single verdict.
\n
Function
What it consumes
What it checks or rejects
validate_projection_protocol
projection_protocol.json
Requires the durable_agent_work_landing_replay_compound source pattern, at least three source refs, two projection result record refs, and three public runtime refs.
validate_landing_policy
landing_policy.json
Requires the four lanes in REQUIRED_LANE_IDS. Rejects a broad-checkpoint lane that skips operator authorization and any lane that sets release_authorized.
validate_work_landing_runs
work_landing_runs.json plus the nine negative files
Runs _validate_run_row over each row and each broken case.
validate_source_module_imports
source_module_manifest.json
Bundle mode only. Verifies copied internal control bodies by class, relation, digest, and required anchors.
\n
Row validation checks evidence and forbidden authority
\n
_validate_run_row requires validation, claimed-path, and completion refs while rejecting stationary HEADs, late validation, uncaptured blockers, private leaks, and authority overclaims._validate_run_row carries the row-level mechanism. For each recorded run it requires validation refs, claimed path refs, and a work log completion ref. It rejects a commit_claimed_landed row whose head_before equals head_after. It rejects a commit-path row (landing_status in COMMIT_ATTEMPT_STATUSES) that does not record validation_precedes_commit_attempt. It rejects a metadata-blocked row with no blocker capture ref, any row that claims live Git mutation, staged unrelated dirty paths, or launch-scope decision, and any row whose keys or strings leak a private body or non-public paths.
\n
Six copied modules must match digest and anchors
\n
validate_source_module_imports accepts the six exact-copy internal control modules only when their class, count, sha256, paths, and required anchors all match the manifest.validate_source_module_imports keeps the source claim honest. The manifest must declare copied_non_secret_macro_body, keep body_in_receipt false, and match its own module count. Each of the six rows names a internal control file under system/lib/, tools/meta/control/, or tools/meta/factory/, states an exact_copy relation, and carries a sha256. The check recomputes the digest of the copied target and rejects a mismatch, then rejects the row if any required anchor string is absent from the copied body. The copied files are workitem_runtime_entrypoint.py, work_landing_status.py, work_landing.py, mission_transaction_preflight.py, scoped_commit.py, and work_ledger.py.
\n
The final verdict requires every independent check
\n
_build_result passes only with all checks green, no missing negative case, and no secret hit, then projects four evidence counts into the board._build_result merges the four checks. The final status is a pass only when no expected negative case is missing, the secret scan has zero blocking hits, and every check passed. _board_from_result then projects four mechanic counts: claimed rows, landed-commit rows, validation-before-commit passes, and metadata-blocked rows. On the first-wave fixture that is three claimed rows, one landed-commit row, two commit-path rows that must record validation order, and one metadata-blocked row, with all nine negative cases observed. Every record carries the AUTHORITY_CEILING constant, and each finding and copied-module row sets body_in_receipt: false.
\nDiagram of the mechanism (6 steps).Diagram source & refs
Source refs
claim, validation order, HEAD advance, blocker capture
validate_work_landing_runs
class, relation, digest, anchors
validate_source_module_imports
flowchart TD Fixture["Fixture inputs projection, policy, runs, nine broken cases"] Source["Copied internal control bodies six files by digest"] Build["_build_result"] Runs["validate_work_landing_runs claim, validation order, HEAD advance, blocker capture"] Imports["validate_source_module_imports class, relation, digest, anchors"] Records["metadata-only records board, result, validation, coverage confirmation"] Fixture --> Build Source --> Imports Build --> Runs Build --> Imports Runs --> Records Imports --> Records
\n
Negative cases
\n
Nine broken rows define the refusal floor
\n
A passing fixture must catch all nine missing-evidence, ordering, mutation, boundary, blocker, launch, and private-path failures through the same row validator. The fixture ships nine broken rows, one per key in EXPECTED_NEGATIVE_CASES. A pass requires all nine to be caught. They are: a row with no validation ref, a row whose completion is missing, a landed-commit claim with no HEAD advance, validation recorded after a commit attempt, a row claiming live Git mutation authority, a row with no dirty-tree boundary, an uncaptured metadata blocker, a launch overclaim, and a non-public paths leak. Catching those refusals is part of the positive claim, not a separate concern.
\n
Prior Art Grounding
\n
Provenance fields stay distinct through landing
\n
W3C PROV and SLSA motivate keeping paths, validation, commit attempts, HEAD movement, blocker capture, and ledger completion as separately checked evidence. The posture comes from provenance and software supply-chain integrity work. The W3C PROV family models the entities, activities, and agents behind an artifact. SLSA tracks source, build process, provenance, and artifact integrity so a consumer can reason about where an artifact came from. This component borrows that separation for agent work landing: claimed paths, validation refs, commit attempts, HEAD before and after, blocker capture, and ledger completion are distinct evidence fields, checked one at a time.
A pass joins negative coverage and source integrity
\n
The test pass requires every negative case, a clean secret scan, and matching copied-source digests; the corpus pass establishes projection sign-off only. The first runs the component over the fixture and its nine broken cases. A pass means every negative case was caught, the secret scan was clean, and the copied source digests matched.
\n
Scope boundary
\n
Scope limit
\n
The supported claim is fixture conformance
\n
The fixture supports a conformance claim that every landing row carries the required paths, ordering, HEAD movement, blocker capture, completion, source digests, and negative-case coverage. The strongest claim the evidence supports is a conformance claim. On the public fixture and exported bundle this component shows that each recorded landing row carries the required evidence and ordering: claimed paths, validation before commit, a HEAD advance behind any landed-commit language, blocker capture for metadata-blocked rows, and work log completion throughout, with six internal control bodies copied by matching digest and nine negative cases caught.
\n
The recorded fixture grants no live authority
\n
The scope limit excludes live Git or ledger mutation, proof that a real commit landed, broad checkpoints, external model access, public sharing, launch, and whole-system correctness. The proof boundary stops at the recorded fixture and its records. The scope limit is the AUTHORITY_CEILING constant: no live Git mutation, no proof that any commit truly landed in a real repository, no live ledger mutation, no broad checkpoint authority, no source-file changes, no external model access, no public sharing or launch, and no whole-system correctness. Broader work-continuity behavior lives in sibling components such as bridge_phase_continuity_runtime and in the internal control source files themselves. This module can cite their copied bodies as evidence; it cannot inherit their authority.
\n",
"components": [
{
"anchor": "component-durable_agent_work_landing_replay",
"component_id": "durable_agent_work_landing_replay",
"public_label": "Durable Agent Work Landing Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 2,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 5
},
"link_count": 11,
"links": [
{
"id": "durable_agent_work_landing_replay",
"kind": "organ",
"kind_label": "component",
"label": "Durable Agent Work Landing Replay",
"relation": "explains",
"url": "component-durable-agent-work-landing-replay.html"
},
{
"id": "mechanism.durable_agent_work_landing_replay.validates_public_work_landing_replay_contract",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public work landing replay contract",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-durable-agent-work-landing-replay-validates-public-work-landing-replay-contract"
},
{
"id": "P-10",
"kind": "principle",
"kind_label": "principle",
"label": "Do not land effects without compensation",
"relation": "governed by",
"url": "doctrine.html#dcard-p-10"
},
{
"id": "P-14",
"kind": "principle",
"kind_label": "principle",
"label": "Carry basis and provenance together",
"relation": "governed by",
"url": "doctrine.html#dcard-p-14"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
},
{
"id": "concept.work_landing_and_continuity_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "work landing and continuity controls as transaction-scoped coordination evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-work-landing-and-continuity-control-bundle"
},
{
"id": "paper_module.bridge_phase_continuity_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Bridge Phase Continuity Runtime",
"relation": "depends on",
"url": "paper-module-bridge-phase-continuity-runtime.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.agent_closeout_faithfulness_audit",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Completion Faithfulness Audit",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[12:paper_module.agent_closeout_faithfulness_audit].depends_on",
"url": "paper-module-agent-closeout-faithfulness-audit.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[20:paper_module.mission_transaction_work_spine].depends_on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.work_landing_control_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Work Landing Control Spine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[17:paper_module.work_landing_control_spine].depends_on",
"url": "paper-module-work-landing-control-spine.html"
}
],
"paper_module_id": "paper_module.durable_agent_work_landing_replay",
"reader_packet": {
"abstract": "The public durable work-landing replay fixture validates recorded agent landing rows, copied source internal control source bodies, validation-before-commit ordering, HEAD-advance evidence, blocker capture, and work log completion without performing live Git work.",
"authority_ceiling": "Public synthetic replay result records and copied source work-landing/internal control source bodies only; no live Git mutation, unrelated staging, broad checkpoint authority, arbitrary commit proof, external model access, non-public body export, public sharing, hosting, launch-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.durable_agent_work_landing_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-durable-agent-work-landing-replay-diagram-1",
"html_url": "docs/paper-module-durable-agent-work-landing-replay.html#paper-module-durable-agent-work-landing-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Fixture[\"Fixture inputs projection, policy, runs, nine broken cases\"] Source[\"Copied internal control bodies six files by digest\"] Build[\"_build_result\"] Runs[\"validate_work_landing_runs claim, validation order, HEAD advance, blocker capture\"] Imports[\"validate_source_module_imports class, relation, digest, anchors\"] Records[\"metadata-only records board, result, validation, coverage confirmation\"] Fixture --> Build Source --> Imports Build --> Runs Build --> Imports Runs --> Records Imports --> Records",
"source_format": "mermaid",
"source_hash": "sha256:f957a1e30e2c8131dc4f622ed61a95dee48dfd80f5ff6ffaf8c97333c13e521f",
"source_line_end": 88,
"source_line_start": 73
}
],
"grouped_relation_feeds": [
{
"count": 2,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.bridge_phase_continuity_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Bridge Phase Continuity Runtime",
"relation": "depends on",
"url": "paper-module-bridge-phase-continuity-runtime.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "durable_agent_work_landing_replay",
"kind": "organ",
"kind_label": "component",
"label": "Durable Agent Work Landing Replay",
"relation": "explains",
"url": "component-durable-agent-work-landing-replay.html"
},
{
"id": "mechanism.durable_agent_work_landing_replay.validates_public_work_landing_replay_contract",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public work landing replay contract",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-durable-agent-work-landing-replay-validates-public-work-landing-replay-contract"
}
],
"relation": "explains"
},
{
"count": 6,
"links": [
{
"id": "P-10",
"kind": "principle",
"kind_label": "principle",
"label": "Do not land effects without compensation",
"relation": "governed by",
"url": "doctrine.html#dcard-p-10"
},
{
"id": "P-14",
"kind": "principle",
"kind_label": "principle",
"label": "Carry basis and provenance together",
"relation": "governed by",
"url": "doctrine.html#dcard-p-14"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "concept.work_landing_and_continuity_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "work landing and continuity controls as transaction-scoped coordination evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-work-landing-and-continuity-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.durable_agent_work_landing_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 143,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/durable_agent_work_landing_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "W3C PROV",
"section_id": "provenance-fields-stay-distinct-through-landing",
"url": "https://www.w3.org/TR/prov-overview/"
},
{
"kind": "external_reference",
"label": "SLSA",
"section_id": "provenance-fields-stay-distinct-through-landing",
"url": "https://slsa.dev/spec/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "An agent's \"I committed the fix\" is a claim, and this component reads recorded landing examples and decides whether each one's evidence supports the words used to describe it. An agent that says \"I committed the fix\" is making a claim, and the claim is the part that is usually wrong. This component reads recorded examples of an agent finishing a piece of work and decides whether the evidence in each example supports the words used to describe it.",
"level": 4,
"order": 1,
"section_id": "landing-language-must-follow-recorded-evidence",
"source_line_end": 7,
"source_line_start": 3,
"title": "Landing language must follow recorded evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It reads a fixture of landing attempts, a policy, a protocol, and broken cases, plus copied internal control bodies in bundle mode, and emits a metadata-only board, result, validation, and negative-coverage record. It takes a fixture of recorded landing attempts, a landing policy, a projection protocol, and a set of deliberately broken cases. In bundle mode it also reads the copied bodies of the internal control source files the model was tested against. It emits a metadata-only result set: a board, a result record, a validation record, and a fourth record confirming every expected negative case was caught.",
"level": 4,
"order": 2,
"section_id": "four-inputs-produce-four-body-free-records",
"source_line_end": 12,
"source_line_start": 8,
"title": "Four inputs produce four metadata-only records"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It runs no Git, stages nothing, moves no ledger, and excludes launch: it reads recorded claims and rejects the ones the evidence does not back. It does not run Git, stage a file, move a ledger, or authorize a launch. It reads recorded claims and rejects the ones the evidence does not back.",
"level": 4,
"order": 3,
"section_id": "recorded-claims-only-with-no-live-mutation",
"source_line_end": 17,
"source_line_start": 13,
"title": "Recorded claims only, with no live mutation"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 19,
"source_line_start": 18,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The narrow question is whether a recorded landing's evidence supports its language, since a loose check passes any row that merely describes itself well: a confident row with nothing behind it. The question is narrow. Given a recorded landing attempt, does the recorded evidence support the landing language? A loose check passes any row that describes itself well, which is exactly the failure mode: a confident row with nothing behind it.",
"level": 4,
"order": 5,
"section_id": "does-the-evidence-support-the-landing-language",
"source_line_end": 24,
"source_line_start": 20,
"title": "Does the evidence support the landing language"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A landed-commit claim needs recorded HEAD movement, a commit-path claim needs validation before the attempt, and every row still needs blocker and ledger evidence.** The component turns two ordinary rules into rejections. A row that uses landed-commit language is rejected unless the recorded Git HEAD moved between before and after, so \"I landed it\" cannot stand on a HEAD that never advanced. A row on a commit path is rejected unless validation is recorded as preceding the commit attempt, so \"it passed\" cannot be back-filled after the fact. Blocker capture for metadata-blocked rows and work log completion for every row are the other two floors.",
"level": 4,
"order": 6,
"section_id": "head-movement-and-validation-order-are-separate-floors",
"source_line_end": 31,
"source_line_start": 25,
"title": "HEAD movement and validation order are separate floors"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Six digest-provenanced internal control files let readers inspect the mechanics behind the landing rules instead of trusting a paraphrase of them.** The check is also source-backed. The mechanics it tests rows against are not paraphrased. The six internal control source files are copied into the bundle with digest provenance, so a reader can see which code the model was checked against instead of trusting a summary.",
"level": 4,
"order": 7,
"section_id": "copied-control-plane-bodies-ground-the-check",
"source_line_end": 37,
"source_line_start": 32,
"title": "Copied internal control bodies ground the check"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 39,
"source_line_start": 38,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "Fixture and bundle entrypoints both call build_result, differing only in negative-case inclusion, manifest requirements, and fresh-record reuse.** Two entry points drive the same builder. run reads the first-wave fixture with includenegative=True and writes four durable records. runworklandingbundle reads an exported bundle with includenegative=False, adds the source-module manifest check, and can reuse a fresh prior record when the input digest is unchanged. Both call build_result, which composes four independent checks and one secret scan into a single verdict.",
"level": 4,
"order": 9,
"section_id": "two-entrypoints-compose-the-same-checks",
"source_line_end": 52,
"source_line_start": 40,
"title": "Two entrypoints compose the same checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validaterun_row requires validation, claimed-path, and completion refs while rejecting stationary HEADs, late validation, uncaptured blockers, private leaks, and authority overclaims.** validaterunrow carries the row-level mechanism. For each recorded run it requires validation refs, claimed path refs, and a work log completion ref. It rejects a commitclaimedlanded row whose headbefore equals headafter. It rejects a commit-path row (landingstatus in COMMITATTEMPTSTATUSES) that does not record validationprecedescommit_attempt. It rejects a metadata-blocked row with no blocker capture ref, any row that claims live Git mutation, staged unrelated dirty paths, or launch-scope decision, and any r\u2026",
"level": 4,
"order": 10,
"section_id": "row-validation-checks-evidence-and-forbidden-authority",
"source_line_end": 59,
"source_line_start": 53,
"title": "Row validation checks evidence and forbidden authority"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatesourcemoduleimports accepts the six exact-copy internal control modules only when their class, count, sha256, paths, and required anchors all match the manifest.** validatesourcemoduleimports keeps the source claim honest. The manifest must declare copiednonsecretmacrobody, keep bodyinreceipt false, and match its own module count. Each of the six rows names a internal control file under system/lib/, tools/meta/control/, or tools/meta/factory/, states an exactcopy relation, and carries a sha256. The check recomputes the digest of the copied target and rejects a mismatch, then rejects the row if any required anchor string is absent from the copied body. The copied files are workitemru\u2026",
"level": 4,
"order": 11,
"section_id": "six-copied-modules-must-match-digest-and-anchors",
"source_line_end": 66,
"source_line_start": 60,
"title": "Six copied modules must match digest and anchors"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "buildresult passes only with all checks green, no missing negative case, and no secret hit, then projects four evidence counts into the board.** buildresult merges the four checks. The final status is a pass only when no expected negative case is missing, the secret scan has zero blocking hits, and every check passed. boardfromresult then projects four mechanic counts: claimed rows, landed-commit rows, validation-before-commit passes, and metadata-blocked rows. On the first-wave fixture that is three claimed rows, one landed-commit row, two commit-path rows that must record validation order, and one metadata-blocked row, with all nine negative cases observed. Every record carries the AUTHOR\u2026",
"level": 4,
"order": 12,
"section_id": "the-final-verdict-requires-every-independent-check",
"source_line_end": 89,
"source_line_start": 67,
"title": "The final verdict requires every independent check"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 91,
"source_line_start": 90,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A passing fixture must catch all nine missing-evidence, ordering, mutation, boundary, blocker, launch, and private-path failures through the same row validator.** The fixture ships nine broken rows, one per key in EXPECTEDNEGATIVECASES. A pass requires all nine to be caught. They are: a row with no validation ref, a row whose completion is missing, a landed-commit claim with no HEAD advance, validation recorded after a commit attempt, a row claiming live Git mutation authority, a row with no dirty-tree boundary, an uncaptured metadata blocker, a launch overclaim, and a non-public paths leak. Catching those refusals is part of the positive claim, not a separate concern.",
"level": 4,
"order": 14,
"section_id": "nine-broken-rows-define-the-refusal-floor",
"source_line_end": 98,
"source_line_start": 92,
"title": "Nine broken rows define the refusal floor"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 100,
"source_line_start": 99,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**W3C PROV and SLSA motivate keeping paths, validation, commit attempts, HEAD movement, blocker capture, and ledger completion as separately checked evidence.** The posture comes from provenance and software supply-chain integrity work. The W3C PROV family models the entities, activities, and agents behind an artifact. SLSA tracks source, build process, provenance, and artifact integrity so a consumer can reason about where an artifact came from. This component borrows that separation for agent work landing: claimed paths, validation refs, commit attempts, HEAD before and after, blocker capture, and ledger completion are distinct evidence fields, checked one at a time.",
"level": 4,
"order": 16,
"section_id": "provenance-fields-stay-distinct-through-landing",
"source_line_end": 106,
"source_line_start": 101,
"title": "Provenance fields stay distinct through landing"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 108,
"source_line_start": 107,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "**The focused test exercises the fixture and nine broken rows, while the corpus record.** Run these from the repository root without mutating durable records or generated projections:",
"level": 4,
"order": 18,
"section_id": "two-read-only-commands-verify-behavior-and-parity",
"source_line_end": 120,
"source_line_start": 109,
"title": "Two read-only commands verify behavior and parity"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The test pass requires every negative case, a clean secret scan, and matching copied-source digests; the corpus pass establishes projection sign-off only.** The first runs the component over the fixture and its nine broken cases. A pass means every negative case was caught, the secret scan was clean, and the copied source digests matched.",
"level": 4,
"order": 19,
"section_id": "a-pass-joins-negative-coverage-and-source-integrity",
"source_line_end": 126,
"source_line_start": 121,
"title": "A pass joins negative coverage and source integrity"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 128,
"source_line_start": 127,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The fixture supports a conformance claim that every landing row carries the required paths, ordering, HEAD movement, blocker capture, completion, source digests, and negative-case coverage.** The strongest claim the evidence supports is a conformance claim. On the public fixture and exported bundle this component shows that each recorded landing row carries the required evidence and ordering: claimed paths, validation before commit, a HEAD advance behind any landed-commit language, blocker capture for metadata-blocked rows, and work log completion throughout, with six internal control bodies copied by matching digest and nine negative cases caught.",
"level": 4,
"order": 21,
"section_id": "the-supported-claim-is-fixture-conformance",
"source_line_end": 135,
"source_line_start": 129,
"title": "The supported claim is fixture conformance"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The scope limit excludes live Git or ledger mutation, proof that a real commit landed, broad checkpoints, external model access, public sharing, launch, and whole-system correctness.** The proof boundary stops at the recorded fixture and its records. The scope limit is the AUTHORITYCEILING constant: no live Git mutation, no proof that any commit truly landed in a real repository, no live ledger mutation, no broad checkpoint authority, no source-file changes, no external model access, no public sharing or launch, and no whole-system correctness. Broader work-continuity behavior lives in sibling components such as bridgephasecontinuityruntime and in the internal control source files themsel\u2026",
"level": 4,
"order": 22,
"section_id": "the-recorded-fixture-grants-no-live-authority",
"source_line_end": 142,
"source_line_start": 136,
"title": "The recorded fixture grants no live authority"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/durable_agent_work_landing_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public durable work-landing replay fixture validates recorded agent landing rows, copied source internal control source bodies, validation-before-commit ordering, HEAD-advance evidence, blocker capture, and work log completion without performing live Git work."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "durable_agent_work_landing_replay.py",
"path": "src/microcosm_core/organs/durable_agent_work_landing_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for durable work-landing replay validation, exported bundle source-module digest checks, negative-case enforcement, scope limits, and metadata-only public result records.",
"symbols": [
"run",
"run_work_landing_bundle",
"validate_source_module_imports",
"validate_projection_protocol",
"validate_landing_policy",
"validate_work_landing_runs",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/durable_agent_work_landing_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/durable_agent_work_landing_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/durable_agent_work_landing_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Durable Agent Work-Landing Replay is the public work-spine replay contract for agent landing claims. It checks claimed paths, owner-native validation refs, commit-attempt order, HEAD-before/after evidence, metadata-blocked rows, work log finalizer evidence, nine negative cases, and six exact copied source bodies while keeping raw diffs, non-public paths, model-output data, and source bodies out of result records.",
"summary_line": "The public durable work-landing replay fixture validates recorded agent landing rows, copied source internal control source bodies, validation-before-commit ordering, HEAD-advance evidence, blocker capture, and work log completion without performing live Git work.",
"title": "Durable Agent Work-Landing Replay"
},
{
"anchor": "paper-module-work-landing-control-spine",
"authority_ceiling": "Copied source internal control source bundle, validator result record, source-manifest digest evidence, required-anchor evidence, no-live-mutation contract flags, and secret-exclusion evidence only; no live Git mutation, live work log or work log mutation, claim launch, shared-index mutation, private-index commit execution, broad staging, external model access, account secret export, public sharing, launch-scope decision, or whole-system correctness.",
"body_html": "
With many agents editing at once, the hard problem is landing one finished commit without sweeping up a neighbour's half-done edits, solved by internal control modules copied into a public bundle. The larger source project this public folder draws from is edited by several agents at the same time. Its hardest everyday problem is dull: how does one agent land a small, finished commit without sweeping up a neighbour's half-done edits in the same files, and without claiming more than it actually did. The answer lives in a handful of internal control source modules. A separate import step copies those bodies, non-secret, into a public bundle so a reader can inspect the real mechanics instead of taking a description on trust.
\n
work_landing_control_spine guards that copy, confirming five copied modules are exact, their load-bearing logic is present, the contract claims no live authority, and nothing secret leaked.work_landing_control_spine is the check that guards that copy. It reads a bundle of five copied source modules plus the manifests and runtime contract that describe them, confirms the copies are exact, confirms the load-bearing logic is still present, confirms the runtime contract claims no live authority, and confirms nothing secret leaked in. It then writes one result record.
\n
It is a witness over copied source, not a runner.
\n
Purpose
\n
A described control plane is easy to fake
\n
A prose description of a control plane is easy to fake, so this component copies the actual bodies and asks whether they are genuine, load-bearing, and within their limits. A description of a control plane is easy to fake. A reader cannot tell whether "we validate scoped commits" means the real actuator or a paragraph of prose. So this component copies the actual bodies into the public bundle and then answers one narrow question: are these copied bodies the genuine ones, is their load-bearing logic still present, and does the bundle stay inside its stated limits.
\n
The copied set is five modules:
\n
tools/meta/control/work_landing.py (the command parser and admission gate)
system/lib/work_landing_status.py (controller action order and finalizer policy)
tools/meta/control/scoped_commit.py (the private-index scoped-commit actuator)
\n
Scoped commit and controller ordering
\n
Two ideas matter: scoped_commit.py commits from a throwaway private index that never touches a neighbour's changes, and work_landing_status.py fixes the controller action order. Two ideas in that source are worth a reader's attention. One is the scoped commit: scoped_commit.py builds a throwaway Git index from the current HEAD, stages only the exact paths an agent declares it owns, writes a tree from that private index, and commits against the captured parent with a compare-and-swap on the branch ref. The shared .git/index is never written. An agent cannot pick up a neighbour's dirty changes, because those changes were never in the index it committed from. The other is ordering: work_landing_status.py fixes a sequence of controller actions so claims launch only after the session is finalised and convergence recomputes only after claims launch. The check confirms both are still present in the copied bodies, which is what separates an honest copy from a stale bag of hashes.
\n
How it works
\n
The validate_work_landing_control_bundle entry point
\n
validate_work_landing_control_bundle runs the stages in order and returns one result record, wrapped by main and leaning on shared JSON, result record, and secret-scan helpers. The check lives in src/microcosm_core/macro_tools/work_landing_control_spine.py. The entry point is validate_work_landing_control_bundle, which runs the stages below in order and returns a single result record. main and build_parser wrap it for the command line. It depends on three shared helpers: the strict JSON reader in microcosm_core.schemas, the atomic record writer in microcosm_core.receipts, and the secret scanner in microcosm_core.secret_exclusion_scan.
\n
The stages run in this order.
\n
Loading the three JSON inputs
\n
First _load_json_input reads the bundle manifest, source manifest, and runtime contract, turning a missing file or non-object JSON into a finding rather than a crash. First it loads the three JSON inputs with _load_json_input: the bundle manifest, the source module manifest, and the runtime contract. A missing file or non-object JSON becomes a finding, not a crash.
\n
_source_manifest_summary checks copy fidelity
\n
_source_manifest_summary recomputes a SHA-256 and line count for each of the seven required inputs against the manifest, and confirms every material and import class. Then _source_manifest_summary checks copy fidelity. For each of the seven required inputs (five source bodies plus the two JSON files, named in REQUIRED_INPUTS) it computes a SHA-256 with _file_sha256 and a line count with _line_count, then compares them against the manifest's expected values. It also confirms each row's material_class is one of ALLOWED_MATERIAL_CLASSES and that each source body is classified copied_non_secret_macro_body (SOURCE_IMPORT_CLASS).
\n
_anchor_summary confirms load-bearing anchors
\n
_anchor_summary confirms every required anchor string in REQUIRED_SOURCE_REFS still survives in each copied body, counting any missing anchor rather than dropping it. Then _anchor_summary reads each copied body and confirms every required anchor string is still in the text. The anchors are declared in REQUIRED_SOURCE_REFS, keyed by module. For example work_landing.py must still contain build_parser, admission-check, and begin; scoped_commit.py must still contain perform_scoped_commit, private-index scoped commit, and the line stating it never invokes git commit against the shared index. A missing anchor is counted, not silently dropped.
\n
_contract_summary checks the scope limit
\n
_contract_summary requires all thirteen FALSE_AUTHORITY_FLAGS to be exactly False, treating any other value as an overclaim, and confirms the required import classifications. Then _contract_summary checks the scope limit. Every flag in FALSE_AUTHORITY_FLAGS (thirteen of them, from live_git_mutation_authorized through publication_authorized and release_authorized) must be exactly False in the contract. Any flag that is not False is an overclaim. The same stage confirms the contract declares the required import classifications (REQUIRED_CLASSIFICATIONS).
\n
scan_paths runs the secret scanner
\n
scan_paths runs the shared secret scanner over all seven inputs, turning any blocking hit into a finding. Then scan_paths runs the shared secret scanner over all seven inputs. Any blocking hit is a finding.
\n
Assembling the metadata-only result record
\n
Finally the record is assembled, pass only when there are no findings, carrying digests, anchor and authority rows, and body_in_receipt: false so no body is copied out. Finally the record is assembled. status is pass when there are no findings and blocked otherwise. The record carries refs, digests, line counts, anchor rows, authority rows, material-class counts, and findings, and it sets body_in_receipt: false so no source body is copied into the output. The record is written atomically as exported_work_landing_control_bundle_validation_result.json and also printed to stdout.
\n
Function
Role
validate_work_landing_control_bundle
Runs the stages, assembles and writes the result record
_load_json_input
Loads and shape-checks each JSON input, or records a finding
_source_manifest_summary
Digest, line-count, material-class, and import-class parity
_anchor_summary
Confirms required anchor strings survive in each copied body
_contract_summary
Confirms all FALSE_AUTHORITY_FLAGS are false and classifications present
build_parser / main
Command-line wrapper returning exit 0 on pass, 1 on block
\nDiagram of the mechanism (7 steps).Diagram source & refs
flowchart TD A["Copied bundle 5 bodies + manifest + contract"] --> B["Load JSON inputs _load_json_input"] B --> C["Copy parity _source_manifest_summary"] C --> D["Anchor scan _anchor_summary"] D --> E["Scope limit _contract_summary"] E --> F["Secret scan scan_paths"] F --> G["Result record status pass or blocked"]
\n
Negative cases
\n
Five rejections from one-field mutations
\n
The focused tests drive five rejections by mutating one field at a time, each producing a blocked status with a stable finding code, and the same shape covers other drifts. The focused tests drive five rejections from the accepted bundle by mutating one field at a time. A digest changed in the manifest gives status blocked with SOURCE_DIGEST_MISMATCH. A source body relabelled to a private material class gives MATERIAL_CLASS_NOT_ALLOWED. A body reclassified away from copied_non_secret_macro_body gives SOURCE_IMPORT_CLASS_MISMATCH. Dropping secret_exclusion from the contract's required classifications gives CLASSIFICATION_FLOOR_MISSING. Setting live_git_mutation_authorized to True in the contract gives AUTHORITY_CEILING_OVERCLAIM. The check emits the same shape of finding for a missing anchor (SOURCE_ANCHOR_MISSING), a missing input (MISSING_REQUIRED_SOURCE_INPUT), a line-count drift (SOURCE_LINE_COUNT_MISMATCH), and a secret hit (SECRET_EXCLUSION_BLOCKING_HIT).
\n
Prior Art Grounding
\n
Git index and durable-workflow ancestry
\n
The mechanics come from Git's index separating selected changes and from durable-workflow action ordering, imported here as an auditable public check with no power to act. The mechanics copied here come from ordinary version-control and workflow engineering. Git's index already separates the changes you have selected from the rest of a dirty worktree, which is the practical ancestor of scoped path ownership. Durable-workflow systems show the value of a recorded action order and deterministic replay for long-running work. This component imports those ideas as a public check: digest parity, anchor presence, a no-live-mutation contract, and a secret scan make the copied control plane auditable without granting it any power to act.
\n
Git staging and index workflow: https://git-scm.com/book/en/v2/Git-Basics-Recording-Changes-to-the-Repository
Durable workflow event history and replay: https://docs.temporal.io/workflows
A pass writes the validation result record and means the copied bodies matched the manifest, every anchor was present, all authority flags stayed false, and the secret scan found nothing. A pass writes receipts/first_wave/work_landing_control_spine/exported_work_landing_control_bundle_validation_result.json and means the copied bodies matched the manifest, every required anchor was present, all authority flags stayed false, and the secret scan found nothing. The focused regression suite that pins the pass case and the four rejections above is:
For a given bundle, the five copied bodies match their digests and line counts, the anchor logic is present, the contract declares no live authority, and no secret is present. The strongest claim the evidence supports is narrow: for a given bundle, the five copied internal control bodies match their recorded digests and line counts, the named anchor logic is still present, the runtime contract declares no live authority, and no secret material is present. That is the whole proof boundary.
\n
What the scope limit stops
\n
The check runs no live Git, mutates no ledger, releases no claim, and calls no provider, so a passing record proves only a faithful copy, never that a future landing is safe. The scope limit stops there. This check does not run live Git, does not mutate any ledger, does not launch a claim, does not stage broadly, does not execute a private-index commit, does not call a provider, does not export account secrets, and does not publish or authorize a launch. A passing record is evidence that a copy is faithful and honest about its limits. It is not evidence that any future landing operation is safe, and it grants no power to perform one.
\n",
"components": [
{
"anchor": "component-macro_projection_import_protocol",
"component_id": "macro_projection_import_protocol",
"public_label": "Source Projection Import Protocol"
}
],
"doctrine_feed": {
"counts": {
"axiom": 2,
"concept": 2,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 5
},
"link_count": 14,
"links": [
{
"id": "macro_projection_import_protocol",
"kind": "organ",
"kind_label": "component",
"label": "Source Projection Import Protocol",
"relation": "explains",
"url": "component-macro-projection-import-protocol.html"
},
{
"id": "mechanism.macro_projection_import_protocol.validates_public_macro_projection_imports",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public source projection imports",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-macro-projection-import-protocol-validates-public-macro-projection-imports"
},
{
"id": "P-10",
"kind": "principle",
"kind_label": "principle",
"label": "Do not land effects without compensation",
"relation": "governed by",
"url": "doctrine.html#dcard-p-10"
},
{
"id": "P-14",
"kind": "principle",
"kind_label": "principle",
"label": "Carry basis and provenance together",
"relation": "governed by",
"url": "doctrine.html#dcard-p-14"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
},
{
"id": "concept.work_landing_and_continuity_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "work landing and continuity controls as transaction-scoped coordination evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-work-landing-and-continuity-control-bundle"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.durable_agent_work_landing_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Durable Agent Work-Landing Replay",
"relation": "depends on",
"url": "paper-module-durable-agent-work-landing-replay.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch9_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 9 Source Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[70:paper_module.batch9_macro_engines_capsule].depends_on",
"url": "paper-module-batch9-macro-engines-capsule.html"
},
{
"id": "paper_module.concurrency_mission_control",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Concurrency Mission Control",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[47:paper_module.concurrency_mission_control].depends_on",
"url": "paper-module-concurrency-mission-control.html"
}
],
"paper_module_id": "paper_module.work_landing_control_spine",
"reader_packet": {
"abstract": "The public work-landing control spine validates copied work-landing internal control source bodies without authorizing live Git, ledger, claim-launch, private-index, public sharing, or launch operations.",
"authority_ceiling": "Copied source internal control source bundle, validator result record, source-manifest digest evidence, required-anchor evidence, no-live-mutation contract flags, and secret-exclusion evidence only; no live Git mutation, live work log or work log mutation, claim launch, shared-index mutation, private-index commit execution, broad staging, external model access, account secret export, public sharing, launch-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.work_landing_control_spine.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-work-landing-control-spine-diagram-1",
"html_url": "docs/paper-module-work-landing-control-spine.html#paper-module-work-landing-control-spine-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Copied bundle 5 bodies + manifest + contract\"] --> B[\"Load JSON inputs _load_json_input\"] B --> C[\"Copy parity _source_manifest_summary\"] C --> D[\"Anchor scan _anchor_summary\"] D --> E[\"Scope limit _contract_summary\"] E --> F[\"Secret scan scan_paths\"] F --> G[\"Result record status pass or blocked\"]",
"source_format": "mermaid",
"source_hash": "sha256:bf0cc3d779c8e9839a4fa7dabca8aeab1b09283b455e175fab77637711849cbb",
"source_line_end": 156,
"source_line_start": 148
}
],
"grouped_relation_feeds": [
{
"count": 2,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.durable_agent_work_landing_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Durable Agent Work-Landing Replay",
"relation": "depends on",
"url": "paper-module-durable-agent-work-landing-replay.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "macro_projection_import_protocol",
"kind": "organ",
"kind_label": "component",
"label": "Source Projection Import Protocol",
"relation": "explains",
"url": "component-macro-projection-import-protocol.html"
},
{
"id": "mechanism.macro_projection_import_protocol.validates_public_macro_projection_imports",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public source projection imports",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-macro-projection-import-protocol-validates-public-macro-projection-imports"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-10",
"kind": "principle",
"kind_label": "principle",
"label": "Do not land effects without compensation",
"relation": "governed by",
"url": "doctrine.html#dcard-p-10"
},
{
"id": "P-14",
"kind": "principle",
"kind_label": "principle",
"label": "Carry basis and provenance together",
"relation": "governed by",
"url": "doctrine.html#dcard-p-14"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "concept.work_landing_and_continuity_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "work landing and continuity controls as transaction-scoped coordination evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-work-landing-and-continuity-control-bundle"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.work_landing_control_spine",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 244,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/work_landing_control_spine.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**With many agents editing at once, the hard problem is landing one finished commit without sweeping up a neighbour's half-done edits, solved by internal control modules copied into a public bundle.** The larger source project this public folder draws from is edited by several agents at the same time. Its hardest everyday problem is dull: how does one agent land a small, finished commit without sweeping up a neighbour's half-done edits in the same files, and without claiming more than it actually did. The answer lives in a handful of internal control source modules. A separate import step copies those bodies, non-secret, into a public bundle so a reader can inspect the real mechanics instea\u2026",
"level": 4,
"order": 1,
"section_id": "with-many-agents-editing-at-once",
"source_line_end": 14,
"source_line_start": 3,
"title": "With many agents editing at once"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "worklandingcontrolspine guards that copy, confirming five copied modules are exact, their load-bearing logic is present, the contract claims no live authority, and nothing secret leaked. worklandingcontrolspine is the check that guards that copy. It reads a bundle of five copied source modules plus the manifests and runtime contract that describe them, confirms the copies are exact, confirms the load-bearing logic is still present, confirms the runtime contract claims no live authority, and confirms nothing secret leaked in. It then writes one result record. It is a witness over copied source, not a runner.",
"level": 4,
"order": 2,
"section_id": "work-landing-control-spine-guards-that-copy",
"source_line_end": 26,
"source_line_start": 15,
"title": "`work_landing_control_spine` guards that copy"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 3,
"section_id": "purpose",
"source_line_end": 28,
"source_line_start": 27,
"title": "Purpose"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**A prose description of a control plane is easy to fake, so this component copies the actual bodies and asks whether they are genuine, load-bearing, and within their limits.** A description of a control plane is easy to fake. A reader cannot tell whether \"we validate scoped commits\" means the real actuator or a paragraph of prose. So this component copies the actual bodies into the public bundle and then answers one narrow question: are these copied bodies the genuine ones, is their load-bearing logic still present, and does the bundle stay inside its stated limits. The copied set is five modules: tools/meta/control/work_landing.py (the command parser and admission gate) system/lib/worklan\u2026",
"level": 4,
"order": 4,
"section_id": "a-described-control-plane-is-easy-to-fake",
"source_line_end": 47,
"source_line_start": 29,
"title": "A described control plane is easy to fake"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Two ideas matter: scopedcommit.py commits from a throwaway private index that never touches a neighbour's changes, and worklandingstatus.py fixes the controller action order.** Two ideas in that source are worth a reader's attention. One is the scoped commit: scoped_commit.py builds a throwaway Git index from the current HEAD, stages only the exact paths an agent declares it owns, writes a tree from that private index, and commits against the captured parent with a compare-and-swap on the branch ref. The shared .git/index is never written. An agent cannot pick up a neighbour's dirty changes, because those changes were never in the index it committed from. The other is ordering: worklandings\u2026",
"level": 4,
"order": 5,
"section_id": "scoped-commit-and-controller-ordering",
"source_line_end": 63,
"source_line_start": 48,
"title": "Scoped commit and controller ordering"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 6,
"section_id": "how-it-works",
"source_line_end": 65,
"source_line_start": 64,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateworklandingcontrol_bundle runs the stages in order and returns one result record, wrapped by main and leaning on shared JSON, result record, and secret-scan helpers.** The check lives in src/microcosmcore/macrotools/worklandingcontrol_spine.py. The entry point is validateworklandingcontrolbundle, which runs the stages below in order and returns a single result record. main and build_parser wrap it for the command line. It depends on three shared helpers: the strict JSON reader in microcosmcore.schemas, the atomic record writer in microcosmcore.result records, and the secret scanner in microcosmcore.secretexclusion_scan. The stages run in this order.",
"level": 4,
"order": 7,
"section_id": "the-validate-work-landing-control-bundle-entry-point",
"source_line_end": 79,
"source_line_start": 66,
"title": "The `validate_work_landing_control_bundle` entry point"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "First loadjsoninput reads the bundle manifest, source manifest, and runtime contract, turning a missing file or non-object JSON into a finding rather than a crash. First it loads the three JSON inputs with loadjsoninput: the bundle manifest, the source module manifest, and the runtime contract. A missing file or non-object JSON becomes a finding, not a crash.",
"level": 4,
"order": 8,
"section_id": "loading-the-three-json-inputs",
"source_line_end": 87,
"source_line_start": 80,
"title": "Loading the three JSON inputs"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourcemanifest_summary recomputes a SHA-256 and line count for each of the seven required inputs against the manifest, and confirms every material and import class. Then sourcemanifestsummary checks copy fidelity. For each of the seven required inputs (five source bodies plus the two JSON files, named in REQUIREDINPUTS) it computes a SHA-256 with file_sha256 and a line count with linecount, then compares them against the manifest's expected values. It also confirms each row's materialclass is one of ALLOWEDMATERIAL_CLASSES and that each source body is classified copiednonsecretmacrobody (SOURCEIMPORTCLASS).",
"level": 4,
"order": 9,
"section_id": "source-manifest-summary-checks-copy-fidelity",
"source_line_end": 98,
"source_line_start": 88,
"title": "`_source_manifest_summary` checks copy fidelity"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "anchorsummary confirms every required anchor string in REQUIREDSOURCEREFS still survives in each copied body, counting any missing anchor rather than dropping it. Then anchor_summary reads each copied body and confirms every required anchor string is still in the text. The anchors are declared in REQUIREDSOURCEREFS, keyed by module. For example worklanding.py must still contain buildparser, admission-check, and begin; scoped_commit.py must still contain performscopedcommit, private-index scoped commit, and the line stating it never invokes git commit against the shared index. A missing anchor is counted, not silently dropped.",
"level": 4,
"order": 10,
"section_id": "anchor-summary-confirms-load-bearing-anchors",
"source_line_end": 110,
"source_line_start": 99,
"title": "`_anchor_summary` confirms load-bearing anchors"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "contractsummary requires all thirteen FALSEAUTHORITYFLAGS to be exactly False, treating any other value as an overclaim, and confirms the required import classifications. Then contract_summary checks the scope limit. Every flag in FALSEAUTHORITYFLAGS (thirteen of them, from livegitmutation_authorized through publicationauthorized and releaseauthorized) must be exactly False in the contract. Any flag that is not False is an overclaim. The same stage confirms the contract declares the required import classifications (REQUIRED_CLASSIFICATIONS).",
"level": 4,
"order": 11,
"section_id": "contract-summary-checks-the-authority-ceiling",
"source_line_end": 121,
"source_line_start": 111,
"title": "`_contract_summary` checks the scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "scanpaths runs the shared secret scanner over all seven inputs, turning any blocking hit into a finding. Then scanpaths runs the shared secret scanner over all seven inputs. Any blocking hit is a finding.",
"level": 4,
"order": 12,
"section_id": "scan-paths-runs-the-secret-scanner",
"source_line_end": 127,
"source_line_start": 122,
"title": "`scan_paths` runs the secret scanner"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "**Finally the record is assembled, pass only when there are no findings, carrying digests, anchor and authority rows, and bodyinreceipt: false so no body is copied out.** Finally the record is assembled. status is pass when there are no findings and blocked otherwise. The record carries refs, digests, line counts, anchor rows, authority rows, material-class counts, and findings, and it sets bodyinreceipt: false so no source body is copied into the output. The record is written atomically as exportedworklandingcontrolbundlevalidationresult.json and also printed to stdout.",
"level": 4,
"order": 13,
"section_id": "assembling-the-body-free-result-record",
"source_line_end": 157,
"source_line_start": 128,
"title": "Assembling the metadata-only result record"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 159,
"source_line_start": 158,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The focused tests drive five rejections by mutating one field at a time, each producing a blocked status with a stable finding code, and the same shape covers other drifts.** The focused tests drive five rejections from the accepted bundle by mutating one field at a time. A digest changed in the manifest gives status blocked with SOURCEDIGESTMISMATCH. A source body relabelled to a private material class gives MATERIALCLASSNOT_ALLOWED. A body reclassified away from copiednonsecretmacrobody gives SOURCEIMPORTCLASS_MISMATCH. Dropping secret_exclusion from the contract's required classifications gives CLASSIFICATIONFLOORMISSING. Setting livegitmutation_authorized to True in the contract gives\u2026",
"level": 4,
"order": 15,
"section_id": "five-rejections-from-one-field-mutations",
"source_line_end": 175,
"source_line_start": 160,
"title": "Five rejections from one-field mutations"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 177,
"source_line_start": 176,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**The mechanics come from Git's index separating selected changes and from durable-workflow action ordering, imported here as an auditable public check with no power to act.** The mechanics copied here come from ordinary version-control and workflow engineering. Git's index already separates the changes you have selected from the rest of a dirty worktree, which is the practical ancestor of scoped path ownership. Durable-workflow systems show the value of a recorded action order and deterministic replay for long-running work. This component imports those ideas as a public check: digest parity, anchor presence, a no-live-mutation contract, and a secret scan make the copied control plane audit\u2026",
"level": 4,
"order": 17,
"section_id": "git-index-and-durable-workflow-ancestry",
"source_line_end": 195,
"source_line_start": 178,
"title": "Git index and durable-workflow ancestry"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from microcosm-substrate/:",
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 206,
"source_line_start": 196,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "**A pass writes the validation result record and means the copied bodies matched the manifest, every anchor was present, all authority flags stayed false, and the secret scan found nothing.** A pass writes result records/firstwave/worklandingcontrolspine/exportedworklandingcontrolbundlevalidationresult.json and means the copied bodies matched the manifest, every required anchor was present, all authority flags stayed false, and the secret scan found nothing. The focused regression suite that pins the pass case and the four rejections above is:",
"level": 4,
"order": 19,
"section_id": "what-a-pass-writes-and-means",
"source_line_end": 221,
"source_line_start": 207,
"title": "What a pass writes and means"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 223,
"source_line_start": 222,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**For a given bundle, the five copied bodies match their digests and line counts, the anchor logic is present, the contract declares no live authority, and no secret is present.** The strongest claim the evidence supports is narrow: for a given bundle, the five copied internal control bodies match their recorded digests and line counts, the named anchor logic is still present, the runtime contract declares no live authority, and no secret material is present. That is the whole proof boundary.",
"level": 4,
"order": 21,
"section_id": "the-strongest-supported-claim",
"source_line_end": 232,
"source_line_start": 224,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The check runs no live Git, mutates no ledger, releases no claim, and calls no provider, so a passing record proves only a faithful copy, never that a future landing is safe.** The scope limit stops there. This check does not run live Git, does not mutate any ledger, does not launch a claim, does not stage broadly, does not execute a private-index commit, does not call a provider, does not export account secrets, and does not publish or authorize a launch. A passing record is evidence that a copy is faithful and honest about its limits. It is not evidence that any future landing operation is safe, and it grants no power to perform one.",
"level": 4,
"order": 22,
"section_id": "what-the-authority-ceiling-stops",
"source_line_end": 243,
"source_line_start": 233,
"title": "What the scope limit stops"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/work_landing_control_spine.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public work-landing control spine validates copied work-landing internal control source bodies without authorizing live Git, ledger, claim-launch, private-index, public sharing, or launch operations."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "work_landing_control_spine.py",
"path": "src/microcosm_core/macro_tools/work_landing_control_spine.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for validating the copied work-landing control source-module bundle, manifest digests, required anchors, no-live-mutation contract, secret-exclusion scan, and scope limit.",
"symbols": [
"validate_work_landing_control_bundle",
"main",
"REQUIRED_SOURCE_REFS",
"FALSE_AUTHORITY_FLAGS",
"ANTI_CLAIM"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/macro_tools/work_landing_control_spine.py"
},
{
"label": "macro_projection_import_protocol.py",
"path": "src/microcosm_core/organs/macro_projection_import_protocol.py",
"resolution": "resolved",
"role": "Accepted import-protocol locus that classifies the work-landing control source-module import as a consumed verified public runtime import.",
"symbols": [
"refresh_exact_copy_source_modules",
"validate_projection_protocol"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/macro_projection_import_protocol.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/work_landing_control_spine.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/work_landing_control_spine.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Work Landing Control Spine is the public source-open import for work-landing internal control bodies. It validates copied work_landing, work_landing_status, mission transaction preflight, landing preflight, and scoped-commit source modules by manifest digest, required anchors, no-live-mutation contract flags, secret-exclusion scan, and metadata-only result record policy while keeping live Git, work log, work log, claim launch, shared-index mutation, private-index execution, external model access, public sharing, and launch-scope decision out of scope.",
"summary_line": "The public work-landing control spine validates copied work-landing internal control source bodies without authorizing live Git, ledger, claim-launch, private-index, public sharing, or launch operations.",
"title": "Work Landing Control Spine"
},
{
"anchor": "paper-module-executable-doctrine-grammar",
"authority_ceiling": "Public fixture, exported standards bundle, and copied source specimen/body digest evidence only; no private standards engine export, source doctrine completeness, source notes export, model-output data export, source-file changes, launch-scope decision, publishing-scope decision, non-public data equivalence, or whole-system correctness.",
"body_html": "
Component turns doctrine
\n
This component turns doctrine shape into something a program can pass or fail, checking each standard and paper-module fixture for a stated purpose, governing rule, result records, and honest scope statement. Executable doctrine grammar is a component that turns doctrine shape into something a program can pass or fail. It reads a folder of example standards and paper-module write-ups and checks each one for the parts doctrine here requires: a stated purpose, a governing rule, the result records it should produce, and an honest statement of what it does not claim. It writes a per-file report saying which entries are well formed and which fail a required check.
\n
Report is set
\n
The report is a set of metadata-only result records carrying references, hashes, counts, and verdicts, and a pass means the fixtures carry the required structure and every expected failure was observed. The report is a set of metadata-only result records. They carry references, hashes, counts, and verdicts. They never inline the document bodies being checked. A pass means the fixtures carry the required structure and that the checker observed every failure it is supposed to catch.
\n
Checks are structural
\n
The checks are structural, not semantic: they judge whether load-bearing fields are present, so a row cannot quietly drop its result-record expectations or scope statement and still pass. They judge whether the load-bearing fields are present, so a row cannot quietly drop its result-record expectations or its scope statement and still pass.
\n
Purpose
\n
Doctrine as prose convention
\n
In most systems doctrine is prose convention that nothing checks, so a loose fixture reads the part then drops a required field with no signal; this component asks per row whether the structure is real. Doctrine in most systems is prose convention. A standard says a rule should hold, a paper module says a section should be present, and nothing checks whether the claim is true. A loose fixture passes review because it reads the part, then drops a required field later with no signal. This component removes that gap. It answers one question per row: does this standard or paper-module fixture carry the structure doctrine here requires, or is it text that looks the part?
\n
Failure is first class
\n
The less obvious half is that failure is first class: five negative cases are part of the contract, so a run that never demonstrates a rejection is treated as blocked. The less obvious half is that failure is first class. Five negative cases are part of the contract. A run that accepts every input but never demonstrates a rejection is treated as blocked, so the checker has to prove it can reject, not only that it can accept.
\n
How it works
\n
validate reads and merges two checks
\n
validate reads standards_registry.json and every paper_modules/*.md file, scans them for forbidden content classes, then runs two structural checks and merges their results. The entry point is validate. It reads standards_registry.json and every paper_modules/*.md file from the input directory, scans them for forbidden content classes, then runs two structural checks and merges their results.
\n
Standard-registry field and overclaim checks
\n
validate_standard_registry records missing-field, duplicate-slug, prose-authority, and doctrine-completeness overclaim findings per row, and accepts a row only when it has no findings and no expected-failure marker.validate_standard_registry walks each standard row. It records MISSING_TELEOLOGY, MISSING_RECEIPT_EXPECTATIONS, MISSING_GOVERNING_STANDARD, and MISSING_ANTI_CLAIM when those fields are absent. It records DUPLICATE_STANDARD_SLUG_CONFLICT when two rows share a slug. It records PROSE_STANDARD_NOT_EXECUTABLE_AUTHORITY when a row marked standard_kind: prose also claims executable authority, and GRAMMAR_PASS_OVERCLAIMS_DOCTRINE_COMPLETE when a row claims doctrine completeness. A row with no findings and no expected-failure marker is accepted.
\n
Paper-module heading and sentinel checks
\n
validate_paper_modules checks each fixture module for the Teleology, Governing Standard, Receipt Expectations, and Scope boundary headings and rejects any that carries the copied-body sentinel or a completeness marker.validate_paper_modules calls validate_paper_module_shape on each fixture module. That function checks for the Teleology, Governing Standard, Receipt Expectations, and Scope boundary headings, and rejects any fixture that contains the copied-body sentinel text or a doctrine-completeness marker.
\n
Merging findings and required negatives
\n
validate passes only when no required negative case is missing and the content scan is clean, writing four result records under result records/sign-off/first_wave/.validate then computes which of the five required negative cases were observed and which are missing. The run status is a pass only when no required case is missing and the content scan is clean. It writes four result records: the standards report, the standards group index, the paper-module report, and the fixture result record under result records/sign-off/first_wave/.
\n
Two bundle entry points reuse the checks
\n
validate_standards_bundle and validate_executable_grammar_metabolism_bundle reuse the same checks over the exported public standards bundle and the exported specimen, each passing only on a clean scan with valid rows. Two more entry points reuse the same checks over exported bundles. validate_standards_bundle validates the exported public standards bundle and passes only when the scan is clean, there are no findings, and at least one standard row and one paper module are valid. validate_executable_grammar_metabolism_bundle validates the exported specimen: it reads bundle_manifest.json, README.md, grammar_board.json, and receipt.json, runs _metabolism_bundle_findings over their required keys and boundary phrases, then calls the copied-body gate.
\n
The copied-body digest gate
\n
validate_source_module_imports requires each of the 12 declared bodies to be a copied, allowlisted, non-result record source body whose SHA-256 recomputes to match the manifest, treating any mismatch as a finding.validate_source_module_imports is that gate. It reads source_module_manifest.json and requires source_import_class to be copied_non_secret_macro_body, body_in_receipt to be false, and the declared module_count to equal the number of rows. For each of the 12 declared bodies it checks the material class against public_macro_standard_body, public_macro_tool_body, and public_macro_receipt_body, requires body_copied true and body_in_receipt false, requires the source reference to sit on the allowlist, confirms the copied target file exists, and recomputes its SHA-256 digest with _sha256_file to match the manifest. A mismatch is a finding, not a silent pass.
\n
result_card projects a metadata-only card
\n
result_card projects any result into a compact card of status, counts, and verdicts, dropping the full finding lists and any body-bearing fields named in CARD_OMITTED_FULL_PAYLOAD_KEYS.result_card projects any of these results into a compact card for a reader or the Atlas. It keeps status, counts, and verdicts and drops the full finding lists and any body-bearing fields named in CARD_OMITTED_FULL_PAYLOAD_KEYS.
\n
Function
Role
validate
Runs the fixture check and writes the four result records
validate_standard_registry
Field, duplicate-slug, prose-authority, and overclaim checks per standard row
validate_paper_module_shape
Required-heading and sentinel checks per fixture module
validate_standards_bundle
Same checks over the exported public standards bundle
validate_executable_grammar_metabolism_bundle
Specimen bundle keys, boundary phrases, and copied-body checks
validate_source_module_imports
Manifest, allowlist, and SHA-256 digest gate for the 12 copied bodies
result_card
metadata-only card projection for readers and the Atlas
\n
To run the fixture check from microcosm-substrate/:
\nDiagram of the mechanism (6 steps).Diagram source & refs
Source refs
Public fixtures
standards_registry.json + paper_modules/*.md
field, slug, authority checks
validate_standard_registry
heading and sentinel checks
validate_paper_module_shape
metadata-only result records
receipts/first_wave/executable_doctrine_grammar/
flowchart TD A["Public fixtures standards_registry.json + paper_modules/*.md"] --> B["validate"] B --> C["validate_standard_registry field, slug, authority checks"] B --> D["validate_paper_module_shape heading and sentinel checks"] C --> E["Merge findings observed vs required negative cases"] D --> E E --> F["metadata-only result records result records/first_wave/executable_doctrine_grammar/"]
\n
Negative cases
\n
Fixture keeps five
\n
The fixture keeps five failures executable rather than described in prose. The fixture keeps five failures executable rather than described in prose. Each maps to an error code the checker must observe:
\n
invalid_standard_and_module: a row missing teleology, result-record expectations, governing standard, and scope statement, raising MISSING_TELEOLOGY, MISSING_RECEIPT_EXPECTATIONS, MISSING_GOVERNING_STANDARD, and MISSING_ANTI_CLAIM.
prose_standard_claims_runtime_authority: a prose-only standard that claims executable authority, raising PROSE_STANDARD_NOT_EXECUTABLE_AUTHORITY.
macro_doctrine_body_copied_into_fixture: a copied private body pasted into a public fixture, raising MACRO_DOCTRINE_BODY_IN_PUBLIC_FIXTURE.
duplicate_standard_slug_conflict: two rows sharing a slug, raising DUPLICATE_STANDARD_SLUG_CONFLICT.
grammar_index_pass_overclaims_doctrine_complete: a passing fixture that claims doctrine completeness, raising GRAMMAR_PASS_OVERCLAIMS_DOCTRINE_COMPLETE.
\n
Required cases enforce provable rejection
\n
The required cases live in EXPECTED_NEGATIVE_CASES, and if a run does not observe every one validate returns a blocked status. The set of required cases lives in EXPECTED_NEGATIVE_CASES. If a run does not observe every one, validate returns a blocked status.
\n
Prior Art Grounding
\n
Schema validation and executable semantics
\n
The design follows schema validation, generated grammars, and executable-semantics work like JSON Schema, Tree-sitter, and the K framework, checking doctrine shape by program instead of leaving it as convention. The design follows schema validation, generated grammars, and executable-semantics work. JSON Schema is the reference for validating document shape against a shared machine contract. Tree-sitter shows the value of generated grammars for inspectable structure. The K framework is a close reference for turning semantic rules into executable artifacts. The borrowed pattern is narrow: doctrine shape, result-record expectations, duplicate slugs, copied bodies, and scope statements are checked by a program instead of left as convention.
A green run here means only that the focused tests and the corpus parity check both accept these fixtures and this projection. A pass means the focused test suite and the corpus check both accept the current fixtures and this projection.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
Given the public fixtures and exported bundles, it checks doctrine shape, rejects the five named failure classes, and verifies the 12 copied bodies match their declared SHA-256 digests, all metadata-only. The strongest claim this component supports is bounded. Given the public fixtures and the exported bundles, it checks doctrine shape, rejects the five named failure classes, and verifies that the 12 copied bodies match their declared SHA-256 digests, all through metadata-only result records. The proof boundary is the public standards and paper-module fixtures, the exported bundles, the copied-body manifest, and the result records those runs produce.
\n
What a grammar pass is not
\n
A grammar pass is not doctrine completeness, publishing-scope decision, or launch-scope decision, and the component judges no prose quality, calls no provider, exports no private bodies, and mutates no source. It does not claim more. A grammar pass is not doctrine completeness, publishing-scope decision, or launch-scope decision. The component does not judge prose quality, prove whole-system correctness, call any provider, export private bodies, or change source files. That scope limit is fixed. A stronger claim has to land in the source record and the generated projection before this page can narrate it.
\n",
"components": [
{
"anchor": "component-executable_doctrine_grammar",
"component_id": "executable_doctrine_grammar",
"public_label": "Executable Doctrine Grammar"
}
],
"doctrine_feed": {
"counts": {
"axiom": 3,
"concept": 2,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 3
},
"link_count": 11,
"links": [
{
"id": "executable_doctrine_grammar",
"kind": "organ",
"kind_label": "component",
"label": "Executable Doctrine Grammar",
"relation": "explains",
"url": "component-executable-doctrine-grammar.html"
},
{
"id": "mechanism.executable_doctrine_grammar.validates_public_doctrine_grammar_bundle",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public doctrine grammar bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-executable-doctrine-grammar-validates-public-doctrine-grammar-bundle"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
},
{
"id": "concept.executable_doctrine_grammar_standard_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "standard as schema contract with explicit scope limit",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-executable-doctrine-grammar-standard-bundle"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.doctrine_fact_claim_audit",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Doctrine Fact Claim Audit",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[48:paper_module.doctrine_fact_claim_audit].depends_on",
"url": "paper-module-doctrine-fact-claim-audit.html"
},
{
"id": "paper_module.microcosm_axiom_substrate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Microcosm Axiom System",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[92:paper_module.microcosm_axiom_substrate].depends_on",
"url": "paper-module-microcosm-axiom-substrate.html"
}
],
"paper_module_id": "paper_module.executable_doctrine_grammar",
"reader_packet": {
"abstract": "The public executable-doctrine grammar fixture validates public standard rows, paper-module sections, negative cases, copied executable-grammar and standards/type-plane source-module bodies, and metadata-only result records without claiming doctrine completeness.",
"authority_ceiling": "Public fixture, exported standards bundle, and copied source specimen/body digest evidence only; no private standards engine export, source doctrine completeness, source notes export, model-output data export, source-file changes, launch-scope decision, publishing-scope decision, non-public data equivalence, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.executable_doctrine_grammar.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-executable-doctrine-grammar-diagram-1",
"html_url": "docs/paper-module-executable-doctrine-grammar.html#paper-module-executable-doctrine-grammar-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Public fixtures standards_registry.json + paper_modules/*.md\"] --> B[\"validate\"] B --> C[\"validate_standard_registry field, slug, authority checks\"] B --> D[\"validate_paper_module_shape heading and sentinel checks\"] C --> E[\"Merge findings observed vs required negative cases\"] D --> E E --> F[\"metadata-only result records result records/first_wave/executable_doctrine_grammar/\"]",
"source_format": "mermaid",
"source_hash": "sha256:bb7424e89e192da58ad1e51703462febda2a783b9278c3bec2f066bb5a6a47c6",
"source_line_end": 161,
"source_line_start": 153
}
],
"grouped_relation_feeds": [
{
"count": 3,
"links": [
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "executable_doctrine_grammar",
"kind": "organ",
"kind_label": "component",
"label": "Executable Doctrine Grammar",
"relation": "explains",
"url": "component-executable-doctrine-grammar.html"
},
{
"id": "mechanism.executable_doctrine_grammar.validates_public_doctrine_grammar_bundle",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public doctrine grammar bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-executable-doctrine-grammar-validates-public-doctrine-grammar-bundle"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.executable_doctrine_grammar_standard_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "standard as schema contract with explicit scope limit",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-executable-doctrine-grammar-standard-bundle"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.executable_doctrine_grammar",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 245,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/executable_doctrine_grammar.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "JSON Schema",
"section_id": "schema-validation-and-executable-semantics",
"url": "https://json-schema.org/specification"
},
{
"kind": "external_reference",
"label": "Tree-sitter",
"section_id": "schema-validation-and-executable-semantics",
"url": "https://github.com/tree-sitter/tree-sitter"
},
{
"kind": "external_reference",
"label": "K framework",
"section_id": "schema-validation-and-executable-semantics",
"url": "https://kframework.org/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component turns doctrine shape into something a program can pass or fail, checking each standard and paper-module fixture for a stated purpose, governing rule, result records, and honest scope statement.** Executable doctrine grammar is a component that turns doctrine shape into something a program can pass or fail. It reads a folder of example standards and paper-module write-ups and checks each one for the parts doctrine here requires: a stated purpose, a governing rule, the result records it should produce, and an honest statement of what it does not claim. It writes a per-file report saying which entries are well formed and which fail a required check.",
"level": 4,
"order": 1,
"section_id": "component-turns-doctrine",
"source_line_end": 14,
"source_line_start": 3,
"title": "Component turns doctrine"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The report is a set of metadata-only result records carrying references, hashes, counts, and verdicts, and a pass means the fixtures carry the required structure and every expected failure was observed.** The report is a set of metadata-only result records. They carry references, hashes, counts, and verdicts. They never inline the document bodies being checked. A pass means the fixtures carry the required structure and that the checker observed every failure it is supposed to catch.",
"level": 4,
"order": 2,
"section_id": "report-is-set",
"source_line_end": 23,
"source_line_start": 15,
"title": "Report is set"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The checks are structural, not semantic: they judge whether load-bearing fields are present, so a row cannot quietly drop its result-record expectations or scope statement and still pass.** They judge whether the load-bearing fields are present, so a row cannot quietly drop its result-record expectations or its scope statement and still pass.",
"level": 4,
"order": 3,
"section_id": "checks-are-structural",
"source_line_end": 31,
"source_line_start": 24,
"title": "Checks are structural"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 33,
"source_line_start": 32,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**In most systems doctrine is prose convention that nothing checks, so a loose fixture reads the part then drops a required field with no signal; this component asks per row whether the structure is real.** Doctrine in most systems is prose convention. A standard says a rule should hold, a paper module says a section should be present, and nothing checks whether the claim is true. A loose fixture passes review because it reads the part, then drops a required field later with no signal. This component removes that gap. It answers one question per row: does this standard or paper-module fixture carry the structure doctrine here requires, or is it text that looks the part?",
"level": 4,
"order": 5,
"section_id": "doctrine-as-prose-convention",
"source_line_end": 45,
"source_line_start": 34,
"title": "Doctrine as prose convention"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The less obvious half is that failure is first class: five negative cases are part of the contract, so a run that never demonstrates a rejection is treated as blocked.** The less obvious half is that failure is first class. Five negative cases are part of the contract. A run that accepts every input but never demonstrates a rejection is treated as blocked, so the checker has to prove it can reject, not only that it can accept.",
"level": 4,
"order": 6,
"section_id": "failure-is-first-class",
"source_line_end": 54,
"source_line_start": 46,
"title": "Failure is first class"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 56,
"source_line_start": 55,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validate reads standardsregistry.json and every papermodules/.md file, scans them for forbidden content classes, then runs two structural checks and merges their results.** The entry point is validate. It reads standardsregistry.json and every papermodules/*.md file from the input directory, scans them for forbidden content classes, then runs two structural checks and merges their results.",
"level": 4,
"order": 8,
"section_id": "validate-reads-and-merges-two-checks",
"source_line_end": 65,
"source_line_start": 57,
"title": "`validate` reads and merges two checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatestandard_registry records missing-field, duplicate-slug, prose-authority, and doctrine-completeness overclaim findings per row, and accepts a row only when it has no findings and no expected-failure marker.** validatestandardregistry walks each standard row. It records MISSINGTELEOLOGY, MISSINGRECEIPTEXPECTATIONS, MISSINGGOVERNING_STANDARD, and MISSINGANTICLAIM when those fields are absent. It records DUPLICATESTANDARDSLUG_CONFLICT when two rows share a slug. It records PROSESTANDARDNOTEXECUTABLEAUTHORITY when a row marked standard_kind: prose also claims executable authority, and GRAMMARPASSOVERCLAIMSDOCTRINECOMPLETE when a row claims doctrine completeness. A row with no findings a\u2026",
"level": 4,
"order": 9,
"section_id": "standard-registry-field-and-overclaim-checks",
"source_line_end": 79,
"source_line_start": 66,
"title": "Standard-registry field and overclaim checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatepaper_modules checks each fixture module for the Teleology, Governing Standard, Result record Expectations, and Scope boundary headings and rejects any that carries the copied-body sentinel or a completeness marker.** validatepapermodules calls validatepapermodule_shape on each fixture module. That function checks for the Teleology, Governing Standard, `Result record Expectations, and Scope boundary` headings, and rejects any fixture that contains the copied-body sentinel text or a doctrine-completeness marker.",
"level": 4,
"order": 10,
"section_id": "paper-module-heading-and-sentinel-checks",
"source_line_end": 89,
"source_line_start": 80,
"title": "Paper-module heading and sentinel checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**validate passes only when no required negative case is missing and the content scan is clean, writing four result records under result records/sign-off/firstwave/. validate then computes which of the five required negative cases were observed and which are missing. The run status is a pass only when no required case is missing and the content scan is clean. It writes four result records: the standards report, the standards group index, the paper-module report, and the fixture result record under result records/sign-off/first_wave/.",
"level": 4,
"order": 11,
"section_id": "merging-findings-and-required-negatives",
"source_line_end": 99,
"source_line_start": 90,
"title": "Merging findings and required negatives"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatestandardsbundle and validateexecutablegrammarmetabolism_bundle reuse the same checks over the exported public standards bundle and the exported specimen, each passing only on a clean scan with valid rows.** Two more entry points reuse the same checks over exported bundles. validatestandardsbundle validates the exported public standards bundle and passes only when the scan is clean, there are no findings, and at least one standard row and one paper module are valid. validateexecutablegrammarmetabolismbundle validates the exported specimen: it reads bundlemanifest.json, README.md, grammarboard.json, and result record.json, runs metabolismbundle_findings over their required keys and bo\u2026",
"level": 4,
"order": 12,
"section_id": "two-bundle-entry-points-reuse-the-checks",
"source_line_end": 112,
"source_line_start": 100,
"title": "Two bundle entry points reuse the checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatesourcemoduleimports requires each of the 12 declared bodies to be a copied, allowlisted, non-result record source body whose SHA-256 recomputes to match the manifest, treating any mismatch as a finding. validatesourcemoduleimports is that gate. It reads sourcemodulemanifest.json and requires sourceimportclass to be copiednonsecretmacrobody, bodyinreceipt to be false, and the declared module_count to equal the number of rows. For each of the 12 declared bodies it checks the material class against publicmacrostandard_body, publicmacrotoolbody, and publicmacroreceiptbody, requires body_copied true and bodyinreceipt false, requires the source reference to sit on the allowlist, confirms\u2026",
"level": 4,
"order": 13,
"section_id": "the-copied-body-digest-gate",
"source_line_end": 127,
"source_line_start": 113,
"title": "The copied-body digest gate"
},
{
"block_kinds": [
"code",
"diagram",
"paragraph",
"table"
],
"excerpt": "resultcard projects any result into a compact card of status, counts, and verdicts, dropping the full finding lists and any body-bearing fields named in CARDOMITTEDFULLPAYLOADKEYS. resultcard projects any of these results into a compact card for a reader or the Atlas. It keeps status, counts, and verdicts and drops the full finding lists and any body-bearing fields named in CARDOMITTEDFULLPAYLOADKEYS. To run the fixture check from microcosm-substrate/:",
"level": 4,
"order": 14,
"section_id": "result-card-projects-a-body-free-card",
"source_line_end": 162,
"source_line_start": 128,
"title": "`result_card` projects a metadata-only card"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "negative-cases",
"source_line_end": 164,
"source_line_start": 163,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The fixture keeps five failures executable rather than described in prose. The fixture keeps five failures executable rather than described in prose. Each maps to an error code the checker must observe: invalidstandardandmodule: a row missing teleology, result-record expectations, governing standard, and scope statement, raising MISSINGTELEOLOGY, MISSINGRECEIPTEXPECTATIONS, MISSINGGOVERNINGSTANDARD, and MISSINGANTICLAIM. prosestandardclaimsruntimeauthority: a prose-only standard that claims executable authority, raising PROSESTANDARDNOTEXECUTABLEAUTHORITY. macrodoctrinebodycopiedintofixture: a copied private body pasted into a public fixture, raising MACRODOCTRINEBODYINPUBLICFIXTURE. duplic\u2026",
"level": 4,
"order": 16,
"section_id": "fixture-keeps-five",
"source_line_end": 176,
"source_line_start": 165,
"title": "Fixture keeps five"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The required cases live in EXPECTEDNEGATIVE_CASES, and if a run does not observe every one validate returns a blocked status.** The set of required cases lives in EXPECTEDNEGATIVECASES. If a run does not observe every one, validate returns a blocked status.",
"level": 4,
"order": 17,
"section_id": "required-cases-enforce-provable-rejection",
"source_line_end": 183,
"source_line_start": 177,
"title": "Required cases enforce provable rejection"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "prior-art-grounding",
"source_line_end": 185,
"source_line_start": 184,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The design follows schema validation, generated grammars, and executable-semantics work like JSON Schema, Tree-sitter, and the K framework, checking doctrine shape by program instead of leaving it as convention.** The design follows schema validation, generated grammars, and executable-semantics work. JSON Schema is the reference for validating document shape against a shared machine contract. Tree-sitter shows the value of generated grammars for inspectable structure. The K framework is a close reference for turning semantic rules into executable artifacts. The borrowed pattern is narrow: doctrine shape, result-record expectations, duplicate slugs, copied bodies, and scope statements are\u2026",
"level": 4,
"order": 19,
"section_id": "schema-validation-and-executable-semantics",
"source_line_end": 200,
"source_line_start": 186,
"title": "Schema validation and executable semantics"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "validation-receipt-path",
"source_line_end": 202,
"source_line_start": 201,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "the repo root without mutating durable result-record or generated projection surfaces:",
"level": 4,
"order": 21,
"section_id": "validate-reader-projection",
"source_line_end": 213,
"source_line_start": 203,
"title": "Validate reader projection"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A green run here means only that the focused tests and the corpus parity check both accept these fixtures and this projection.** A pass means the focused test suite and the corpus check both accept the current fixtures and this projection.",
"level": 4,
"order": 22,
"section_id": "what-a-pass-means-here",
"source_line_end": 219,
"source_line_start": 214,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 23,
"section_id": "claim-ceiling",
"source_line_end": 221,
"source_line_start": 220,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Given the public fixtures and exported bundles, it checks doctrine shape, rejects the five named failure classes, and verifies the 12 copied bodies match their declared SHA-256 digests, all metadata-only.** The strongest claim this component supports is bounded. Given the public fixtures and the exported bundles, it checks doctrine shape, rejects the five named failure classes, and verifies that the 12 copied bodies match their declared SHA-256 digests, all through metadata-only result records. The proof boundary is the public standards and paper-module fixtures, the exported bundles, the copied-body manifest, and the result records those runs produce.",
"level": 4,
"order": 24,
"section_id": "the-strongest-supported-claim",
"source_line_end": 233,
"source_line_start": 222,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A grammar pass is not doctrine completeness, public sharing decision, or launch readiness, and the component judges no prose quality, calls no provider, exports no private bodies, and mutates no source.** It does not claim more. A grammar pass is not doctrine completeness, public sharing decision, or launch-scope decision. The component does not judge prose quality, prove whole-system correctness, call any provider, export private bodies, or change source files. That scope limit is fixed. A stronger claim has to land in the source record and the generated projection before this page can narrate it.",
"level": 4,
"order": 25,
"section_id": "what-a-grammar-pass-is-not",
"source_line_end": 244,
"source_line_start": 234,
"title": "What a grammar pass is not"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/executable_doctrine_grammar.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public executable-doctrine grammar fixture validates public standard rows, paper-module sections, negative cases, copied executable-grammar and standards/type-plane source-module bodies, and metadata-only result records without claiming doctrine completeness."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "executable_doctrine_grammar.py",
"path": "src/microcosm_core/organs/executable_doctrine_grammar.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public doctrine grammar fixture validation, exported bundle validation, copied source-module digest checks, negative-case enforcement, cards, and scope limits.",
"symbols": [
"validate",
"validate_standards_bundle",
"validate_executable_grammar_metabolism_bundle",
"validate_source_module_imports",
"result_card",
"EXPECTED_NEGATIVE_CASES",
"GRAMMAR_AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/executable_doctrine_grammar.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/executable_doctrine_grammar.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/executable_doctrine_grammar.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Executable Doctrine Grammar is the public grammar membrane for doctrine-shaped runtime fixtures. It validates standard row fields, paper-module teleology and result record sections, duplicate-slug and overclaim negative cases, exact copied executable-grammar, standards-registry/type-plane, lattice, kind-atlas, and standards option-surface bodies, then emits metadata-only result records with source refs, digests, counts, and scope limits.",
"summary_line": "The public executable-doctrine grammar fixture validates public standard rows, paper-module sections, negative cases, copied executable-grammar and standards/type-plane source-module bodies, and metadata-only result records without claiming doctrine completeness.",
"title": "Executable Doctrine Grammar"
},
{
"anchor": "paper-module-macro-projection-import-protocol",
"authority_ceiling": "Verified source body imports, fixture result records, exported projection-bundle result records, per-slice manifest refs, metadata-only formal-math metadata and validation refs, and metadata-only result record fields only; no theorem/formal-result correctness authority, private target status, challenge-result public sharing, live source authority, whole-system equivalence, launch, public sharing, hosted deployment, recipient work, provider/oracle payload authority, provider or Lean/Lake execution, secret/non-public body export, source-file changes, or whole-system correctness.",
"body_html": "
The checkpoint for copying source material
\n
macro_projection_import_protocol is the checkpoint for copying material out of a much larger source project into the public Microcosm folder.macro_projection_import_protocol is the checkpoint for copying material out of a much larger source project into the public Microcosm folder. It reads a proposed set of copied files, classifies each one, and decides whether the copy is allowed out and honestly tied to the source it claims to come from.
\n
It exists so the public folder can stay dense and real without becoming a place where a secret, an operator conversation, a model-output data, or launch material leaks out alongside genuinely useful code. It exists so the public folder can stay dense and real without becoming a place where a secret, an operator conversation, a model-output data, or launch material leaks out alongside genuinely useful code. Every run answers one question per copied file: was this body allowed out, and does the public copy match the source it names.
\n
The output is a metadata-only result record. It lists what was imported, what was deliberately left out with a note saying so, and what was refused. The output is a metadata-only result record. It lists what was imported, what was deliberately left out with a note saying so, and what was refused. The record carries paths, hashes, counts, and validation refs, never the private bodies themselves.
\n
Purpose
\n
Microcosm grows by copying real files out of a private codebase. A loose copy is the exact artifact that quietly carries a account secret or an operator thread along with the code you wanted. Microcosm grows by copying real files out of a private codebase. A loose copy is the exact artifact that quietly carries a account secret or an operator thread along with the code you wanted. A trust statement ("we checked, it is fine") does not scale and cannot be re-run by a reader.
\n
So the check is an accounting check, not a trust statement. Each copied row declares its source ref, its public target ref, a content digest, and a material class. So the check is an accounting check, not a trust statement. Each copied row declares its source ref, its public target ref, a content digest, and a material class. The protocol sorts the class into one of two sets. PUBLIC_SAFE_BODY_MATERIAL_CLASSES holds five classes that may be copied with provenance: pattern, standard, tool, result-record, and proof bodies. TRUE_FORBIDDEN_MATERIAL_CLASSES holds fifteen classes that can never appear as an imported body, including raw operator source, operator thread, model-output data, account secret, secret, recipient material, launch material, and the six private formal-math body classes. Anything claiming to be must also carry a verification record naming the digest, the source-to-target relation, and the mode used to check the copy.
\n
How it works
\n
main dispatches four commands. run validates a first-wave fixture directory (the positive slices plus the named negative cases) and writes result records through write_receipts.main dispatches four commands. run validates a first-wave fixture directory (the positive slices plus the named negative cases) and writes result records through write_receipts. run_projection_bundle validates an exported bundle and writes one exported_projection_import_bundle_validation_result.json. plan calls preview_import_plan, which scores each proposed cell and writes nothing. refresh-exact-copy-source-modules calls refresh_exact_copy_source_modules to re-copy exact-copy slices whose source has moved on. run and run_projection_bundle both assemble their answer through _build_result, which runs four validators over the fixture: validate_projection_protocol for the copied and omitted rows, validate_cleaning_policy for the cleaned rows, validate_standalone_runtime_severance for standalone-runtime dependency leaks, and validate_import_plan for the forward plan.
\n
The core is validate_projection_protocol. It reads the protocol payload and runs these checks in order: The core is validate_projection_protocol. It reads the protocol payload and runs these checks in order:
\n
Function
What it does
validate_projection_protocol
Sorts every copied row and emits the pass/blocked record
_public_safe_import_status
Classifies a row whose class is public-safe; returns nothing for other classes
_body_import_verification_findings
Checks the verification record: mode, verified flag, and digest binding
_forbidden_body_request
Detects a request to import a account secret or account-bound body
_authority_upgrade
Detects a claim to live private-source authority
_release_or_equivalence_overclaim
Detects a launch, public sharing, or private-equivalence claim
_formal_math_claim_boundary_subjects
Detects a formal-math theorem, proof, or solver-trace authority claim
\n
First it enforces a density floor: the protocol must cite at least two source refs, two public runtime refs, and two validation refs, or the row fails as MACRO_PROJECTION_PROTOCOL_DENSITY_MISSING. First it enforces a density floor: the protocol must cite at least two source refs, two public runtime refs, and two validation refs, or the row fails as MACRO_PROJECTION_PROTOCOL_DENSITY_MISSING. Then it walks each copied row. A row whose class is goes to _public_safe_import_status; anything it flags becomes a blocking finding. A row whose class is in TRUE_FORBIDDEN_MATERIAL_CLASSES, or any row that sets body_copied without a valid verification record, is rejected as MACRO_PROJECTION_FORBIDDEN_BODY_IMPORT. Omitted rows must carry an omission_receipt_ref. The result is PASS only when the floors are met, copied, omitted, cleaned, and step rows are all present, and no blocking finding remains.
\n
Digest requirements for a copied body
\n
For a body that is copied, _body_import_verification_findings does the real integrity work. For a body that is copied, _body_import_verification_findings does the real integrity work. The verification record must be marked verified and must use one of two modes in BODY_IMPORT_VERIFICATION_MODES: exact_source_digest_match or verified_light_edit_recipe. The target body digest must equal the declared target digest. For an exact copy it also re-hashes the named source, requires a relation from EXACT_COPY_SOURCE_TO_TARGET_RELATIONS, and checks the source digest.
\n
The load-bearing detail is how the protocol treats an exact copy whose source has since changed. A mismatch on an exact-copy row is not a failed import. The load-bearing detail is how the protocol treats an exact copy whose source has since changed. A mismatch on an exact-copy row is not a failed import. It is recorded as live source drift: the original copy was still honest, the source has simply moved on, and the row is flagged for the refresh path rather than failed. That separates a dishonest import from a stale one, so a routine upstream edit is not mistaken for a broken proof, and the public copy stays faithful without tracking every upstream change in lock-step.
\n
validate_import_plan guards the forward direction. Each proposed cell must name source and target refs, must name validation refs, must not use a private-system- shaped ref, and must only reference body material that the protocol already carries.validate_import_plan guards the forward direction. Each proposed cell must name source and target refs, must name validation refs, must not use a private-system- shaped ref, and must only reference body material that the protocol already carries. A plan needs at least three cells to pass. preview_import_plan wraps this as the macro_projection_import_intake_preview_v1 record: it scores each cell, projects the intake and severance boards, and lists blocking surfaces without writing anything.
\nDiagram of the mechanism (9 steps).Diagram source & refs
Source refs
Validate projection protocol
validate_projection_protocol
flowchart TD A["Copied row source ref, target ref, digest, material class"] --> B["validate_projection_protocol"] B --> C["Class in forbidden set Reject: forbidden body import"] B --> D["Class _public_safe_import_status"] D --> E["_body_import_verification_findings mode, verified flag, digest binding"] E --> F["Exact copy: re-hash live source"] F --> G["Match: body floor"] F --> H["Mismatch: flag live source drift, refresh later"] H --> G G --> I["metadata-only result record"]
\n
Negative cases
\n
The first-wave fixture drives these named rejections through validate_projection_protocol, validate_import_plan, and validate_standalone_runtime_severance: The first-wave fixture drives these named rejections through validate_projection_protocol, validate_import_plan, and validate_standalone_runtime_severance:
\n
forbidden_body_import_overclaim: a request to import a account secret or account-bound body.
missing_omission_receipt: omitted material with no omission result record ref.
authority_upgrade_overclaim: a claim that a public import grants live private-source authority.
release_or_private_equivalence_overclaim: a launch, public sharing, recipient, or private-equivalence claim.
formal_math_claim_boundary_overclaim: a formal-math theorem, proof, private target status, solver-trace, or provider-oracle authority claim.
missing_validation_ref: a proposed cell with no validation refs.
standalone_dependency_leak: a runtime dependency on the live source root that a standalone public checkout cannot resolve.
\n
Prior Art Grounding
\n
The check follows established provenance and software-supply-chain patterns: a copied or refactored artifact needs source refs, target refs, digests, validation refs, omission records, and a stated claim boundary. The check follows established provenance and software-supply-chain patterns: a copied or refactored artifact needs source refs, target refs, digests, validation refs, omission records, and a stated claim boundary. The closest public anchors are W3C PROV for entity, activity, and agent provenance, the SLSA specification for artifact integrity and provenance in software supply chains, and in-toto for linking supply-chain steps through signed metadata. This component applies those patterns to a public and private projection boundary rather than to launch attestation, and it stays a local validator over public fixtures and exported bundles.
\n
Validation Result record Path
\n
From microcosm-substrate/, reproduce this page's checks with temporary result records:
\n
What the fixture and bundle validate
\n
A pass means the fixture and bundle slices carried their required refs and digests, the named negative cases were rejected, and the result records stayed metadata-only. A pass means the fixture and bundle slices carried their required refs and digests, the named negative cases were rejected, and the result records stayed metadata-only. Re-enter this module when a new projection cell lands, a source-module manifest is refreshed, or a result record count changes.
\n
Scope boundary
\n
Scope limit
\n
The strongest import-accounting claim
\n
The strongest claim this component supports: for a named slice, it validated classification, source-to-target digest binding, per-slice manifests, omission records, the named negative cases, and a metadata-only result-record policy. The strongest claim this component supports: for a named slice, it validated classification, source-to-target digest binding, per-slice manifests, omission records, the named negative cases, and a metadata-only result-record policy. Accepted records expose the current public_safe_body_import_count, public_safe_body_import_status, projection_status_counts, open_actionable_cell_count, and secret_exclusion_scan fields.
\n
The import protocol's scope limit
\n
That is the proof boundary and the scope limit. A passing scan is a declared-input regression gate over the material classes the protocol recognizes, not a complete account secret audit and bounded evidence that no private material exists anywhere. That is the proof boundary and the scope limit. A passing scan is a declared-input regression gate over the material classes the protocol recognizes, not a complete account secret audit and bounded evidence that no private material exists anywhere. It does not show the public copy is complete, equal to the private root, or ready to launch. It does not grant authority over live private source, and it does not permit public sharing, hosted deployment, secret export, source-file changes, external model access, or Lean and Lake execution. The formal-math lane may carry statement-shape metadata, evidence-cell refs, and synthetic fixtures, but never theorem or formal-result correctness. Any stronger claim has to come from the owning result record, standard, or launch gate.
\n",
"components": [
{
"anchor": "component-macro_projection_import_protocol",
"component_id": "macro_projection_import_protocol",
"public_label": "Source Projection Import Protocol"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 7,
"principle": 3
},
"link_count": 14,
"links": [
{
"id": "macro_projection_import_protocol",
"kind": "organ",
"kind_label": "component",
"label": "Source Projection Import Protocol",
"relation": "explains",
"url": "component-macro-projection-import-protocol.html"
},
{
"id": "mechanism.macro_projection_import_protocol.validates_public_macro_projection_imports",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public source projection imports",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-macro-projection-import-protocol-validates-public-macro-projection-imports"
},
{
"id": "P-14",
"kind": "principle",
"kind_label": "principle",
"label": "Carry basis and provenance together",
"relation": "governed by",
"url": "doctrine.html#dcard-p-14"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
},
{
"id": "paper_module.formal_math_readiness_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Readiness Gate",
"relation": "depends on",
"url": "paper-module-formal-math-readiness-gate.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.batch11_saturation_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 11 Saturation Engines Bundle",
"relation": "depends on",
"url": "paper-module-batch11-saturation-engines-capsule.html"
},
{
"id": "paper_module.batch8_compliance_pipeline_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Compliance Pipeline Bundle",
"relation": "depends on",
"url": "paper-module-batch8-compliance-pipeline-capsule.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.agent_monitor_redteam_falsification_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Monitor Redteam Falsification Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[33:paper_module.agent_monitor_redteam_falsification_replay].depends_on",
"url": "paper-module-agent-monitor-redteam-falsification-replay.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[2:paper_module.agent_route_observability_runtime].depends_on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.batch10_governance_compilers_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 10 Governance And Compilers Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[72:paper_module.batch10_governance_compilers_capsule].depends_on",
"url": "paper-module-batch10-governance-compilers-capsule.html"
},
{
"id": "paper_module.batch10_live_source_drift_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 10 Live Source Drift Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[68:paper_module.batch10_live_source_drift_capsule].depends_on",
"url": "paper-module-batch10-live-source-drift-capsule.html"
},
{
"id": "paper_module.batch11_saturation_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 11 Saturation Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[73:paper_module.batch11_saturation_engines_capsule].depends_on",
"url": "paper-module-batch11-saturation-engines-capsule.html"
},
{
"id": "paper_module.batch4_proof_authority_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 4 Proof, Authority, and Runtime Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[74:paper_module.batch4_proof_authority_runtime].depends_on",
"url": "paper-module-batch4-proof-authority-runtime.html"
},
{
"id": "paper_module.batch5_authority_systems_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 5 Authority and Systems Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[77:paper_module.batch5_authority_systems_capsule].depends_on",
"url": "paper-module-batch5-authority-systems-capsule.html"
},
{
"id": "paper_module.batch6_unsurfaced_primitives_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 6 Unsurfaced Primitives Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[75:paper_module.batch6_unsurfaced_primitives_capsule].depends_on",
"url": "paper-module-batch6-unsurfaced-primitives-capsule.html"
},
{
"id": "paper_module.batch7_demo_take_console_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Demo Take Console Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[79:paper_module.batch7_demo_take_console_capsule].depends_on",
"url": "paper-module-batch7-demo-take-console-capsule.html"
},
{
"id": "paper_module.batch7_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Source Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[69:paper_module.batch7_macro_engines_capsule].depends_on",
"url": "paper-module-batch7-macro-engines-capsule.html"
},
{
"id": "paper_module.batch7_oracle_sibling_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Oracle Sibling Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[78:paper_module.batch7_oracle_sibling_capsule].depends_on",
"url": "paper-module-batch7-oracle-sibling-capsule.html"
},
{
"id": "paper_module.batch7_secondary_runtime_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Secondary Runtime Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[91:paper_module.batch7_secondary_runtime_capsule].depends_on",
"url": "paper-module-batch7-secondary-runtime-capsule.html"
},
{
"id": "paper_module.batch8_audio_level_rms_port",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Audio Level RMS Port",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[58:paper_module.batch8_audio_level_rms_port].depends_on",
"url": "paper-module-batch8-audio-level-rms-port.html"
},
{
"id": "paper_module.batch8_compliance_pipeline_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Compliance Pipeline Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[59:paper_module.batch8_compliance_pipeline_capsule].depends_on",
"url": "paper-module-batch8-compliance-pipeline-capsule.html"
},
{
"id": "paper_module.batch8_policy_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Policy Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[60:paper_module.batch8_policy_engines_capsule].depends_on",
"url": "paper-module-batch8-policy-engines-capsule.html"
},
{
"id": "paper_module.batch9_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 9 Source Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[70:paper_module.batch9_macro_engines_capsule].depends_on",
"url": "paper-module-batch9-macro-engines-capsule.html"
},
{
"id": "paper_module.bridge_phase_continuity_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Bridge Phase Continuity Runtime",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[10:paper_module.bridge_phase_continuity_runtime].depends_on",
"url": "paper-module-bridge-phase-continuity-runtime.html"
},
{
"id": "paper_module.cognitive_operator_registry",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cognitive Operator Registry",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[11:paper_module.cognitive_operator_registry].depends_on",
"url": "paper-module-cognitive-operator-registry.html"
},
{
"id": "paper_module.computer_use_action_trace_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Computer-Use Action Trace Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[46:paper_module.computer_use_action_trace_replay].depends_on",
"url": "paper-module-computer-use-action-trace-replay.html"
},
{
"id": "paper_module.engine_room_generated_projection_drift_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Generated Projection Drift Gate",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[80:paper_module.engine_room_generated_projection_drift_gate].depends_on",
"url": "paper-module-engine-room-generated-projection-drift-gate.html"
},
{
"id": "paper_module.executable_doctrine_grammar",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Executable Doctrine Grammar",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[18:paper_module.executable_doctrine_grammar].depends_on",
"url": "paper-module-executable-doctrine-grammar.html"
},
{
"id": "paper_module.finance_forecast_evaluation_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Finance Forecast Evaluation Spine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[30:paper_module.finance_forecast_evaluation_spine].depends_on",
"url": "paper-module-finance-forecast-evaluation-spine.html"
},
{
"id": "paper_module.indirect_prompt_injection_information_flow_policy_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Indirect Prompt-Injection Information-Flow Policy Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[38:paper_module.indirect_prompt_injection_information_flow_policy_replay].depends_on",
"url": "paper-module-indirect-prompt-injection-information-flow-policy-replay.html"
},
{
"id": "paper_module.materials_chemistry_closed_loop_lab_safety_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Materials Chemistry Closed-Loop Lab-Safety Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[6:paper_module.materials_chemistry_closed_loop_lab_safety_replay].depends_on",
"url": "paper-module-materials-chemistry-closed-loop-lab-safety-replay.html"
},
{
"id": "paper_module.mcp_tool_authority_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "MCP Tool Authority Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[39:paper_module.mcp_tool_authority_replay].depends_on",
"url": "paper-module-mcp-tool-authority-replay.html"
},
{
"id": "paper_module.standards_meta_diagnostics",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Standards Meta Diagnostics",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[29:paper_module.standards_meta_diagnostics].depends_on",
"url": "paper-module-standards-meta-diagnostics.html"
},
{
"id": "paper_module.work_landing_control_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Work Landing Control Spine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[17:paper_module.work_landing_control_spine].depends_on",
"url": "paper-module-work-landing-control-spine.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[27:paper_module.world_model_projection_drift_control_room].depends_on",
"url": "paper-module-world-model-projection-drift-control-room.html"
}
],
"paper_module_id": "paper_module.macro_projection_import_protocol",
"reader_packet": {
"abstract": "The public source-projection import protocol validates classified source-to-Microcosm projection cells, per-slice manifests, digest relations, omission result records, intake statuses, formal-math claim boundaries, and scope limits without claiming source or launch-scope decision.",
"authority_ceiling": "Verified source body imports, fixture result records, exported projection-bundle result records, per-slice manifest refs, metadata-only formal-math metadata and validation refs, and metadata-only result record fields only; no theorem/formal-result correctness authority, private target status, challenge-result public sharing, live source authority, whole-system equivalence, launch, public sharing, hosted deployment, recipient work, provider/oracle payload authority, provider or Lean/Lake execution, secret/non-public body export, source-file changes, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.macro_projection_import_protocol.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-macro-projection-import-protocol-diagram-1",
"html_url": "docs/paper-module-macro-projection-import-protocol.html#paper-module-macro-projection-import-protocol-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Copied row source ref, target ref, digest, material class\"] --> B[\"validate_projection_protocol\"] B --> C[\"Class in forbidden set Reject: forbidden body import\"] B --> D[\"Class _public_safe_import_status\"] D --> E[\"_body_import_verification_findings mode, verified flag, digest binding\"] E --> F[\"Exact copy: re-hash live source\"] F --> G[\"Match: body floor\"] F --> H[\"Mismatch: flag live source drift, refresh later\"] H --> G G --> I[\"metadata-only result record\"]",
"source_format": "mermaid",
"source_hash": "sha256:0f0b59e6f87006b57ac293fb05ba80365e43cd3116c51d3aa894868cc59b5d7d",
"source_line_end": 145,
"source_line_start": 134
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
}
],
"relation": "abides by"
},
{
"count": 7,
"links": [
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
},
{
"id": "paper_module.formal_math_readiness_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Readiness Gate",
"relation": "depends on",
"url": "paper-module-formal-math-readiness-gate.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.batch11_saturation_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 11 Saturation Engines Bundle",
"relation": "depends on",
"url": "paper-module-batch11-saturation-engines-capsule.html"
},
{
"id": "paper_module.batch8_compliance_pipeline_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Compliance Pipeline Bundle",
"relation": "depends on",
"url": "paper-module-batch8-compliance-pipeline-capsule.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "macro_projection_import_protocol",
"kind": "organ",
"kind_label": "component",
"label": "Source Projection Import Protocol",
"relation": "explains",
"url": "component-macro-projection-import-protocol.html"
},
{
"id": "mechanism.macro_projection_import_protocol.validates_public_macro_projection_imports",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public source projection imports",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-macro-projection-import-protocol-validates-public-macro-projection-imports"
}
],
"relation": "explains"
},
{
"count": 4,
"links": [
{
"id": "P-14",
"kind": "principle",
"kind_label": "principle",
"label": "Carry basis and provenance together",
"relation": "governed by",
"url": "doctrine.html#dcard-p-14"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.macro_projection_import_protocol",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 232,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/macro_projection_import_protocol.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "W3C PROV",
"section_id": "the-check-follows-established-provenance-and-software-supply-chain-patterns",
"url": "https://www.w3.org/TR/prov-overview/"
},
{
"kind": "external_reference",
"label": "SLSA specification",
"section_id": "the-check-follows-established-provenance-and-software-supply-chain-patterns",
"url": "https://slsa.dev/spec/"
},
{
"kind": "external_reference",
"label": "in-toto",
"section_id": "the-check-follows-established-provenance-and-software-supply-chain-patterns",
"url": "https://in-toto.io/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "macroprojectionimportprotocol is the checkpoint for copying material out of a much larger source project into the public Microcosm folder.** macroprojectionimport_protocol is the checkpoint for copying material out of a much larger source project into the public Microcosm folder. It reads a proposed set of copied files, classifies each one, and decides whether the copy is allowed out and honestly tied to the source it claims to come from.",
"level": 4,
"order": 1,
"section_id": "the-checkpoint-for-copying-macro-material",
"source_line_end": 10,
"source_line_start": 3,
"title": "The checkpoint for copying source material"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It exists so the public folder can stay dense and real without becoming a place where a secret, an operator conversation, a model-output data, or launch material leaks out alongside genuinely useful code. It exists so the public folder can stay dense and real without becoming a place where a secret, an operator conversation, a model-output data, or launch material leaks out alongside genuinely useful code. Every run answers one question per copied file: was this body allowed out, and does the public copy match the source it names.",
"level": 4,
"order": 2,
"section_id": "it-exists-so-the-public-folder-can-stay-dense",
"source_line_end": 19,
"source_line_start": 11,
"title": "It exists so the public folder can stay dense"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The output is a metadata-only result record. It lists what was imported, what was deliberately left out with a note saying so, and what was refused. The output is a metadata-only result record. It lists what was imported, what was deliberately left out with a note saying so, and what was refused. The record carries paths, hashes, counts, and validation refs, never the private bodies themselves.",
"level": 4,
"order": 3,
"section_id": "the-output-is-a-body-free-result-record",
"source_line_end": 27,
"source_line_start": 20,
"title": "The output is a metadata-only result record"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 29,
"source_line_start": 28,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Microcosm grows by copying real files out of a private codebase. A loose copy is the exact artifact that quietly carries a account secret or an operator thread along with the code you wanted. Microcosm grows by copying real files out of a private codebase. A loose copy is the exact artifact that quietly carries a account secret or an operator thread along with the code you wanted. A trust statement (\"we checked, it is fine\") does not scale and cannot be re-run by a reader.",
"level": 4,
"order": 5,
"section_id": "microcosm-grows-by-copying-real-files-out-of-a-private-codebase",
"source_line_end": 37,
"source_line_start": 30,
"title": "Microcosm grows by copying real files out of a private codebase"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "So the check is an accounting check, not a trust statement. Each copied row declares its source ref, its public target ref, a content digest, and a material class. So the check is an accounting check, not a trust statement. Each copied row declares its source ref, its public target ref, a content digest, and a material class. The protocol sorts the class into one of two sets. PUBLICSAFEBODYMATERIALCLASSES holds five classes that may be copied with provenance: pattern, standard, tool, result-record, and proof bodies. TRUEFORBIDDENMATERIAL_CLASSES holds fifteen classes that can never appear as an imported body, including raw operator source, operator thread, model-output data, account secret,\u2026",
"level": 4,
"order": 6,
"section_id": "so-the-check-is-an-accounting-check-not-a-trust-statement",
"source_line_end": 52,
"source_line_start": 38,
"title": "So the check is an accounting check, not a trust statement"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 54,
"source_line_start": 53,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "main dispatches four commands. run validates a first-wave fixture directory (the positive slices plus the named negative cases) and writes result records through writereceipts.** main dispatches four commands. run validates a first-wave fixture directory (the positive slices plus the named negative cases) and writes result records through writereceipts. runprojection_bundle validates an exported bundle and writes one exportedprojectionimportbundlevalidation_result.json. plan calls previewimportplan, which scores each proposed cell and writes nothing. refresh-exact-copy-source-modules calls refreshexactcopysourcemodules to re-copy exact-copy slices whose source has moved on. run and runproje\u2026",
"level": 4,
"order": 8,
"section_id": "main-dispatches-four-commands",
"source_line_end": 70,
"source_line_start": 55,
"title": "`main` dispatches four commands"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "The core is validateprojectionprotocol. It reads the protocol payload and runs these checks in order: The core is validateprojectionprotocol. It reads the protocol payload and runs these checks in order:",
"level": 4,
"order": 9,
"section_id": "the-core-is-validate-projection-protocol",
"source_line_end": 86,
"source_line_start": 71,
"title": "The core is `validate_projection_protocol`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "First it enforces a density floor: the protocol must cite at least two source refs, two public runtime refs, and two validation refs, or the row fails as MACROPROJECTIONPROTOCOLDENSITYMISSING. First it enforces a density floor: the protocol must cite at least two source refs, two public runtime refs, and two validation refs, or the row fails as MACROPROJECTIONPROTOCOLDENSITYMISSING. Then it walks each copied row. A row whose class is goes to publicsafeimportstatus; anything it flags becomes a blocking finding. A row whose class is in TRUEFORBIDDENMATERIALCLASSES, or any row that sets bodycopied without a valid verification record, is rejected as MACROPROJECTIONFORBIDDENBODYIMPORT. Omitted r\u2026",
"level": 4,
"order": 10,
"section_id": "first-it-enforces-a-density-floor",
"source_line_end": 100,
"source_line_start": 87,
"title": "First it enforces a density floor"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "For a body that is copied, bodyimportverificationfindings does the real integrity work. For a body that is copied, bodyimportverificationfindings does the real integrity work. The verification record must be marked verified and must use one of two modes in BODYIMPORTVERIFICATIONMODES: exactsourcedigestmatch or verifiedlightedit_recipe. The target body digest must equal the declared target digest. For an exact copy it also re-hashes the named source, requires a relation from EXACTCOPYSOURCETOTARGET_RELATIONS, and checks the source digest.",
"level": 4,
"order": 11,
"section_id": "digest-requirements-for-a-copied-body",
"source_line_end": 111,
"source_line_start": 101,
"title": "Digest requirements for a copied body"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The load-bearing detail is how the protocol treats an exact copy whose source has since changed. A mismatch on an exact-copy row is not a failed import. The load-bearing detail is how the protocol treats an exact copy whose source has since changed. A mismatch on an exact-copy row is not a failed import. It is recorded as live source drift: the original copy was still honest, the source has simply moved on, and the row is flagged for the refresh path rather than failed. That separates a dishonest import from a stale one, so a routine upstream edit is not mistaken for a broken proof, and the public copy stays faithful without tracking every upstream change in lock-step.",
"level": 4,
"order": 12,
"section_id": "the-load-bearing-detail-is-how-the-protocol-treats-an-exact-copy-whose-source-has-since-changed",
"source_line_end": 122,
"source_line_start": 112,
"title": "The load-bearing detail is how the protocol treats an exact copy whose source has since changed"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "validateimportplan guards the forward direction. Each proposed cell must name source and target refs, must name validation refs, must not use a private-system- shaped ref, and must only reference body material that the protocol already carries. validateimportplan guards the forward direction. Each proposed cell must name source and target refs, must name validation refs, must not use a private-system- shaped ref, and must only reference body material that the protocol already carries. A plan needs at least three cells to pass. previewimportplan wraps this as the macroprojectionimportintakepreview_v1 record: it scores each cell, projects the intake and severance boards, and lists blocking su\u2026",
"level": 4,
"order": 13,
"section_id": "validate-import-plan-guards-the-forward-direction",
"source_line_end": 146,
"source_line_start": 123,
"title": "`validate_import_plan` guards the forward direction"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 148,
"source_line_start": 147,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The first-wave fixture drives these named rejections through validateprojectionprotocol, validateimportplan, and validatestandaloneruntimeseverance:** The first-wave fixture drives these named rejections through validateprojectionprotocol, validateimportplan, and validatestandaloneruntime_severance: forbiddenbodyimport_overclaim: a request to import a account secret or account-bound body. missingomissionreceipt: omitted material with no omission result record ref. authorityupgradeoverclaim: a claim that a public import grants live private-source authority. releaseorprivateequivalenceoverclaim: a launch, public sharing, recipient, or private-equivalence claim. formalmathclaimboundaryoverclai\u2026",
"level": 4,
"order": 15,
"section_id": "the-first-wave-fixture-drives-these-named-rejections-through-validate-projection-protocol",
"source_line_end": 168,
"source_line_start": 149,
"title": "The first-wave fixture drives these named rejections through `validate_projection_protocol`"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 170,
"source_line_start": 169,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The check follows established provenance and software-supply-chain patterns: a copied or refactored artifact needs source refs, target refs, digests, validation refs, omission records, and a stated claim boundary. The check follows established provenance and software-supply-chain patterns: a copied or refactored artifact needs source refs, target refs, digests, validation refs, omission records, and a stated claim boundary. The closest public anchors are W3C PROV for entity, activity, and agent provenance, the SLSA specification for artifact integrity and provenance in software supply chains, and in-toto for linking supply-chain steps through signed metadata. This component applies those pa\u2026",
"level": 4,
"order": 17,
"section_id": "the-check-follows-established-provenance-and-software-supply-chain-patterns",
"source_line_end": 184,
"source_line_start": 171,
"title": "The check follows established provenance and software-supply-chain patterns"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate/, reproduce this page's checks with temporary result records:",
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 197,
"source_line_start": 185,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means the fixture and bundle slices carried their required refs and digests, the named negative cases were rejected, and the result records stayed metadata-only. A pass means the fixture and bundle slices carried their required refs and digests, the named negative cases were rejected, and the result records stayed metadata-only. Re-enter this module when a new projection cell lands, a source-module manifest is refreshed, or a result record count changes.",
"level": 4,
"order": 19,
"section_id": "what-the-fixture-and-bundle-validate",
"source_line_end": 205,
"source_line_start": 198,
"title": "What the fixture and bundle validate"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 207,
"source_line_start": 206,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim this component supports: for a named slice, it validated classification, source-to-target digest binding, per-slice manifests, omission records, the named negative cases, and a metadata-only result-record policy. The strongest claim this component supports: for a named slice, it validated classification, source-to-target digest binding, per-slice manifests, omission records, the named negative cases, and a metadata-only result-record policy. Accepted records expose the current publicsafebodyimportcount, publicsafebodyimportstatus, projectionstatuscounts, openactionablecellcount, and secretexclusion_scan fields.",
"level": 4,
"order": 21,
"section_id": "the-strongest-import-accounting-claim",
"source_line_end": 217,
"source_line_start": 208,
"title": "The strongest import-accounting claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "That is the proof boundary and the scope limit. A passing scan is a declared-input regression gate over the material classes the protocol recognizes, not a complete account secret audit and bounded evidence that no private material exists anywhere. That is the proof boundary and the scope limit. A passing scan is a declared-input regression gate over the material classes the protocol recognizes, not a complete account secret audit and bounded evidence that no private material exists anywhere. It does not show the public copy is complete, equal to the private root, or ready to launch. It does not grant authority over live private source, and it does not permit public sharing, hosted deployme\u2026",
"level": 4,
"order": 22,
"section_id": "the-import-protocol-s-authority-ceiling",
"source_line_end": 231,
"source_line_start": 218,
"title": "The import protocol's scope limit"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/macro_projection_import_protocol.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public source-projection import protocol validates classified source-to-Microcosm projection cells, per-slice manifests, digest relations, omission result records, intake statuses, formal-math claim boundaries, and scope limits without claiming source or launch-scope decision."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "macro_projection_import_protocol.py",
"path": "src/microcosm_core/organs/macro_projection_import_protocol.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for source projection import validation, intake previews, exported bundle validation, per-slice source-module refresh, and metadata-only result records.",
"symbols": [
"run",
"run_projection_bundle",
"preview_import_plan",
"validate_projection_protocol",
"validate_import_plan",
"refresh_exact_copy_source_modules",
"main"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/macro_projection_import_protocol.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/macro_projection_import_protocol.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/macro_projection_import_protocol.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Source Projection Import Protocol is the public import membrane for source-backed Microcosm growth. It validates fixture and exported projection bundles by checking source refs, public target refs, content digests, source-to-target relations, per-slice source-module manifests, validation refs, omitted-material result records, metadata-only result record policy, secret-exclusion scans, projection cell statuses, formal-math claim-boundary negatives, and other negative cases while keeping true private bodies, theorem/formal-result correctness claims, provider/oracle payloads, launch material, and static count claims out of public authority.",
"summary_line": "The public source-projection import protocol validates classified source-to-Microcosm projection cells, per-slice manifests, digest relations, omission result records, intake statuses, formal-math claim boundaries, and scope limits without claiming source or launch-scope decision.",
"title": "Source Projection Import Protocol"
},
{
"anchor": "paper-module-mission-transaction-work-spine",
"authority_ceiling": "Public fixture and exported-bundle result records only; no live work log mutation, live work log mutation, live git mutation, private backup execution, broad checkpoint authorization, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
The riskiest moment in agentic code work is the one that feels safest. An agent runs a few checks, sees no errors, and concludes its work is finished and committed. The riskiest moment in agentic code work is the one that feels safest. An agent runs a few checks, sees no errors, and concludes its work is finished and committed. Those are two different facts. A clean preflight describes the state of the checks. It says nothing about whether a competing claim already owns the same file, whether the branch moved underneath the agent, or whether the commit ever landed.
\n
mission_transaction_work_spine is a fixture that replays a fixed set of work-landing situations against a toy repository and records, for each one, whether the change would be allowed to land or would be blocked.mission_transaction_work_spine is a fixture that replays a fixed set of work-landing situations against a toy repository and records, for each one, whether the change would be allowed to land or would be blocked. It reads recorded inputs, recomputes each decision from copied internal control source code, and writes one result record per situation that says exactly why the situation was permitted or refused.
\n
Recompute decisions instead of trusting labels
\n
The design choice that matters is that the validator refuses to trust the labels baked into its own inputs. The design choice that matters is that the validator refuses to trust the labels baked into its own inputs. It then perturbs the input one field at a time and requires the decision to move.
\n
Purpose
\n
A loose check passes whenever its inputs carry the right labels. That is the failure this component is built against. A loose check passes whenever its inputs carry the right labels. That is the failure this component is built against. Agentic work fails most often at transaction boundaries, not at isolated syntax errors: a clean preflight read as a landed commit, a competing claim on the same file ignored, a claim whose expected parent no longer matches the repository, a downstream item marked ready without its hard dependency, a dirty tree used to justify a broad commit with no operator intent, or a result record that smuggles private ledger bodies into a public artifact.
\n
The single question the fixture answers is narrow. What evidence has to hold before a unit of work is allowed to land, and is that evidence checkable rather than asserted? The single question the fixture answers is narrow. What evidence has to hold before a unit of work is allowed to land, and is that evidence checkable rather than asserted? A cold reader can open the inputs, the copied source bodies that implement the checks, and the result records, and see which decision came from which check.
\n
How it works
\n
The module has two entry paths, selected by subcommand. run validates the first-wave replay fixture. The module has two entry paths, selected by subcommand. run validates the first-wave replay fixture. run_mission_transaction_bundle validates the exported bundle of copied source modules. Both write metadata-only result records.
\n
run loads the fixture payloads, scans every input file for forbidden private classes, then calls one validator per concern and merges their findings.run loads the fixture payloads, scans every input file for forbidden private classes, then calls one validator per concern and merges their findings. Status is pass only when no expected negative case is missing, the secret scan passes, and the real-snapshot replay passes. The validators it calls carry the mechanism.
\n
validate_claim_preflight reads the active claims and the toy repository state. It walks the claims, records the first owner of each path, and emits SAME_PATH_CLAIM_CONFLICT when a second active claim wants a path that is already owned.validate_claim_preflight reads the active claims and the toy repository state. It walks the claims, records the first owner of each path, and emits SAME_PATH_CLAIM_CONFLICT when a second active claim wants a path that is already owned. It compares each claim's expected_parent_sha against the current parent in the repository state and emits EXPECTED_PARENT_MISMATCH when they differ. A claim with no owned path emits MISSING_OWNED_PATH.
\n
validate_checkpoint_lane_policy reads a set of commit-lane cases. For each case it computes the recommended lane and compares it to the selected lane.validate_checkpoint_lane_policy reads a set of commit-lane cases. For each case it computes the recommended lane and compares it to the selected lane. A broad_checkpoint lane without operator_authorized_broad_checkpoint set to true emits CHECKPOINT_BROAD_AUTH_REQUIRED. A suspected secret not routed to the hard-stop lane emits CHECKPOINT_SECRET_REQUIRES_HARD_STOP. A dirty tree with isolated owned paths must not block a scoped commit, so a case that treats it as a scoped blocker emits CHECKPOINT_DIRTY_TREE_NOT_SCOPED_BLOCKER. A case with no selected lane emits CHECKPOINT_LANE_DECISION_REQUIRED.
\n
validate_real_active_claims_snapshot is the discriminating check
\n
**validate_real_active_claims_snapshot is the discriminating check. It then calls active_claim_collisions_for_paths three times: on the real status, on a status mutated to add a same-path conflict, and on a status mutated to add a disjoint-path claim. The real status must return no collision for the requested path, the same-path mutation must return one, and the disjoint mutation must return none. It also mutates the expected parent and the checkpoint lane row and requires each mutation to block while the equal-parent and disjoint cases stay clear. Only when the good case passes, every mutation blocks, every clear case clears, the source hash binds, and the session heartbeat binds does it award the R3 rung public_safe_real_work_ledger_session_snapshot_replay. Otherwise it falls to R2 or R1.
\n
run_mission_transaction_bundle validates copied control source. The four import validators (validate_task_ledger_source_import, validate_work_ledger_source_import, validate_checkpoint_source_import, validate_mission_control_source_import) each read a source-module manifest and a runtime contract, check that the manifest declares exactly the expected module ids and count, then for each module confirm the copied file exists, contains its required internal control anchor strings, matches the manifest sha256, and carries classification: copied_non_secret_macro_body with body_in_receipt: false.run_mission_transaction_bundle validates copied control source. The four import validators (validate_task_ledger_source_import, validate_work_ledger_source_import, validate_checkpoint_source_import, validate_mission_control_source_import) each read a source-module manifest and a runtime contract, check that the manifest declares exactly the expected module ids and count, then for each module confirm the copied file exists, contains its required internal control anchor strings, matches the manifest sha256, and carries classification: copied_non_secret_macro_body with body_in_receipt: false. The bundle path also requires the real claims snapshot to be present and to pass.
\n
Nine metadata-only transaction records
\n
write_receipts writes the nine result records for a fixture run (preflight, dependency-blocked, work-landing attempt, claim preflight, scoped mutation, checkpoint lane, completion projection, dependency-unlock scheduler, reconcile plan).write_receipts writes the nine result records for a fixture run (preflight, dependency-blocked, work-landing attempt, claim preflight, scoped mutation, checkpoint lane, completion projection, dependency-unlock scheduler, reconcile plan). Each record carries a schema version, refs, counts, hashes, verdicts, and body_in_receipt: false. Source bodies live only in the exported bundle's source_modules/ tree; they never enter a result record.
\nDiagram of the mechanism (6 steps).Diagram source & refs
Source refs
+ checkpoint lane policy
validate_claim_preflight
Replay
validate_real_active_claims_snapshot
flowchart TD Fixture["First-wave fixture claims, deps, lane cases, snapshot"] Bundle["Exported bundle copied source modules"] Preflight["validate_claim_preflight + checkpoint lane policy"] Replay["validate_real_active_claims_snapshot runtime-rebuilt snapshot + perturbations"] Imports["four source-import validators manifest, anchors, sha256"] Records["metadata-only result records refs, hashes, verdicts, limits"] Fixture --> Preflight Fixture --> Replay Bundle --> Imports Preflight --> Records Replay --> Records Imports --> Records
\n
Negative cases
\n
The positive claim is not that the fixture passes. It is that the fixture accepts good evidence and rejects targeted one-field perturbations. The positive claim is not that the fixture passes. It is that the fixture accepts good evidence and rejects targeted one-field perturbations. Fourteen named negative cases are required, each bound to its error code:
\n
competing_claim_and_stale_parent must observe both SAME_PATH_CLAIM_CONFLICT and EXPECTED_PARENT_MISMATCH
\n
**competing_claim_and_stale_parent must observe both SAME_PATH_CLAIM_CONFLICT and EXPECTED_PARENT_MISMATCH. mission_claim_missing_owned_path observes MISSING_OWNED_PATH. scoped_commit_receipt_claims_global_authority observes SCOPED_RECEIPT_AUTHORITY_UPGRADE. mission_fixture_private_task_ledger_body observes LIVE_TASK_LEDGER_BODY_IN_FIXTURE. clean_preflight_overclaims_landing_complete observes PREFLIGHT_PASS_OVERCLAIMS_WORK_LANDED. dependency_unlock_without_resolution_receipt observes DANGLING_DEPENDENCY_REF and ready_workitem_with_unsatisfied_hard_dep observes READY_WITH_INCOMPLETE_HARD_DEP. The four checkpoint-lane cases (broad_checkpoint_without_operator_authorization, suspected_secret_without_hard_stop, dirty_tree_blocks_scoped_lane, missing_selected_lane_for_state) observe the lane error codes above. A run is blocked if any of these expected cases is not observed.
\n
Prior Art Grounding
\n
This is the mission-transaction member of a small family of work-landing fixtures. Its closest sibling is durable_agent_work_landing_replay, which checks recorded landing rows, validation-before-commit ordering, HEAD movement, blocker capture, and completion evidence without running live git. This is the mission-transaction member of a small family of work-landing fixtures. Its closest sibling is durable_agent_work_landing_replay, which checks recorded landing rows, validation-before-commit ordering, HEAD movement, blocker capture, and completion evidence without running live git. This module narrows that pattern to the transaction preflight and the work-ledger claims membrane: same-path conflicts, expected-parent mismatches, checkpoint-lane selection, dependency unlocks, and session finalisation. The lineage is transactional-integrity engineering, specifically compare-and-set on a parent commit and single-writer path claims. There is no external citation to invent here; the pattern is standard optimistic-concurrency control applied to a repo.
A pass means the fixture rows, the copied source bodies, the real snapshot replay, and the fourteen negative cases all hold together, and the result records carry no private body. A pass means the fixture rows, the copied source bodies, the real snapshot replay, and the fourteen negative cases all hold together, and the result records carry no private body.
\n
Scope boundary
\n
Scope limit
\n
The R3 replay-and-evidence-shape claim
\n
The strongest claim the evidence supports: the public fixture rows, the copied control source bodies, the real work-ledger session-snapshot replay, the discriminating negative cases, and the metadata-only result records preserve the work-landing contract at the R3 rung, and the checks that discriminate still discriminate. The strongest claim the evidence supports: the public fixture rows, the copied control source bodies, the real work-ledger session-snapshot replay, the discriminating negative cases, and the metadata-only result records preserve the work-landing contract at the R3 rung, and the checks that discriminate still discriminate. That is a replay-and-evidence-shape claim. The proof boundary is those inputs and their recomputation. Nothing runs live.
\n
Within that scope limit the module refuses more. It does not mutate the live work log, the live work log, or git. Within that scope limit the module refuses more. It does not mutate the live work log, the live work log, or git. It does not certify completion of the current repository. It grants no authority to commit, checkpoint broadly, back up, or launch, and it makes no claim about provider behaviour, browser state, hosted-product readiness, or whole-system correctness.
\n",
"components": [
{
"anchor": "component-mission_transaction_work_spine",
"component_id": "mission_transaction_work_spine",
"public_label": "Mission Transaction Work Spine"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 2
},
"link_count": 7,
"links": [
{
"id": "mission_transaction_work_spine",
"kind": "organ",
"kind_label": "component",
"label": "Mission Transaction Work Spine",
"relation": "explains",
"url": "component-mission-transaction-work-spine.html"
},
{
"id": "mechanism.mission_transaction_work_spine.validates_public_mission_transaction_bundle",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public mission transaction bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-mission-transaction-work-spine-validates-public-mission-transaction-bundle"
},
{
"id": "P-10",
"kind": "principle",
"kind_label": "principle",
"label": "Do not land effects without compensation",
"relation": "governed by",
"url": "doctrine.html#dcard-p-10"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
},
{
"id": "concept.work_landing_and_continuity_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "work landing and continuity controls as transaction-scoped coordination evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-work-landing-and-continuity-control-bundle"
},
{
"id": "paper_module.durable_agent_work_landing_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Durable Agent Work-Landing Replay",
"relation": "depends on",
"url": "paper-module-durable-agent-work-landing-replay.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.agent_benchmark_integrity_anti_gaming_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Benchmark Integrity Anti-Gaming Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[3:paper_module.agent_benchmark_integrity_anti_gaming_replay].depends_on",
"url": "paper-module-agent-benchmark-integrity-anti-gaming-replay.html"
},
{
"id": "paper_module.agentic_vulnerability_discovery_patch_proof_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agentic Vulnerability Discovery Patch-Proof Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[5:paper_module.agentic_vulnerability_discovery_patch_proof_replay].depends_on",
"url": "paper-module-agentic-vulnerability-discovery-patch-proof-replay.html"
},
{
"id": "paper_module.batch10_live_source_drift_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 10 Live Source Drift Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[68:paper_module.batch10_live_source_drift_capsule].depends_on",
"url": "paper-module-batch10-live-source-drift-capsule.html"
},
{
"id": "paper_module.batch4_proof_authority_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 4 Proof, Authority, and Runtime Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[74:paper_module.batch4_proof_authority_runtime].depends_on",
"url": "paper-module-batch4-proof-authority-runtime.html"
},
{
"id": "paper_module.batch8_tools_tail_primitives_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Tools-Tail Primitives Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[62:paper_module.batch8_tools_tail_primitives_capsule].depends_on",
"url": "paper-module-batch8-tools-tail-primitives-capsule.html"
},
{
"id": "paper_module.bounded_autonomy_campaign_packet",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Bounded Autonomy Campaign Packet",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[45:paper_module.bounded_autonomy_campaign_packet].depends_on",
"url": "paper-module-bounded-autonomy-campaign-packet.html"
},
{
"id": "paper_module.concurrency_mission_control",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Concurrency Mission Control",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[47:paper_module.concurrency_mission_control].depends_on",
"url": "paper-module-concurrency-mission-control.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[19:paper_module.macro_projection_import_protocol].depends_on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.proof_derived_governed_mutation_authorization",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof-Derived Governed Mutation Authorization",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[15:paper_module.proof_derived_governed_mutation_authorization].depends_on",
"url": "paper-module-proof-derived-governed-mutation-authorization.html"
},
{
"id": "paper_module.tool_server_pressure_inventory",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tool Server Pressure Inventory",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[50:paper_module.tool_server_pressure_inventory].depends_on",
"url": "paper-module-tool-server-pressure-inventory.html"
},
{
"id": "paper_module.work_landing_control_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Work Landing Control Spine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[17:paper_module.work_landing_control_spine].depends_on",
"url": "paper-module-work-landing-control-spine.html"
}
],
"paper_module_id": "paper_module.mission_transaction_work_spine",
"reader_packet": {
"abstract": "The public mission-transaction fixture validates work-landing, claim, dependency, scoped-commit, checkpoint-lane, result record-drain, completion, and copied control source-module contracts without mutating live ledgers or git.",
"authority_ceiling": "Public fixture and exported-bundle result records only; no live work log mutation, live work log mutation, live git mutation, private backup execution, broad checkpoint authorization, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.mission_transaction_work_spine.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-mission-transaction-work-spine-diagram-1",
"html_url": "docs/paper-module-mission-transaction-work-spine.html#paper-module-mission-transaction-work-spine-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Fixture[\"First-wave fixture claims, deps, lane cases, snapshot\"] Bundle[\"Exported bundle copied source modules\"] Preflight[\"validate_claim_preflight + checkpoint lane policy\"] Replay[\"validate_real_active_claims_snapshot runtime-rebuilt snapshot + perturbations\"] Imports[\"four source-import validators manifest, anchors, sha256\"] Records[\"metadata-only result records refs, hashes, verdicts, limits\"] Fixture --> Preflight Fixture --> Replay Bundle --> Imports Preflight --> Records Replay --> Records Imports --> Records",
"source_format": "mermaid",
"source_hash": "sha256:0d4b8f050afcad738ea49bc4fb13db6236bbdf0ac31a1d5b242b3cab73fb7031",
"source_line_end": 152,
"source_line_start": 137
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.durable_agent_work_landing_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Durable Agent Work-Landing Replay",
"relation": "depends on",
"url": "paper-module-durable-agent-work-landing-replay.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "mission_transaction_work_spine",
"kind": "organ",
"kind_label": "component",
"label": "Mission Transaction Work Spine",
"relation": "explains",
"url": "component-mission-transaction-work-spine.html"
},
{
"id": "mechanism.mission_transaction_work_spine.validates_public_mission_transaction_bundle",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public mission transaction bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-mission-transaction-work-spine-validates-public-mission-transaction-bundle"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-10",
"kind": "principle",
"kind_label": "principle",
"label": "Do not land effects without compensation",
"relation": "governed by",
"url": "doctrine.html#dcard-p-10"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "concept.work_landing_and_continuity_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "work landing and continuity controls as transaction-scoped coordination evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-work-landing-and-continuity-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.mission_transaction_work_spine",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 255,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/mission_transaction_work_spine.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "The riskiest moment in agentic code work is the one that feels safest. An agent runs a few checks, sees no errors, and concludes its work is finished and committed. The riskiest moment in agentic code work is the one that feels safest. An agent runs a few checks, sees no errors, and concludes its work is finished and committed. Those are two different facts. A clean preflight describes the state of the checks. It says nothing about whether a competing claim already owns the same file, whether the branch moved underneath the agent, or whether the commit ever landed.",
"level": 4,
"order": 1,
"section_id": "the-riskiest-moment-in-agentic-code-work-is-the-one-that-feels-safest",
"source_line_end": 12,
"source_line_start": 3,
"title": "The riskiest moment in agentic code work is the one that feels safest"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "missiontransactionworkspine is a fixture that replays a fixed set of work-landing situations against a toy repository and records, for each one, whether the change would be allowed to land or would be blocked.** missiontransactionwork_spine is a fixture that replays a fixed set of work-landing situations against a toy repository and records, for each one, whether the change would be allowed to land or would be blocked. It reads recorded inputs, recomputes each decision from copied internal control source code, and writes one result record per situation that says exactly why the situation was permitted or refused.",
"level": 4,
"order": 2,
"section_id": "mission-transaction-work-spine-is-a-fixture-that-replays-a-fixed-set-of-work-landing-situations-against-a-toy-repository",
"source_line_end": 22,
"source_line_start": 13,
"title": "`mission_transaction_work_spine` is a fixture that replays a fixed set of work-landing situations against a toy repository"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The design choice that matters is that the validator refuses to trust the labels baked into its own inputs. The design choice that matters is that the validator refuses to trust the labels baked into its own inputs. It imports a copied, copy of the payload, and re-derives each pass-or-block decision from that recomputation. It then perturbs the input one field at a time and requires the decision to move.",
"level": 4,
"order": 3,
"section_id": "recompute-decisions-instead-of-trusting-labels",
"source_line_end": 31,
"source_line_start": 23,
"title": "Recompute decisions instead of trusting labels"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 33,
"source_line_start": 32,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A loose check passes whenever its inputs carry the right labels. That is the failure this component is built against. A loose check passes whenever its inputs carry the right labels. That is the failure this component is built against. Agentic work fails most often at transaction boundaries, not at isolated syntax errors: a clean preflight read as a landed commit, a competing claim on the same file ignored, a claim whose expected parent no longer matches the repository, a downstream item marked ready without its hard dependency, a dirty tree used to justify a broad commit with no operator intent, or a result record that smuggles private ledger bodies into a public artifact.",
"level": 4,
"order": 5,
"section_id": "a-loose-check-passes-whenever-its-inputs-carry-the-right-labels",
"source_line_end": 45,
"source_line_start": 34,
"title": "A loose check passes whenever its inputs carry the right labels"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The single question the fixture answers is narrow. What evidence has to hold before a unit of work is allowed to land, and is that evidence checkable rather than asserted? The single question the fixture answers is narrow. What evidence has to hold before a unit of work is allowed to land, and is that evidence checkable rather than asserted? A cold reader can open the inputs, the copied source bodies that implement the checks, and the result records, and see which decision came from which check.",
"level": 4,
"order": 6,
"section_id": "the-single-question-the-fixture-answers-is-narrow",
"source_line_end": 54,
"source_line_start": 46,
"title": "The single question the fixture answers is narrow"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 56,
"source_line_start": 55,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The module has two entry paths, selected by subcommand. run validates the first-wave replay fixture. The module has two entry paths, selected by subcommand. run validates the first-wave replay fixture. runmissiontransaction_bundle validates the exported bundle of copied source modules. Both write metadata-only result records.",
"level": 4,
"order": 8,
"section_id": "the-module-has-two-entry-paths-selected-by-subcommand",
"source_line_end": 63,
"source_line_start": 57,
"title": "The module has two entry paths, selected by subcommand"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run loads the fixture payloads, scans every input file for forbidden private classes, then calls one validator per concern and merges their findings. run loads the fixture payloads, scans every input file for forbidden private classes, then calls one validator per concern and merges their findings. Status is pass only when no expected negative case is missing, the secret scan passes, and the real-snapshot replay passes. The validators it calls carry the mechanism.",
"level": 4,
"order": 9,
"section_id": "run-loads-the-fixture-payloads",
"source_line_end": 72,
"source_line_start": 64,
"title": "`run` loads the fixture payloads"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateclaimpreflight reads the active claims and the toy repository state. It walks the claims, records the first owner of each path, and emits SAMEPATHCLAIMCONFLICT when a second active claim wants a path that is already owned.** validateclaimpreflight reads the active claims and the toy repository state. It walks the claims, records the first owner of each path, and emits SAMEPATHCLAIM_CONFLICT when a second active claim wants a path that is already owned. It compares each claim's expectedparentsha against the current parent in the repository state and emits EXPECTEDPARENTMISMATCH when they differ. A claim with no owned path emits MISSINGOWNEDPATH.",
"level": 4,
"order": 10,
"section_id": "validate-claim-preflight-reads-the-active-claims-and-the-toy-repository-state",
"source_line_end": 82,
"source_line_start": 73,
"title": "`validate_claim_preflight` reads the active claims and the toy repository state"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatecheckpointlanepolicy reads a set of commit-lane cases. For each case it computes the recommended lane and compares it to the selected lane.** validatecheckpointlane_policy reads a set of commit-lane cases. For each case it computes the recommended lane and compares it to the selected lane. A broadcheckpoint lane without operatorauthorizedbroadcheckpoint set to true emits CHECKPOINTBROADAUTH_REQUIRED. A suspected secret not routed to the hard-stop lane emits CHECKPOINTSECRETREQUIRESHARDSTOP. A dirty tree with isolated owned paths must not block a scoped commit, so a case that treats it as a scoped blocker emits CHECKPOINTDIRTYTREENOTSCOPED_BLOCKER. A case with no selected lane emits\u2026",
"level": 4,
"order": 11,
"section_id": "validate-checkpoint-lane-policy-reads-a-set-of-commit-lane-cases",
"source_line_end": 94,
"source_line_start": 83,
"title": "`validate_checkpoint_lane_policy` reads a set of commit-lane cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validaterealactiveclaimssnapshot is the discriminating check. validaterealactiveclaimssnapshot is the discriminating check. It imports the copied workledgerruntime from the exported source modules, calls runtimestatus, and requires the declared sourcesnapshot_hash to equal the recomputed source_hash (a 64-character hex digest, not all zeros). It then calls activeclaimcollisionsforpaths three times: on the real status, on a status mutated to add a same-path conflict, and on a status mutated to add a disjoint-path claim. The real status must return no collision for the requested path, the same-path mutation must return one, and the disjoint mutation must return none. It also mutates the expec\u2026",
"level": 4,
"order": 12,
"section_id": "validate-real-active-claims-snapshot-is-the-discriminating-check",
"source_line_end": 113,
"source_line_start": 95,
"title": "`validate_real_active_claims_snapshot` is the discriminating check"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "runmissiontransactionbundle validates copied control source. The four import validators (validatetaskledgersourceimport, validateworkledgersourceimport, validatecheckpointsourceimport, validatemissioncontrolsourceimport) each read a source-module manifest and a runtime contract, check that the manifest declares exactly the expected module ids and count, then for each module confirm the copied file exists, contains its required internal control anchor strings, matches the manifest sha256, and carries classification: copiednonsecretmacrobody with bodyinreceipt: false. runmissiontransaction_bundle validates copied control source. The four import validators (validatetaskledgersourceimport, vali\u2026",
"level": 4,
"order": 13,
"section_id": "run-mission-transaction-bundle-validates-copied-non-secret-control-source",
"source_line_end": 126,
"source_line_start": 114,
"title": "`run_mission_transaction_bundle` validates copied control source"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "writereceipts writes the nine result records for a fixture run (preflight, dependency-blocked, work-landing attempt, claim preflight, scoped mutation, checkpoint lane, completion projection, dependency-unlock scheduler, reconcile plan).** write_receipts writes the nine result records for a fixture run (preflight, dependency-blocked, work-landing attempt, claim preflight, scoped mutation, checkpoint lane, completion projection, dependency-unlock scheduler, reconcile plan). Each record carries a schema version, refs, counts, hashes, verdicts, and bodyinreceipt: false. Source bodies live only in the exported bundle's source_modules/ tree; they never enter a result record.",
"level": 4,
"order": 14,
"section_id": "nine-body-free-transaction-records",
"source_line_end": 153,
"source_line_start": 127,
"title": "Nine metadata-only transaction records"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "negative-cases",
"source_line_end": 155,
"source_line_start": 154,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The positive claim is not that the fixture passes. It is that the fixture accepts good evidence and rejects targeted one-field perturbations. The positive claim is not that the fixture passes. It is that the fixture accepts good evidence and rejects targeted one-field perturbations. Fourteen named negative cases are required, each bound to its error code:",
"level": 4,
"order": 16,
"section_id": "the-positive-claim-is-not-that-the-fixture-passes",
"source_line_end": 162,
"source_line_start": 156,
"title": "The positive claim is not that the fixture passes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "competingclaimandstaleparent must observe both SAMEPATHCLAIMCONFLICT and EXPECTEDPARENTMISMATCH. competingclaimandstaleparent must observe both SAMEPATHCLAIM_CONFLICT and EXPECTEDPARENTMISMATCH. realworkledgersamepathclaimconflict, realworkledgerexpectedparent_mismatch, and plus a broad-checkpoint landing-row mutation that must observe CHECKPOINTBROADAUTHREQUIRED. missionclaimmissingowned_path observes MISSINGOWNEDPATH. scopedcommitreceiptclaimsglobal_authority observes SCOPEDRECEIPTAUTHORITYUPGRADE. missionfixtureprivatetaskledgerbody observes LIVETASKLEDGERBODYIN_FIXTURE. cleanpreflightoverclaimslandingcomplete observes PREFLIGHTPASSOVERCLAIMSWORKLANDED. dependencyunlockwithoutresolutionr\u2026",
"level": 4,
"order": 17,
"section_id": "competing-claim-and-stale-parent-must-observe-both-same-path-claim-conflict-and-expected-parent-mismatch",
"source_line_end": 184,
"source_line_start": 163,
"title": "`competing_claim_and_stale_parent` must observe both `SAME_PATH_CLAIM_CONFLICT` and `EXPECTED_PARENT_MISMATCH`"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "prior-art-grounding",
"source_line_end": 186,
"source_line_start": 185,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "This is the mission-transaction member of a small family of work-landing fixtures. Its closest sibling is durableagentworklandingreplay, which checks recorded landing rows, validation-before-commit ordering, HEAD movement, blocker capture, and completion evidence without running live git. This is the mission-transaction member of a small family of work-landing fixtures. Its closest sibling is durableagentworklandingreplay, which checks recorded landing rows, validation-before-commit ordering, HEAD movement, blocker capture, and completion evidence without running live git. This module narrows that pattern to the transaction preflight and the work-ledger claims membrane: same-path conflicts,\u2026",
"level": 4,
"order": 19,
"section_id": "this-is-the-mission-transaction-member-of-a-small-family-of-work-landing-fixtures",
"source_line_end": 200,
"source_line_start": 187,
"title": "This is the mission-transaction member of a small family of work-landing fixtures"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "To validate the exported source bundle: For the focused regression tests and the paper-module corpus check:",
"level": 2,
"order": 20,
"section_id": "validation-receipt-path",
"source_line_end": 226,
"source_line_start": 201,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means the fixture rows, the copied source bodies, the real snapshot replay, and the fourteen negative cases all hold together, and the result records carry no private body. A pass means the fixture rows, the copied source bodies, the real snapshot replay, and the fourteen negative cases all hold together, and the result records carry no private body.",
"level": 4,
"order": 21,
"section_id": "a-pass-means-the-fixture-rows",
"source_line_end": 233,
"source_line_start": 227,
"title": "A pass means the fixture rows"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 235,
"source_line_start": 234,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim the evidence supports: the public fixture rows, the copied control source bodies, the real work-ledger session-snapshot replay, the discriminating negative cases, and the metadata-only result records preserve the work-landing contract at the R3 rung, and the checks that discriminate still discriminate. The strongest claim the evidence supports: the public fixture rows, the copied control source bodies, the real work-ledger session-snapshot replay, the discriminating negative cases, and the metadata-only result records preserve the work-landing contract at the R3 rung, and the checks that discriminate still discriminate. That is a replay-and-evidence-shape claim. The proo\u2026",
"level": 4,
"order": 23,
"section_id": "the-r3-replay-and-evidence-shape-claim",
"source_line_end": 245,
"source_line_start": 236,
"title": "The `R3` replay-and-evidence-shape claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Within that scope limit the module refuses more. It does not mutate the live work log, the live work log, or git. Within that scope limit the module refuses more. It does not mutate the live work log, the live work log, or git. It does not certify completion of the current repository. It grants no authority to commit, checkpoint broadly, back up, or launch, and it makes no claim about provider behaviour, browser state, hosted-product readiness, or whole-system correctness.",
"level": 4,
"order": 24,
"section_id": "within-that-authority-ceiling-the-module-refuses-more",
"source_line_end": 254,
"source_line_start": 246,
"title": "Within that scope limit the module refuses more"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/mission_transaction_work_spine.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public mission-transaction fixture validates work-landing, claim, dependency, scoped-commit, checkpoint-lane, result record-drain, completion, and copied control source-module contracts without mutating live ledgers or git."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "mission_transaction_work_spine.py",
"path": "src/microcosm_core/organs/mission_transaction_work_spine.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public mission-transaction fixture validation, copied source-module manifest checks, negative-case enforcement, and metadata-only result records.",
"symbols": [
"run",
"run_mission_transaction_bundle",
"validate_task_ledger_source_import",
"validate_work_ledger_source_import",
"validate_checkpoint_source_import",
"validate_mission_control_source_import",
"validate_claim_preflight",
"validate_checkpoint_lane_policy"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/mission_transaction_work_spine.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/mission_transaction_work_spine.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/mission_transaction_work_spine.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Mission Transaction Work Spine is the public replay membrane for Microcosm work-landing discipline. It checks fixed Work item, claim, dependency, transaction, result record-drain, completion, scoped mutation, and checkpoint-lane rows; validates copied work log, work log, checkpoint, scoped-commit, and mission-preflight source modules by manifest; and writes metadata-only result records with secret-exclusion and scope limits.",
"summary_line": "The public mission-transaction fixture validates work-landing, claim, dependency, scoped-commit, checkpoint-lane, result record-drain, completion, and copied control source-module contracts without mutating live ledgers or git.",
"title": "Mission Transaction Work Spine"
},
{
"anchor": "paper-module-formal-math-readiness-gate",
"authority_ceiling": "Public readiness metadata, copied PROVER smoke-run readiness/probe artifacts, public component source body floor, fixture result records, and exported-bundle result records only; no Lean/Lake execution, Mathlib availability beyond probe status, formal proof authority, formal-result correctness, external model access, private proof body, oracle premise id, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
Teleology
\n
From deferred slogan to executable boundary
\n
This is the public runtime cell that turns the formal-math slice into an executable boundary, validating synthetic readiness metadata for corpus, tactics, premises, routing, and provider recipes before any Lean witness claims authority.formal_math_readiness_gate is the public runtime cell that turns the formal math slice from a deferred slogan into an executable boundary. It validates synthetic readiness metadata for corpus availability, tactic probes, premise indexes, target-shape routing, and provider context recipes before any future Lean witness can claim authority.
\n
One question for a cold reader
\n
The page should let a cold reader answer one question without rereading the component: what evidence has Microcosm validated, and where does that evidence stop? The page should let a cold reader answer one question without rereading the component: what evidence has Microcosm actually validated, and where does that evidence stop?
\n
Purpose
\n
Making the ready-versus-proven boundary explicit
\n
Formal-math tooling fails quietly when a library or corpus is assumed rather than checked, so this component makes the ready-versus-proven boundary explicit by asking which declared inputs a later proof witness could safely consume. Formal-math tooling fails quietly when a library, tactic, or corpus is assumed present rather than checked. A pipeline that routes a proof to aesop when aesop is not actually available, or that treats a premise index as proof evidence because it happens to carry a proof body, has already lost the boundary between "ready to attempt" and "proven". This component exists to make that boundary explicit before any downstream proof work begins. It answers one question: which declared formal-math inputs are well-formed and honest enough that a later proof witness could safely consume them, and where exactly does that warrant stop?
\n
A reducer that refuses specific lies
\n
The mechanism is a deterministic reducer over five public JSON inputs that runs no Lean and instead refuses the specific ways they can lie, emitting a readiness board rather than theorem evidence. The mechanism is a deterministic reducer over five public JSON inputs: corpus readiness, tactic-portfolio availability, a premise index, target-shape routing, and provider context recipes. It does not run Lean or Lake. Instead it reads what those inputs declare and refuses the specific ways they can lie. A corpus that claims Mathlib is available without a passing probe is rejected. A tactic marked available without a probe result record is rejected. A premise row carrying a proof_body or oracle_needed_premise_ids field is rejected. A route that admits a tactic the portfolio probe already marked unavailable is rejected. The output is a readiness board, not theorem evidence.
\n
Proving discipline through negative cases
\n
The gate proves its own discipline through five overclaim inputs that each run must catch, while a digest manifest keeps copied probe bodies source-open without leaking proof bodies, model-output data, or private state. The design choice worth noticing is that the gate proves its own discipline through negative cases. Alongside the positive inputs, the fixture carries five inputs that each commit a known overclaim, and the run passes only when every one of those overclaims is caught and no unexpected finding appears. The gate is therefore not merely asserting "we check Mathlib availability"; it is demonstrating, on each run, that a falsified Mathlib claim is actually refused. A second guard keeps the floor source-open without leaking: copied prover probe bodies are verified by digest through a manifest, while proof bodies, model-output data, and private state stay out of the result records entirely.
\n
Shape
\nDiagram of the mechanism (11 steps).Diagram source & refs
Source refs
reject Mathlib-availability overclaim
validate_corpus_readiness
each available tactic needs a probe result record
validate_tactic_portfolio
reject route admitting an unavailable tactic
validate_target_shape_routing
reject over-budget or proof-body recipe
validate_provider_context_recipes
copied probe bodies, digest-checked
validate_source_module_imports
flowchart TD Inputs["Five public JSON inputs corpus, tactics, premises, routes, provider recipes"] Scan["Secret-exclusion scan zero blocking hits required"] Corpus["validate_corpus_readiness reject Mathlib-availability overclaim"] Tactics["validate_tactic_portfolio each available tactic needs a probe result record"] Premises["validate_premise_index reject proof_body / oracle premise ids"] Routing["validate_target_shape_routing reject route admitting an unavailable tactic"] Provider["validate_provider_context_recipes reject over-budget or proof-body recipe"] SourceFloor["validate_source_module_imports copied probe bodies, digest-checked"] Reconcile["Reconcile findings vs EXPECTED_NEGATIVE_CASES every known overclaim must be caught"] Board["Readiness board + extension board available / blocked capabilities, counts"] Ceiling["Scope limit no Lean/Lake, proof, provider, launch, or private-system authority"] Inputs --> Scan Scan --> Corpus Scan --> Tactics Scan --> Premises Scan --> Provider Tactics -->|unavailable tactic ids| Routing Corpus --> Reconcile Tactics --> Reconcile Premises --> Reconcile Routing --> Reconcile Provider --> Reconcile SourceFloor --> Reconcile Reconcile --> Board Board --> Ceiling
\n
The hand-authored diagram is not the machine graph
\n
The machine graph remains the generated paper_module.formal_math_readiness_gate.mermaid projection derived from the source record, not from this hand-authored Mermaid block.
\n
Reader Evidence Routing
\n
Read this module in evidence order:
\n
Start at core/paper_module_capsules.json::paper_modules[21:paper_module.formal_math_readiness_gate]. That row names the source authority, subjects, mechanism refs, code locus, Microcosm concept/principle/axiom refs, generated projection statuses, and the bundle scope limit.
Check the generated structured source record paper_modules/formal_math_readiness_gate.json. Its relationships.edges cite the bundle source refs and show the generated Mermaid status, Atlas status, source_authority: json_capsule, and unresolved selective-relation count.
Inspect the runtime locus src/microcosm_core/organs/formal_math_readiness_gate.py, especially run, run_readiness_bundle, validate_source_module_imports, write_receipts, EXPECTED_NEGATIVE_CASES, AUTHORITY_CEILING, and SOURCE_MODULE_MANIFEST_NAME.
Use fixture evidence for the gate behavior: fixtures/first_wave/formal_math_readiness_gate/input, receipts/first_wave/formal_math_readiness_gate/readiness_gate_result.json, formal_math_readiness_board.json, formal_math_readiness_extension_board.json, formal_math_readiness_validation_receipt.json, and result records/sign-off/first_wave/formal_math_readiness_gate_fixture_acceptance.json.
Use exported-bundle evidence for source-open body-floor claims: examples/formal_math_readiness_gate/exported_formal_math_readiness_bundle/source_module_manifest.json, bundle_manifest.json, source_artifacts/, source_body_floor/source_modules/, and receipts/runtime_shell/demo_project/organs/formal_math_readiness_gate/exported_formal_math_readiness_bundle_validation_result.json.
Use tests/test_formal_math_readiness_gate.py for the behavioral result record boundary. The tests cover negative cases, exported bundle sign-off, source-module digest and target-ref mismatch rejection, bounded command-card output, source-body omission from result records, secret-exclusion/public-relative result record paths, and non-writing plan preview.
\n
This page routes readiness evidence, bounded evidence claims
\n
A proof claim must go elsewhere because this page routes only readiness evidence, result record integrity, and source-body-floor accounting. Do not route a proof claim through this page. It routes readiness evidence, result record integrity, and source-body-floor accounting only.
\n
Technical Mechanism
\n
One reducer, two entrypoints
\n
run() evaluates the fixture with negative cases and run_readiness_bundle() evaluates the exported bundle, both calling _build_result() so they share one scope limit, secret scan, digest checker, and board schema. The runtime is a deterministic readiness reducer over declared public inputs. run() evaluates the first-wave fixture directory with positive and negative JSON cases enabled; run_readiness_bundle() evaluates the exported public bundle without fixture-negative cases and requires the bundle source-module manifest. Both entrypoints call _build_result(), so the fixture and exported bundle result records share one scope limit, one secret scan, one source-module digest checker, and one readiness-board schema.
\n
Loading inputs and the non-advisory scan
\n
_build_result() loads the five public input families and scans them for forbidden material, and the result passes only with zero blocking hits, passing source-module imports, all negatives observed, and no stray positive findings._build_result() first loads the five public input families: corpus_readiness.json, tactic_portfolio_availability.json, premise_index.json, target_shape_tactic_routing.json, and provider_context_recipes.json. It then scans those inputs plus any declared source artifacts through secret_exclusion_scan.scan_paths, using the public Microcosm forbidden-class policy. The scan is not advisory: the result can pass only when the scan has zero blocking hits, source-module imports pass, all expected fixture-negative cases are observed, and no unexpected positive-case findings remain.
\n
The mechanism is split into six validators:
\n
validate_corpus_readiness() records Lean and Mathlib readiness metadata and adds lean_std_synthetic_core:mathlib to blocked capabilities when Mathlib is unavailable. A Mathlib availability claim without a passing probe becomes MATHLIB_AVAILABILITY_OVERCLAIM.
validate_tactic_portfolio() separates available from unavailable tactics and requires every available tactic to carry a probe result record. Synthetic probe labels are accepted only when _tactic_probe_realness_evidence() binds them to copied source modules or fixture-manifest source-open evidence.
validate_premise_index() admits premise rows as metadata only. It counts premises, namespaces, retrieval terms, and split eligibility, but rejects proof_body, ground_truth_proof, provider_output_body, and oracle_needed_premise_ids.
validate_target_shape_routing() intersects each route case's allowed tactics with the unavailable tactics emitted by the portfolio validator. Any overlap becomes ROUTING_ALLOWS_UNAVAILABLE_TACTIC, so routing cannot silently re-enable a tactic that the probe plane blocked.
validate_provider_context_recipes() records byte budgets and deliverable shape while rejecting public recipes over 32,768 bytes or recipes that allow proof bodies or provider-body material.
validate_source_module_imports() verifies the exported bundle's source_module_manifest.json, target refs, source refs, line counts, target digests, source digests, exact-copy rows, and the two permitted private-path rewrites. It reports digest/ref failures without placing copied source bodies in result records.
\n
Reconciling findings and projecting the board
\n
_merge_observed() and _merge_findings() compare observed failures against EXPECTED_NEGATIVE_CASES, then _build_extension_board() projects accepted metadata into pattern ids, counts, blocked routes, budgets, and the scope limit. After the validators run, _merge_observed() and _merge_findings() compare observed fixture failures against EXPECTED_NEGATIVE_CASES. This is the local scope limit: the fixture run must prove that the known overclaims are caught, while the exported-bundle run must prove that the positive public bundle has no unexpected findings. _build_extension_board() then projects the accepted metadata into the extension board: selected pattern ids, namespace and split counts, tactic availability counts, Mathlib-dependent unavailable tactics, blocked route cases, provider budgets, source-body import counts, the scope limit, and the scope boundary.
\n
Result records and the asserting test suite
\n
write_receipts() emits the gate result and boards while preserving the boundary, and the focused test asserts the mechanism itself, from negative-case ids to digest-mismatch rejection and result record omission of copied bodies. Result record writing preserves the same boundary. write_receipts() emits the gate result, readiness board, extension board, validation result record, and sign-off result record for fixture mode. run_readiness_bundle() emits the exported-bundle result record. The focused test suite asserts the mechanism rather than just file existence: it checks the five expected negative case ids, local Lean/Lake probe metadata with Mathlib unavailable, six available tactics with aesop blocked, eleven premises, five route cases, three provider recipes, thirteen verified source artifacts, source/target digest mismatch rejection, target-ref mismatch rejection, secret-exclusion/public-relative result record paths, and result record omission of copied body text.
\n
Public Contract
\n
What the component consumes and rejects
\n
The component runs no Lean or Lake; it records available and blocked capabilities and rejects Mathlib overclaims, unprobed tactics, proof-body premise rows, unavailable-tactic routes, and over-budget or proof-body recipes. The component does not run Lean or Lake. It consumes public JSON fixtures and exported bundles, records which capabilities are available or blocked, rejects Mathlib availability overclaims, rejects unprobed tactics, rejects premise rows that contain proof bodies, rejects routes that admit unavailable tactics, and rejects provider recipes that exceed the public budget or allow proof bodies.
\n
The result is a readiness board
\n
The accepted result is a readiness board that can tell a later component what is safe to attempt, but it is bounded evidence evidence, benchmark evidence, or permission to run a prover. The accepted result is a readiness board. That board can tell a later component what is safe to attempt, but it is bounded evidence evidence, benchmark evidence, or permission to execute a theorem prover.
\n
Prior Art Grounding
\n
Formal-math readiness and environment work
\n
The component is grounded in formal-math benchmark and environment-readiness work where a library or corpus being present is not enough, drawing on miniF2F, LeanDojo, and mathlib. This component is grounded in formal-math benchmark and environment-readiness work where the presence of a library, tactic, or corpus is not enough by itself. miniF2F motivates explicit benchmark split discipline for formal mathematics, LeanDojo motivates reproducible theorem-proving environments, and mathlib makes the availability of library imports a concrete precondition rather than a vague capability claim.
\n
Borrowing the readiness-gate pattern
\n
Microcosm borrows the readiness-gate pattern, requiring corpus availability, Mathlib and tactic probes, premise indexes, routing, and context budgets to be checked before downstream proof language, without authorizing Lean execution. Microcosm borrows the readiness-gate pattern: corpus availability, Mathlib probes, tactic probes, premise indexes, target-shape routing, and context budgets must be checked before downstream proof or retrieval language is allowed. It excludes Lean execution or proof authority.
\n
Runtime Surfaces
\n
python -m microcosm_core.organs.formal_math_readiness_gate run --input fixtures/first_wave/formal_math_readiness_gate/input --out receipts/first_wave/formal_math_readiness_gate
python -m microcosm_core.organs.formal_math_readiness_gate plan --input fixtures/first_wave/formal_math_readiness_gate/input
microcosm formal-math-readiness-gate run --input fixtures/first_wave/formal_math_readiness_gate/input --out receipts/first_wave/formal_math_readiness_gate
microcosm formal-math-readiness-gate plan --input fixtures/first_wave/formal_math_readiness_gate/input
\n
Relationship To Lean Witness
\n
The deferral is typed and testable
\n
formal_math_lean_proof_witness remains deferred, and this gate makes the deferral typed and testable: absent Mathlib, unroutable unavailable tactics, no proof or oracle bodies in premises, and no proof-body provider recipes.formal_math_lean_proof_witness remains deferred. This gate makes the deferral typed and testable: Mathlib is absent until a passing probe says otherwise, unavailable tactics cannot be routed, premise indexes cannot carry proof or oracle bodies, and provider recipes cannot smuggle proof-body deliverables.
\n
Validation Result record Path
\n
Three read-only checks validate body and projection
The generated row should expose fifteen resolved edges
\n
A correct row reports fifteen edges, available Mermaid, owner-blocked Atlas, JSON-bundle authority, and no unresolved selective relation. Expected generated-row proof: edge_count: 15, mermaid_status: available_from_capsule_edges, atlas_status: blocked_until_organ_atlas_owner_lane_binds_edges, source_authority: json_capsule, and unresolved_selective_relation_count: 0.
\n
Scope boundary
\n
Scope limit
\n
The bounded valid claim
\n
The module may claim a public readiness gate for formal-math preparation, bounded to corpus availability, probe metadata, premise coverage, routing, budgets, pattern ids, source artifacts, an component-source floor, and result records. This module may claim that Microcosm has a public readiness gate for formal math system preparation. The valid claim is bounded to corpus availability, Mathlib and tactic probe metadata, premise-index coverage, target-shape tactic routing, provider context budget checks, extension-board pattern ids, public PROVER smoke-run source artifacts, an exact public component-source body floor, and fixture or exported-bundle result records.
\n
What the module must not claim
\n
It must not claim Lean/Lake execution, theorem proving, proof authority, benchmark performance, external model access, private proof-body import, source-file changes, or launch, and its strongest statement is readiness-boundary enforcement. The module must not claim Lean/Lake execution, theorem proving, formal proof authority, formal-result correctness, Mathlib-dependent proof success, benchmark performance, provider-call execution, private proof-body import, oracle-needed premise disclosure, source-file changes, publishing-scope decision, hosted deployment, recipient work, secret export, or whole-system correctness. Its strongest launch-facing statement is readiness-boundary enforcement over public metadata and copied source artifacts.
\n
Limitations
\n
Finite fixtures, no execution
\n
The runtime validates finite public fixtures without executing Lean or Lake, importing Mathlib, or calling a provider, so a blocked capability is a downstream boundary, not an invitation to route around the gate. The runtime validates finite public fixtures and exported-bundle manifests. It does not execute Lean or Lake, import Mathlib in the current environment, call a provider, or check theorem statements. When the result reports blocked capabilities such as lean_std_synthetic_core:mathlib, that is a readiness boundary for downstream components, not an invitation to route around the gate.
\n
Source artifacts are body-floor evidence only
\n
The copied source artifacts are source-open body-floor evidence only: digest and target-ref checks show they match their manifests but authorize no source-file changes, proof-body disclosure, deployment, or public sharing. The copied source artifacts are source-open body-floor evidence only. Digest and target-ref checks show that selected PROVER readiness/probe bodies and the public component source copy match their manifests; they do not authorize source-file changes, private source-root export, proof-body disclosure, recipient work, hosted deployment, or public sharing. Result records intentionally carry counts, digests, paths, negative-case coverage, and authority flags instead of copied body text.
\n
The negative cases are finite, covering the known overclaims in EXPECTED_NEGATIVE_CASES, so a new formal-math claim needs a new source-backed negative case or a different proof consumer, not this page. The negative cases are also finite. They cover the known overclaims encoded in EXPECTED_NEGATIVE_CASES: Mathlib availability without a passing probe, unprobed tactic availability, premise rows with proof bodies, target routes that admit unavailable tactics, and provider recipes that exceed public budgets or permit proof bodies. A new formal-math claim needs either a new source-backed negative case here or a different proof consumer; this page should not be used as a generic formal-proof claim surface.
\n
Scope boundary
\n
What this module excludes
\n
This module documents a public readiness gate only, authorizing no Lean/Lake execution, proof authority, external model access, launch, or deployment, and adding no private source-root, account, or proof-body material to the public floor. This module documents a public readiness gate only. It excludes Lean/Lake execution, formal proof authority, Mathlib-dependent proof attempts, external model access, benchmark claims, public launch, hosted deployment, public sharing, recipient work, secret export, or whole-system correctness. It also does not make private source-root material, browser UI state, account or browser material, browser state, account secrets, source notes, model-output data bodies, recipient-send state, or private proof bodies part of the public Microcosm body floor.
\n\nSource and projection details\n
Source-Open Body Floor
\n
Thirteen probe bodies preserve the readiness floor
\n
The exported bundle carries thirteen probe bodies plus two digest-bound path rewrites covering corpus, tactic, Mathlib, trace, and Lean portfolio readiness. The exported readiness bundle carries thirteen PROVER smoke-run readiness/probe bodies under source_artifacts. They cover corpus readiness, tactic-affordance probe metadata, Mathlib and trace probes, and the copied portfolio-core Lean probes used to decide which tactics are blocked or available. Two JSON rows are private-path rewrites; those rows retain source and target digests plus the rewrite mode.
\n
An exact component copy remains separate from result records
\n
The public component-source copy is inspectable as body-floor evidence, while result records exclude copied text, private proof material, provider bodies, live session state, account secrets, and send state. The bundle also carries an exact public component-source copy for src/microcosm_core/organs/formal_math_readiness_gate.py under source_body_floor/source_modules. Generated state/runs Lean artifacts are runnable readiness evidence, not source-body authority. Neither floor places body text in result records or workingness cards, and neither imports model-output data bodies, account or browser state, browser UI live access, recipient-send state, account secrets, private proof bodies, or oracle-needed premise ids.
\n
Manifests carry bodies while result records carry references
\n
Body-floor inspection belongs in the source and bundle manifests; validation result records deliberately retain only status, digests, counts, and public-relative references. The source-module manifest and bundle manifest are the right surfaces for body-floor inspection. The validation result records intentionally carry status, digests, counts, and public-relative refs rather than copied source bodies.
\n
The extension board projects typed readiness metadata
\n
Wave 011's extension board records selected patterns, intake and validation refs, premise coverage, tactic and Mathlib status, route admissibility, and provider budgets without becoming proof evidence. Wave 011 adds the explicit extension board for the source intake cell formal_math_readiness_extensions. The board is still metadata-only, but it is more useful than the older flat counts: it records the selected pattern ids (lean_std_toolchain_premise_index, tactic_portfolio_availability_probe, target_shape_tactic_routing_gate), the source projection intake ref, public target refs, validation refs, namespace and split coverage for the premise index, tactic availability status counts, Mathlib-dependent unavailable tactics, target-shape routing admissibility, and provider context budgets.
\n
Governing Lattice Relation
\n
Two mechanisms separate positive readiness from refusal
\n
The bundle's concept and two mechanisms bind the positive bundle-validation path separately from the validators that reject readiness-to-proof overclaims. The bundle binds this module to concept.formal_math_and_proof_witness_bundle because the component is not a theorem prover; it is the membrane that decides which public formal-math inputs are safe enough for a later proof witness to consume. The governing mechanisms split that membrane in two. The validates_public_formal_math_readiness_bundle mechanism names the positive bundle path: run, run_readiness_bundle, validate_source_module_imports, and write_receipts validate the declared corpus, tactic, premise, routing, provider-budget, source-module-manifest, and source-body-floor evidence before writing readiness boards. The validates_public_readiness_boundary mechanism names the negative path: validate_corpus_readiness, validate_tactic_portfolio, validate_premise_index, validate_target_shape_routing, and validate_provider_context_recipes reject the cases that would turn readiness metadata into proof authority.
\n
Principles and axioms enforce authority separation
\n
The cited principles separate bundle, structured source record, runtime, result record, and body classes, while the axioms keep private material excluded and the downstream proof witness deferred. The principle and axiom refs are therefore operational, not decorative. P-1, P-2, and P-3 are expressed by keeping the JSON bundle, generated structured source record, runtime code locus, and result records as separate authority classes. P-6 and P-8 are expressed by the body-floor and secret-exclusion contracts: copied PROVER probe bodies and the public component source copy can be inspected through digests and manifests, while private proof bodies, model-output data bodies, and browser or account state stay outside the public floor. AX-1, AX-2, AX-5, and AX-7 are the local reason the downstream paper_module.formal_math_lean_proof_witness remains a dependency rather than an already-proven conclusion.
\n
Fifteen generated edges prove connection, not capability
\n
The deliberately small edge set establishes bundle, source, and dependency bindings without proving witness execution, Mathlib availability, or a fully bound Atlas card. The generated lattice edge count is small on purpose: it proves that this page is bundle-backed, source-bound, and connected to one deferred proof-witness module.
\n",
"components": [
{
"anchor": "component-formal_math_readiness_gate",
"component_id": "formal_math_readiness_gate",
"public_label": "Formal Math Readiness Gate"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 2,
"organ": 1,
"paper_module": 1,
"principle": 5
},
"link_count": 14,
"links": [
{
"id": "formal_math_readiness_gate",
"kind": "organ",
"kind_label": "component",
"label": "Formal Math Readiness Gate",
"relation": "explains",
"url": "component-formal-math-readiness-gate.html"
},
{
"id": "mechanism.formal_math_readiness_gate.validates_public_formal_math_readiness_bundle",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public formal math readiness bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-readiness-gate-validates-public-formal-math-readiness-bundle"
},
{
"id": "mechanism.formal_math_readiness_gate.validates_public_readiness_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public readiness boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-readiness-gate-validates-public-readiness-boundary"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.formal_math_lean_proof_witness",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Lean Proof Witness",
"relation": "depends on",
"url": "paper-module-formal-math-lean-proof-witness.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch4_proof_authority_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 4 Proof, Authority, and Runtime Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[74:paper_module.batch4_proof_authority_runtime].depends_on",
"url": "paper-module-batch4-proof-authority-runtime.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[19:paper_module.macro_projection_import_protocol].depends_on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.provider_context_recipe_budget",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Provider Context Recipe Budget",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[54:paper_module.provider_context_recipe_budget].depends_on",
"url": "paper-module-provider-context-recipe-budget.html"
},
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[40:paper_module.tactic_portfolio_availability].depends_on",
"url": "paper-module-tactic-portfolio-availability.html"
},
{
"id": "paper_module.target_shape_tactic_routing",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Target Shape Tactic Routing",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[41:paper_module.target_shape_tactic_routing].depends_on",
"url": "paper-module-target-shape-tactic-routing.html"
}
],
"paper_module_id": "paper_module.formal_math_readiness_gate",
"reader_packet": {
"abstract": "The public formal-math readiness gate validates declared corpus, tactic, premise, routing, provider-budget, source-module manifest, copied PROVER probe body, and negative-case boundaries without claiming Lean/Lake or proof authority.",
"authority_ceiling": "Public readiness metadata, copied PROVER smoke-run readiness/probe artifacts, public component source body floor, fixture result records, and exported-bundle result records only; no Lean/Lake execution, Mathlib availability beyond probe status, formal proof authority, formal-result correctness, external model access, private proof body, oracle premise id, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.formal_math_readiness_gate.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-formal-math-readiness-gate-diagram-1",
"html_url": "docs/paper-module-formal-math-readiness-gate.html#paper-module-formal-math-readiness-gate-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Inputs[\"Five public JSON inputs corpus, tactics, premises, routes, provider recipes\"] Scan[\"Secret-exclusion scan zero blocking hits required\"] Corpus[\"validate_corpus_readiness reject Mathlib-availability overclaim\"] Tactics[\"validate_tactic_portfolio each available tactic needs a probe result record\"] Premises[\"validate_premise_index reject proof_body / oracle premise ids\"] Routing[\"validate_target_shape_routing reject route admitting an unavailable tactic\"] Provider[\"validate_provider_context_recipes reject over-budget or proof-body recipe\"] SourceFloor[\"validate_source_module_imports copied probe bodies, digest-checked\"] Reconcile[\"Reconcile findings vs EXPECTED_NEGATIVE_CASES every known overclaim must be caught\"] Board[\"Readiness board + extension board available / blocked capabilities, counts\"] Ceiling[\"Scope limit no Lean/Lake, proof, provider, launch, or private-system authority\"] Inputs --> Scan Scan --> Corpus Scan --> Tactics Scan --> Premises Scan --> Provider Tactics -->|unavailable tactic ids| Routing Corpus --> Reconcile Tactics --> Reconcile Premises --> Reconcile Routing --> Reconcile Provider --> Reconcile SourceFloor --> Reconcile Reconcile --> Board Board --> Ceiling",
"source_format": "mermaid",
"source_hash": "sha256:80af348c44dc0b502c9b53c0c0370728f18d18af44fee689e95cdb769bab53fc",
"source_line_end": 120,
"source_line_start": 92
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.formal_math_lean_proof_witness",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Lean Proof Witness",
"relation": "depends on",
"url": "paper-module-formal-math-lean-proof-witness.html"
}
],
"relation": "depends on"
},
{
"count": 3,
"links": [
{
"id": "formal_math_readiness_gate",
"kind": "organ",
"kind_label": "component",
"label": "Formal Math Readiness Gate",
"relation": "explains",
"url": "component-formal-math-readiness-gate.html"
},
{
"id": "mechanism.formal_math_readiness_gate.validates_public_formal_math_readiness_bundle",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public formal math readiness bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-readiness-gate-validates-public-formal-math-readiness-bundle"
},
{
"id": "mechanism.formal_math_readiness_gate.validates_public_readiness_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public readiness boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-readiness-gate-validates-public-readiness-boundary"
}
],
"relation": "explains"
},
{
"count": 6,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.formal_math_readiness_gate",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 69,
"title": "JSON Bundle Binding"
},
{
"reason": "public_projection_internal_section_denylist",
"section_id": "structured-lattice-bindings",
"source_line_start": 361,
"title": "Structured Lattice Bindings"
},
{
"reason": "public_projection_internal_section_denylist",
"section_id": "receipt-expectations",
"source_line_start": 505,
"title": "Result record Expectations"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_math_readiness_gate.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "miniF2F",
"section_id": "formal-math-readiness-and-environment-work",
"url": "https://arxiv.org/abs/2109.00110"
},
{
"kind": "external_reference",
"label": "LeanDojo",
"section_id": "formal-math-readiness-and-environment-work",
"url": "https://arxiv.org/abs/2306.15626"
},
{
"kind": "external_reference",
"label": "mathlib",
"section_id": "formal-math-readiness-and-environment-work",
"url": "https://arxiv.org/abs/1910.09336"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 1,
"section_id": "teleology",
"source_line_end": 4,
"source_line_start": 3,
"title": "Teleology"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This is the public runtime cell that turns the formal-math slice into an executable boundary, validating synthetic readiness metadata for corpus, tactics, premises, routing, and provider recipes before any Lean witness claims authority.** formalmathreadiness_gate is the public runtime cell that turns the formal math slice from a deferred slogan into an executable boundary. It validates synthetic readiness metadata for corpus availability, tactic probes, premise indexes, target-shape routing, and provider context recipes before any future Lean witness can claim authority.",
"level": 4,
"order": 2,
"section_id": "from-deferred-slogan-to-executable-boundary",
"source_line_end": 15,
"source_line_start": 5,
"title": "From deferred slogan to executable boundary"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The page should let a cold reader answer one question without rereading the component: what evidence has Microcosm validated, and where does that evidence stop?** The page should let a cold reader answer one question without rereading the component: what evidence has Microcosm actually validated, and where does that evidence stop?",
"level": 4,
"order": 3,
"section_id": "one-question-for-a-cold-reader",
"source_line_end": 22,
"source_line_start": 16,
"title": "One question for a cold reader"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 24,
"source_line_start": 23,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Formal-math tooling fails quietly when a library or corpus is assumed rather than checked, so this component makes the ready-versus-proven boundary explicit by asking which declared inputs a later proof witness could safely consume.** Formal-math tooling fails quietly when a library, tactic, or corpus is assumed present rather than checked. A pipeline that routes a proof to aesop when aesop is not actually available, or that treats a premise index as proof evidence because it happens to carry a proof body, has already lost the boundary between \"ready to attempt\" and \"proven\". This component exists to make that boundary explicit before any downstream proof work begins. It answers one quest\u2026",
"level": 4,
"order": 5,
"section_id": "making-the-ready-versus-proven-boundary-explicit",
"source_line_end": 38,
"source_line_start": 25,
"title": "Making the ready-versus-proven boundary explicit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The mechanism is a deterministic reducer over five public JSON inputs that runs no Lean and instead refuses the specific ways they can lie, emitting a readiness board rather than theorem evidence.** The mechanism is a deterministic reducer over five public JSON inputs: corpus readiness, tactic-portfolio availability, a premise index, target-shape routing, and provider context recipes. It does not run Lean or Lake. Instead it reads what those inputs declare and refuses the specific ways they can lie. A corpus that claims Mathlib is available without a passing probe is rejected. A tactic marked available without a probe result record is rejected. A premise row carrying a proofbody or oracle\u2026",
"level": 4,
"order": 6,
"section_id": "a-reducer-that-refuses-specific-lies",
"source_line_end": 53,
"source_line_start": 39,
"title": "A reducer that refuses specific lies"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The gate proves its own discipline through five overclaim inputs that each run must catch, while a digest manifest keeps copied probe bodies source-open without leaking proof bodies, model-output data, or private state.** The design choice worth noticing is that the gate proves its own discipline through negative cases. Alongside the positive inputs, the fixture carries five inputs that each commit a known overclaim, and the run passes only when every one of those overclaims is caught and no unexpected finding appears. The gate is therefore not merely asserting \"we check Mathlib availability\"; it is demonstrating, on each run, that a falsified Mathlib claim is actually refused. A second g\u2026",
"level": 4,
"order": 7,
"section_id": "proving-discipline-through-negative-cases",
"source_line_end": 68,
"source_line_start": 54,
"title": "Proving discipline through negative cases"
},
{
"block_kinds": [
"diagram"
],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "shape",
"source_line_end": 121,
"source_line_start": 90,
"title": "Shape"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This diagram explains the reader flow, while the machine graph remains the The machine graph remains the generated source record, not from this hand-authored Mermaid block.",
"level": 4,
"order": 9,
"section_id": "the-hand-authored-diagram-is-not-the-machine-graph",
"source_line_end": 129,
"source_line_start": 122,
"title": "The hand-authored diagram is not the machine graph"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "Read this module in evidence order: Start at core/papermodulecapsules.json::papermodules[21:papermodule.formalmathreadiness_gate]. That row names the source authority, subjects, mechanism refs, code locus, Microcosm concept/principle/axiom refs, generated projection statuses, and the bundle scope limit. Check the generated structured source record papermodules/formalmathreadinessgate.json. Its relationships.edges cite the bundle source refs and show the generated Mermaid status, Atlas status, sourceauthority: jsoncapsule, and unresolved selective-relation count. Inspect the runtime locus src/microcosmcore/components/formalmathreadinessgate.py, especially run, runreadinessbundle, validatesou\u2026",
"level": 2,
"order": 10,
"section_id": "reader-evidence-routing",
"source_line_end": 140,
"source_line_start": 130,
"title": "Reader Evidence Routing"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A proof claim must go elsewhere because this page routes only readiness evidence, result record integrity, and source-body-floor accounting.** Do not route a proof claim through this page. It routes readiness evidence, result record integrity, and source-body-floor accounting only.",
"level": 4,
"order": 11,
"section_id": "this-page-routes-readiness-evidence-not-proof-claims",
"source_line_end": 147,
"source_line_start": 141,
"title": "This page routes readiness evidence, bounded evidence claims"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "technical-mechanism",
"source_line_end": 149,
"source_line_start": 148,
"title": "Technical Mechanism"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run() evaluates the fixture with negative cases and runreadiness_bundle() evaluates the exported bundle, both calling buildresult() so they share one scope limit, secret scan, digest checker, and board schema.** The runtime is a deterministic readiness reducer over declared public inputs. run() evaluates the first-wave fixture directory with positive and negative JSON cases enabled; runreadinessbundle() evaluates the exported public bundle without fixture-negative cases and requires the bundle source-module manifest. Both entrypoints call buildresult(), so the fixture and exported bundle result records share one scope limit, one secret scan, one source-module digest checker, and one readine\u2026",
"level": 4,
"order": 13,
"section_id": "one-reducer-two-entrypoints",
"source_line_end": 162,
"source_line_start": 150,
"title": "One reducer, two entrypoints"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "buildresult() loads the five public input families and scans them for forbidden material, and the result passes only with zero blocking hits, passing source-module imports, all negatives observed, and no stray positive findings.** buildresult() first loads the five public input families: corpusreadiness.json, tacticportfolio_availability.json, premiseindex.json, targetshapetacticrouting.json, and providercontextrecipes.json. It then scans those inputs plus any declared source artifacts through secretexclusionscan.scan_paths, using the public Microcosm forbidden-class policy. The scan is not advisory: the result can pass only when the scan has zero blocking hits, source-module imports pass,\u2026",
"level": 4,
"order": 14,
"section_id": "loading-inputs-and-the-non-advisory-scan",
"source_line_end": 204,
"source_line_start": 163,
"title": "Loading inputs and the non-advisory scan"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "mergeobserved() and mergefindings() compare observed failures against EXPECTEDNEGATIVECASES, then buildextension_board() projects accepted metadata into pattern ids, counts, blocked routes, budgets, and the scope limit. After the validators run, mergeobserved() and merge_findings() compare observed fixture failures against EXPECTEDNEGATIVECASES. This is the local scope limit: the fixture run must prove that the known overclaims are caught, while the exported-bundle run must prove that the positive public bundle has no unexpected findings. buildextension_board() then projects the accepted metadata into the extension board: selected pattern ids, namespace and split counts, tactic availability\u2026",
"level": 4,
"order": 15,
"section_id": "reconciling-findings-and-projecting-the-board",
"source_line_end": 220,
"source_line_start": 205,
"title": "Reconciling findings and projecting the board"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "writereceipts() emits the gate result and boards while preserving the boundary, and the focused test asserts the mechanism itself, from negative-case ids to digest-mismatch rejection and result record omission of copied bodies.** Result record writing preserves the same boundary. write_receipts() emits the gate result, readiness board, extension board, validation result record, and sign-off result record for fixture mode. runreadinessbundle() emits the exported-bundle result record. The focused test suite asserts the mechanism rather than just file existence: it checks the five expected negative case ids, local Lean/Lake probe metadata with Mathlib unavailable, six available tactics with ae\u2026",
"level": 4,
"order": 16,
"section_id": "receipts-and-the-asserting-test-suite",
"source_line_end": 236,
"source_line_start": 221,
"title": "Result records and the asserting test suite"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "public-contract",
"source_line_end": 238,
"source_line_start": 237,
"title": "Public Contract"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component runs no Lean or Lake; it records available and blocked capabilities and rejects Mathlib overclaims, unprobed tactics, proof-body premise rows, unavailable-tactic routes, and over-budget or proof-body recipes.** The component does not run Lean or Lake. It consumes public JSON fixtures and exported bundles, records which capabilities are available or blocked, rejects Mathlib availability overclaims, rejects unprobed tactics, rejects premise rows that contain proof bodies, rejects routes that admit unavailable tactics, and rejects provider recipes that exceed the public budget or allow proof bodies.",
"level": 4,
"order": 18,
"section_id": "what-the-organ-consumes-and-rejects",
"source_line_end": 250,
"source_line_start": 239,
"title": "What the component consumes and rejects"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The accepted result is a readiness board that can tell a later component what is safe to attempt, but it is bounded evidence evidence, benchmark evidence, or permission to run a prover.** The accepted result is a readiness board. That board can tell a later component what is safe to attempt, but it is bounded evidence evidence, benchmark evidence, or permission to execute a theorem prover.",
"level": 4,
"order": 19,
"section_id": "the-result-is-a-readiness-board",
"source_line_end": 258,
"source_line_start": 251,
"title": "The result is a readiness board"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "prior-art-grounding",
"source_line_end": 260,
"source_line_start": 259,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component is grounded in formal-math benchmark and environment-readiness work where a library or corpus being present is not enough, drawing on miniF2F, LeanDojo, and mathlib.** This component is grounded in formal-math benchmark and environment-readiness work where the presence of a library, tactic, or corpus is not enough by itself. miniF2F motivates explicit benchmark split discipline for formal mathematics, LeanDojo motivates reproducible theorem-proving environments, and mathlib makes the availability of library imports a concrete precondition rather than a vague capability claim.",
"level": 4,
"order": 21,
"section_id": "formal-math-readiness-and-environment-work",
"source_line_end": 273,
"source_line_start": 261,
"title": "Formal-math readiness and environment work"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Microcosm borrows the readiness-gate pattern, requiring corpus availability, Mathlib and tactic probes, premise indexes, routing, and context budgets to be checked before downstream proof language, without authorizing Lean execution.** Microcosm borrows the readiness-gate pattern: corpus availability, Mathlib probes, tactic probes, premise indexes, target-shape routing, and context budgets must be checked before downstream proof or retrieval language is allowed. It excludes Lean execution or proof authority.",
"level": 4,
"order": 22,
"section_id": "borrowing-the-readiness-gate-pattern",
"source_line_end": 283,
"source_line_start": 274,
"title": "Borrowing the readiness-gate pattern"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 23,
"section_id": "source-open-body-floor",
"source_line_end": 285,
"source_line_start": 284,
"title": "Source-Open Body Floor"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The exported bundle carries thirteen probe bodies plus two digest-bound path rewrites covering corpus, tactic, Mathlib, trace, and Lean portfolio readiness.** The exported readiness bundle carries thirteen PROVER smoke-run readiness/probe bodies under source_artifacts. They cover corpus readiness, tactic-affordance probe metadata, Mathlib and trace probes, and the copied portfolio-core Lean probes used to decide which tactics are blocked or available. Two JSON rows are private-path rewrites; those rows retain source and target digests plus the rewrite mode.",
"level": 4,
"order": 24,
"section_id": "thirteen-probe-bodies-preserve-the-readiness-floor",
"source_line_end": 297,
"source_line_start": 286,
"title": "Thirteen probe bodies preserve the readiness floor"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The public component-source copy is inspectable as body-floor evidence, while result records exclude copied text, private proof material, provider bodies, live session state, account secrets, and send state.** The bundle also carries an exact public component-source copy for src/microcosmcore/components/formalmathreadinessgate.py under sourcebodyfloor/source_modules. Generated state/runs Lean artifacts are runnable readiness evidence, not source-body authority. Neither floor places body text in result records or workingness cards, and neither imports model-output data bodies, account or browser state, browser UI live access, recipient-send state, account secrets, private proof bodies, or\u2026",
"level": 4,
"order": 25,
"section_id": "an-exact-organ-copy-remains-separate-from-receipts",
"source_line_end": 310,
"source_line_start": 298,
"title": "An exact component copy remains separate from result records"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Body-floor inspection belongs in the source and bundle manifests; validation result records deliberately retain only status, digests, counts, and public-relative references.** The source-module manifest and bundle manifest are the right surfaces for body-floor inspection. The validation result records intentionally carry status, digests, counts, and public-relative refs rather than copied source bodies.",
"level": 4,
"order": 26,
"section_id": "manifests-carry-bodies-while-receipts-carry-references",
"source_line_end": 319,
"source_line_start": 311,
"title": "Manifests carry bodies while result records carry references"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Wave 011's extension board records selected patterns, intake and validation refs, premise coverage, tactic and Mathlib status, route admissibility, and provider budgets without becoming proof evidence.** Wave 011 adds the explicit extension board for the source intake cell formalmathreadiness_extensions. The board is still metadata-only, but it is more useful than the older flat counts: it records the selected pattern ids (leanstdtoolchainpremiseindex, tacticportfolioavailability_probe, targetshapetacticroutinggate), the source projection intake ref, public target refs, validation refs, namespace and split coverage for the premise index, tactic availability status counts, Mathlib-dependen\u2026",
"level": 4,
"order": 27,
"section_id": "the-extension-board-projects-typed-readiness-metadata",
"source_line_end": 333,
"source_line_start": 320,
"title": "The extension board projects typed readiness metadata"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 28,
"section_id": "claim-ceiling",
"source_line_end": 335,
"source_line_start": 334,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The module may claim a public readiness gate for formal-math preparation, bounded to corpus availability, probe metadata, premise coverage, routing, budgets, pattern ids, source artifacts, an component-source floor, and result records.** This module may claim that Microcosm has a public readiness gate for formal math system preparation. The valid claim is bounded to corpus availability, Mathlib and tactic probe metadata, premise-index coverage, target-shape tactic routing, provider context budget checks, extension-board pattern ids, public PROVER smoke-run source artifacts, an exact public component-source body floor, and fixture or exported-bundle result records.",
"level": 4,
"order": 29,
"section_id": "the-bounded-valid-claim",
"source_line_end": 347,
"source_line_start": 336,
"title": "The bounded valid claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It must not claim Lean/Lake execution, theorem proving, proof authority, benchmark performance, external model access, private proof-body import, source-file changes, or launch, and its strongest statement is readiness-boundary enforcement.** The module must not claim Lean/Lake execution, theorem proving, formal proof authority, formal-result correctness, Mathlib-dependent proof success, benchmark performance, provider-call execution, private proof-body import, oracle-needed premise disclosure, source-file changes, public sharing decision, hosted deployment, recipient work, secret export, or whole-system correctness. Its strongest launch-facing statement is readiness-boundary enforcement\u2026",
"level": 4,
"order": 30,
"section_id": "what-the-module-must-not-claim",
"source_line_end": 360,
"source_line_start": 348,
"title": "What the module must not claim"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 31,
"section_id": "governing-lattice-relation",
"source_line_end": 385,
"source_line_start": 384,
"title": "Governing Lattice Relation"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The bundle's concept and two mechanisms bind the positive bundle-validation path separately from the validators that reject readiness-to-proof overclaims.** The bundle binds this module to concept.formalmathandproofwitness_bundle because the component is not a theorem prover; it is the membrane that decides which public formal-math inputs are safe enough for a later proof witness to consume. The governing mechanisms split that membrane in two. The validatespublicformalmathreadiness_bundle mechanism names the positive bundle path: run, runreadinessbundle, validatesourcemodule_imports, and write_receipts validate the declared corpus, tactic, premise, routing, provider-budget, source-module-\u2026",
"level": 4,
"order": 32,
"section_id": "two-mechanisms-separate-positive-readiness-from-refusal",
"source_line_end": 404,
"source_line_start": 386,
"title": "Two mechanisms separate positive readiness from refusal"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The cited principles separate bundle, structured source record, runtime, result record, and body classes, while the axioms keep private material excluded and the downstream proof witness deferred.** The principle and axiom refs are therefore operational, not decorative. P-1, P-2, and P-3 are expressed by keeping the JSON bundle, generated structured source record, runtime code locus, and result records as separate authority classes. P-6 and P-8 are expressed by the body-floor and secret-exclusion contracts: copied PROVER probe bodies and the public component source copy can be inspected through digests and manifests, while private proof bodies, model-output data bodies, and browser or acc\u2026",
"level": 4,
"order": 33,
"section_id": "principles-and-axioms-enforce-authority-separation",
"source_line_end": 420,
"source_line_start": 405,
"title": "Principles and axioms enforce authority separation"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The deliberately small edge set establishes bundle, source, and dependency bindings without proving witness execution, Mathlib availability, or a fully bound Atlas card.** The generated lattice edge count is small on purpose: it proves that this page is bundle-backed, source-bound, and connected to one deferred proof-witness module. It does not establish that the deferred witness can run, that Mathlib is available, or that the generated Atlas card is fully bound; the structured source record still those edges.",
"level": 4,
"order": 34,
"section_id": "fifteen-generated-edges-prove-connection-not-capability",
"source_line_end": 432,
"source_line_start": 421,
"title": "Fifteen generated edges prove connection, not capability"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 35,
"section_id": "limitations",
"source_line_end": 434,
"source_line_start": 433,
"title": "Limitations"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The runtime validates finite public fixtures without executing Lean or Lake, importing Mathlib, or calling a provider, so a blocked capability is a downstream boundary, not an invitation to route around the gate.** The runtime validates finite public fixtures and exported-bundle manifests. It does not execute Lean or Lake, import Mathlib in the current environment, call a provider, or check theorem statements. When the result reports blocked capabilities such as leanstdsynthetic_core:mathlib, that is a readiness boundary for downstream components, not an invitation to route around the gate.",
"level": 4,
"order": 36,
"section_id": "finite-fixtures-no-execution",
"source_line_end": 445,
"source_line_start": 435,
"title": "Finite fixtures, no execution"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The copied source artifacts are source-open body-floor evidence only: digest and target-ref checks show they match their manifests but authorize no source-file changes, proof-body disclosure, deployment, or public sharing.** The copied source artifacts are source-open body-floor evidence only. Digest and target-ref checks show that selected PROVER readiness/probe bodies and the public component source copy match their manifests; they do not authorize source-file changes, private source-root export, proof-body disclosure, recipient work, hosted deployment, or public sharing. Result records intentionally carry counts, digests, paths, negative-case coverage, and authority flags instead of co\u2026",
"level": 4,
"order": 37,
"section_id": "source-artifacts-are-body-floor-evidence-only",
"source_line_end": 458,
"source_line_start": 446,
"title": "Source artifacts are body-floor evidence only"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The negative cases are finite, covering the known overclaims in EXPECTEDNEGATIVECASES, so a new formal-math claim needs a new source-backed negative case or a different proof consumer, not this page.** The negative cases are also finite. They cover the known overclaims encoded in EXPECTEDNEGATIVECASES: Mathlib availability without a passing probe, unprobed tactic availability, premise rows with proof bodies, target routes that admit unavailable tactics, and provider recipes that exceed public budgets or permit proof bodies. A new formal-math claim needs either a new source-backed negative case here or a different proof consumer; this page should not be used as a generic formal-proof claim\u2026",
"level": 4,
"order": 38,
"section_id": "the-negative-cases-are-finite",
"source_line_end": 471,
"source_line_start": 459,
"title": "The negative cases are finite"
},
{
"block_kinds": [
"list"
],
"excerpt": "python -m microcosmcore.components.formalmathreadinessgate run --input fixtures/firstwave/formalmathreadinessgate/input --out result records/firstwave/formalmathreadinessgate python -m microcosmcore.components.formalmathreadinessgate run-readiness-bundle --input examples/formalmathreadinessgate/exportedformalmathreadinessbundle --out result records/runtimeshell/demoproject/components/formalmathreadinessgate python -m microcosmcore.components.formalmathreadinessgate plan --input fixtures/firstwave/formalmathreadinessgate/input microcosm formal-math-readiness-gate run --input fixtures/firstwave/formalmathreadinessgate/input --out result records/firstwave/formalmathreadinessgate microcosm form\u2026",
"level": 2,
"order": 39,
"section_id": "runtime-surfaces",
"source_line_end": 479,
"source_line_start": 472,
"title": "Runtime Surfaces"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 40,
"section_id": "validation-receipt-path",
"source_line_end": 481,
"source_line_start": 480,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "**The focused test, corpus parity check, and generated-row query validate this result record or generated projection surfaces:",
"level": 4,
"order": 41,
"section_id": "three-read-only-checks-validate-body-and-projection",
"source_line_end": 494,
"source_line_start": 482,
"title": "Three read-only checks validate body and projection"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A correct row reports fifteen edges, available Mermaid, owner-blocked Atlas, JSON-bundle authority, and no unresolved selective relation.** Expected generated-row proof: edge_count: 15, mermaidstatus: availablefromcapsuleedges, atlasstatus: blockeduntilorganatlasownerlanebindsedges, sourceauthority: jsoncapsule, and unresolvedselectiverelation_count: 0.",
"level": 4,
"order": 42,
"section_id": "the-generated-row-should-expose-fifteen-resolved-edges",
"source_line_end": 504,
"source_line_start": 495,
"title": "The generated row should expose fifteen resolved edges"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 43,
"section_id": "relationship-to-lean-witness",
"source_line_end": 535,
"source_line_start": 534,
"title": "Relationship To Lean Witness"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "formalmathleanproof_witness remains deferred, and this gate makes the deferral typed and testable: absent Mathlib, unroutable unavailable tactics, no proof or oracle bodies in premises, and no proof-body provider recipes.** formalmathleanproofwitness remains deferred. This gate makes the deferral typed and testable: Mathlib is absent until a passing probe says otherwise, unavailable tactics cannot be routed, premise indexes cannot carry proof or oracle bodies, and provider recipes cannot smuggle proof-body deliverables.",
"level": 4,
"order": 44,
"section_id": "the-deferral-is-typed-and-testable",
"source_line_end": 545,
"source_line_start": 536,
"title": "The deferral is typed and testable"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 45,
"section_id": "scope-boundary",
"source_line_end": 547,
"source_line_start": 546,
"title": "Scope boundary"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This module documents a public readiness gate only, authorizing no Lean/Lake execution, proof authority, external model access, launch, or deployment, and adding no private source-root, account, or proof-body material to the public floor.** This module documents a public readiness gate only. It excludes Lean/Lake execution, formal proof authority, Mathlib-dependent proof attempts, external model access, benchmark claims, public launch, hosted deployment, public sharing, recipient work, secret export, or whole-system correctness. It also does not make private source-root material, browser UI state, account or browser material, browser state, account secrets, source notes, model-output data\u2026",
"level": 4,
"order": 46,
"section_id": "what-this-module-does-not-authorize",
"source_line_end": 560,
"source_line_start": 548,
"title": "What this module excludes"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/formal_math_readiness_gate.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "teleology",
"source": "markdown_section",
"text": null
}
},
"source_drilldown": {
"source_manifests": [
{
"label": "Source manifest",
"path": "examples/formal_math_readiness_gate/exported_formal_math_readiness_bundle/source_module_manifest.json",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/examples/formal_math_readiness_gate/exported_formal_math_readiness_bundle/source_module_manifest.json"
}
],
"source_modules": [
{
"label": "formal_math_readiness_gate.py",
"path": "src/microcosm_core/organs/formal_math_readiness_gate.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public formal-math readiness fixture validation, exported readiness-bundle validation, copied source-module manifest checks, negative-case enforcement, readiness boards, and scope limits.",
"symbols": [
"run",
"run_readiness_bundle",
"validate_source_module_imports",
"write_receipts",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING",
"SOURCE_MODULE_MANIFEST_NAME"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/formal_math_readiness_gate.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_math_readiness_gate.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_math_readiness_gate.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Formal Math Readiness Gate is the public readiness membrane before downstream proof work. It checks declared Mathlib and corpus readiness, tactic probe result records, proof-metadata-only premise indexes, target-shape routing, provider context budgets, formal_math_readiness_extensions intake rows, copied PROVER smoke-run artifacts, public component source body imports, digest manifests, secret exclusion, and negative cases, then emits readiness boards rather than theorem evidence.",
"summary_line": "The public formal-math readiness gate validates declared corpus, tactic, premise, routing, provider-budget, source-module manifest, copied PROVER probe body, and negative-case boundaries without claiming Lean/Lake or proof authority.",
"title": "Formal Math Readiness Gate"
},
{
"anchor": "paper-module-formal-math-lean-proof-witness",
"authority_ceiling": "Declared public toy Lean/Lake fixture result records only; no Mathlib/Aesop/Batteries authority, general formal-result correctness, private proof import, external model access, benchmark performance, launch-scope decision, publishing-scope decision, source-file changes, or whole-system correctness.",
"body_html": "
Component checks whether
\n
This component checks whether the installed Lean toolchain actually compiles a tiny purpose-built project end to end and whether that run can be recorded without copying the proof into the record. This component answers one narrow question and refuses the larger ones. The question is whether the installed Lean toolchain will actually compile a tiny, purpose-built Lean project end to end, and whether that run can be recorded without copying the proof into the record. It is not a theorem prover, a benchmark, a Mathlib check, or an import of any private proof. It runs lake build over a synthetic fixture and writes down what happened.
\n
Emits metadata-only record
\n
It emits a metadata-only record of tool availability, build return code, output line counts, source hashes, declaration names, and per-file counts, while the proof text and raw tool output never reach it. What it emits is a metadata-only result record: whether Lean and Lake were found, the build return code, line counts for standard output and standard error, source file hashes, the theorem and definition names it parsed out, and per-file line counts. The proof text and the raw tool output never reach that record. A reader gets evidence that the build ran and what it contained, not a copy of the proof.
\n
Point is discipline
\n
The point is the discipline around the run: a real subprocess ran, a deliberately broken proof was rejected, and no answer was smuggled in through the inputs. The point of the component is the discipline around the run. Anyone can print "Lean says it compiles." The harder claim is that a real subprocess ran, that a deliberately broken proof was rejected, and that no answer was smuggled in through the inputs.
\n
Purpose
\n
Guarding against a checker-free pass
\n
The first failure mode is an integration that reports success without invoking the checker, so the witness runs a real lake build subprocess and separately compiles an invalid file it must reject. Two failure modes shape the design. The first is an integration that reports success without ever invoking the checker. The witness guards against that by running a real lake build subprocess, recording its exit status, and separately compiling an intentionally invalid Lean file to confirm the toolchain rejects it. A green result that never touched Lean is not reachable through this path.
\n
Rejecting the circular pass
\n
The second failure mode is a circular pass where the manifest carries the answer, so the component rejects any manifest embedding a proof body, ground truth, provider output, oracle ids, or private body. The second failure mode is a circular pass, where the input manifest quietly carries the answer. The component rejects manifests that embed a proof_body, a ground_truth_proof, provider output, oracle premise ids, or a private source body. Each of those has a named negative case. A loose witness would accept a manifest that already states the conclusion; this one does not.
\n
The scope is small on purpose
\n
The scope is deliberately small: Mathlib, Aesop, and Batteries imports are rejected before anything runs, so this witnesses a toy theorem under a local toolchain, not library-dependent proof work. The scope is small on purpose. Imports of Mathlib, Aesop, and Batteries are rejected before anything runs. This is a witness for a toy theorem under a local toolchain, not a claim about library-dependent proof work. That boundary is the deliverable: it shows the record-keeping a larger formal-math component would need, without borrowing authority it has not earned.
\n
How it works
\n
Two entry points, one builder
\n
run drives the first-wave fixture with negative cases on and run_witness_bundle drives the exported bundle with them off, both delegating to _build_result, which runs in a fixed order. The runtime lives in src/microcosm_core/organs/formal_math_lean_proof_witness.py. Two entry points exist. run drives the first-wave fixture with include_negative=True. run_witness_bundle drives the exported public bundle with include_negative=False, and reuses a fresh prior result record through _fresh_bundle_receipt when the inputs have not changed. Both delegate the real work to _build_result, which executes in this order.
\n
Validating copied source modules
\n
In bundle mode validate_source_module_imports recomputes each copied module's SHA-256 against its live source, checks the declared exact-copy or replacement relation, material class, and anchors, and blocks on any mismatch. First it collects the input file set with _input_paths and, for the bundle mode, validates copied source modules with validate_source_module_imports. That function reads source_module_manifest.json, and for each declared module it recomputes the SHA-256 of the copied target, recomputes the SHA-256 of the live source it was copied from, and checks that the declared source_to_target_relation is exact_copy (matching digests, sha256_match: true) or public_replacement_source_body (distinct digests, sha256_match: false). It also checks the material class is one of the four public classes, that body_in_receipt is false, that the source ref sits under an allowed public prefix, and that every declared anchor string is present in the target. Any mismatch becomes a finding, and any finding blocks the module.
\n
Scanning inputs and the witness manifest
\n
It scans every input and copied artifact for forbidden material, then _validate_manifest rejects forbidden keys and non-public refs and requires the positive witness rows, Lean files, and source pattern ids. Next it scans every input and copied artifact for forbidden material with scan_paths against the public forbidden-class policy, then validates the main witness_manifest.json with _validate_manifest. That check walks the manifest for the forbidden keys, rejects non-public source refs, and requires the positive witness rows, declared Lean files, source refs, public replacement refs, projection result record refs, and source pattern ids to all be present.
\n
Probing the toolchain and running the build
\n
_tool_versions records whether lean and lake are on the path, then _run_command runs lake build in a temporary copy, keeping only the return code, output line counts, and timeout flag. Then it probes the toolchain. _tool_versions uses shutil.which to record whether lean and lake are on the path, and caches the answer. For the live path, _copy_project_to_temp copies the Lake project into a temporary workspace and _run_command runs lake build MicrocosmProofWitness with a 90-second timeout. _run_command captures the subprocess but keeps only the return code, the standard-output and standard-error line counts, and the timeout flag. It never keeps the output body.
\n
Redacted per-file source metadata
\n
_source_metadata records each positive Lean file's ref, hash, line count, declaration names, and imports with body_redacted: True, and sums the compiled declaration count across files. For each positive Lean file, _source_metadata reads the text and records its public-relative ref, its hash, its line count, the declaration names found by a theorem|lemma|def regular expression, and its imports, with body_redacted: True. The compiled declaration count is the sum across those files.
\n
Running the rejection checks
\n
With negatives enabled, _build_result compiles the invalid proof and forbidden-import files for real and runs _validate_negative_manifest over four tampered manifests, recording the expected code for each. When negative cases are enabled, _build_result runs the rejection checks. It copies invalid_proof.lean into the built project and runs lake env lean on it; a non-zero exit records LEAN_WITNESS_INVALID_PROOF_REJECTED. It scans mathlib_import_forbidden.lean with _forbidden_imports and records LEAN_WITNESS_FORBIDDEN_IMPORT when a blocked import is present. It then runs _validate_negative_manifest over four tampered manifests for the private-source-ref, proof-body, provider-output, and oracle-premise-id cases.
\n
Computing the final status
\n
A result is pass only when Lean and Lake are present, the build returned 0, the manifest and bundle modules validated, no forbidden import appeared, no negative case is missing, and the scan is clean. Finally the status is computed. A result is pass only when Lean and Lake are available, the Lake build returned 0, the manifest validated, bundle source modules validated when in bundle mode, no forbidden imports appeared in the positive files, no expected negative case is missing, and the non-public-state scan found zero blocking hits. Otherwise the status is blocked. _build_result returns the record with the scope limit, the compiled declaration count, the redacted tool metadata, and the source file metadata attached.
\n
Function
Role
run
First-wave fixture entry point; negative cases on.
run_witness_bundle
Exported bundle entry point; reuses a fresh record when inputs are unchanged.
_build_result
Orchestrates validation, toolchain probe, Lake build, negative cases, and status.
validate_source_module_imports
Checks copied public source digests, relations, and anchors in the bundle manifest.
_validate_manifest
Rejects forbidden keys and non-public refs; requires the positive witness fields.
_run_command
Runs a subprocess and keeps only return code, line counts, and timeout state.
_tool_versions
Records lean and lake availability without running the checker.
_source_metadata
Emits hash, line count, declarations, and imports with the body redacted.
\nDiagram of the mechanism (11 steps).Diagram source & refs
Source refs
Fixture input
witness_manifest.json + lake_project
Exported bundle
source_module_manifest.json
recompute copied digests, check exact_copy vs replacement
validate_source_module_imports
flowchart TD A["Fixture input witness_manifest.json + lake_project"] --> B["run() include_negative=true"] C["Exported bundle source_module_manifest.json"] --> D["run_witness_bundle() include_negative=false"] B --> E["_build_result"] D --> E E --> F["_validate_manifest reject proof bodies, non-public source refs"] E --> G["validate_source_module_imports recompute copied digests, check exact_copy vs replacement"] E --> H["_tool_versions lean/lake on path?"] H --> I["_run_command lake build MicrocosmProofWitness"] I --> J["Negative cases: invalid proof rejected, forbidden import blocked"] F --> K["metadata-only result record: return code, line counts, hashes, declaration names"] G --> K I --> K J --> K
\n
Negative cases
\n
Fixture ships six
\n
The fixture ships six named negative cases, each with an expected error code, defined in EXPECTED_NEGATIVE_CASES. The fixture ships six named negative cases, each with an expected error code, defined in EXPECTED_NEGATIVE_CASES.
\n
Six refusals, each with a code
\n
A real invalid proof, a blocked import, and four tampered manifests each raise their named error code, and if any expected case is not observed the run is blocked. An invalid Lean proof (invalid_proof.lean) is compiled for real and must fail: LEAN_WITNESS_INVALID_PROOF_REJECTED. A file importing a blocked library (mathlib_import_forbidden.lean) is caught before execution: LEAN_WITNESS_FORBIDDEN_IMPORT. A manifest citing a non-public source refs is refused: LEAN_WITNESS_PRIVATE_SOURCE_REF_FORBIDDEN. A manifest embedding a proof body is refused: LEAN_WITNESS_PROOF_BODY_IN_MANIFEST_FORBIDDEN. A manifest carrying provider output is refused: LEAN_WITNESS_PROVIDER_OUTPUT_IN_MANIFEST_FORBIDDEN. A manifest carrying oracle premise ids is refused: LEAN_WITNESS_ORACLE_PREMISE_IDS_IN_MANIFEST_FORBIDDEN. If any expected case is not observed, the run is blocked.
\n
Prior Art Grounding
\n
The Lean proof-assistant lineage
\n
The component sits in the Lean small-kernel lineage, borrowing the discipline of a local toolchain run with source hashes, declaration names, negative cases, and metadata-only records, not Mathlib-dependent authority. This component sits in the Lean proof-assistant lineage and the small-kernel theorem-proving tradition it belongs to. The Lean theorem prover system description describes the toolchain this witness invokes, and the Lean mathematical library shows why real proof authority depends on explicit imports, declarations, and a checked environment. The witness borrows the discipline, a local toolchain run with source hashes, declaration names, negative cases, and metadata-only records, before any Lean-witness language is allowed. It does not borrow Mathlib-dependent proof authority.
A pass means the focused test saw a real Lake build for the toy witness, the six negative cases, redacted records, and a generated row reporting edge_count: 8 and source_authority: json_capsule. A pass means the focused test saw a successful Lake build for the toy witness when Lean and Lake were available, saw the six negative cases, saw public-relative redacted records, and the generated row reports edge_count: 8, mermaid_status: available_from_capsule_edges, source_authority: json_capsule, and zero unresolved selective relations.
\n
To run the component directly:
\n
PYTHONPATH=src python3 -m microcosm_core.organs.formal_math_lean_proof_witness run --input fixtures/first_wave/formal_math_lean_proof_witness/input --out receipts/first_wave/formal_math_lean_proof_witness
\n
Scope boundary
\n
Scope limit
\n
The narrow proof boundary
\n
The strongest claim is that a tiny public Lean fixture compiles in a temporary local workspace and the run is recorded as hashes, declarations, tool metadata, and verdicts without exporting any proof body. The strongest claim the evidence supports is narrow: a tiny public Lean fixture compiles in a temporary local workspace using the installed Lean and Lake, and the run can be recorded as source hashes, declaration names, tool-return metadata, and negative-case verdicts without exporting a single proof body. That is the whole proof boundary. The positive evidence is one declared toy fixture, one exported public bundle, the copied source-module rows, and the compiled declarations when the toolchain is present.
\n
What the witness excludes
\n
It proves no arbitrary Lean goals, Mathlib coverage, or general correctness, and excludes external model access, private proof import, benchmark claim, launch, deployment, or source-file changes, even on a green run. It does not establish arbitrary Lean goals, Mathlib coverage, or general formal-result correctness. It excludes external model access, no private proof import, no benchmark performance claim, no launch, no hosted deployment, no public sharing, no source-file changes, and no whole-system correctness. Records may cite refs, hashes, material classes, declaration names, counts, and tool-return summaries; they may not embed proof bodies, model-output data, oracle answers, non-public source refs, or raw command output. This is the scope limit, and it holds even when the focused test and the corpus check are green.
\n",
"components": [
{
"anchor": "component-formal_math_lean_proof_witness",
"component_id": "formal_math_lean_proof_witness",
"public_label": "Formal Math Lean Proof Witness"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 2,
"organ": 1,
"paper_module": 1,
"principle": 1
},
"link_count": 7,
"links": [
{
"id": "formal_math_lean_proof_witness",
"kind": "organ",
"kind_label": "component",
"label": "Formal Math Lean Proof Witness",
"relation": "explains",
"url": "component-formal-math-lean-proof-witness.html"
},
{
"id": "mechanism.formal_math_lean_proof_witness.validates_public_lean_lake_witness",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public lean lake witness",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-lean-proof-witness-validates-public-lean-lake-witness"
},
{
"id": "mechanism.formal_math_lean_proof_witness.validates_public_lean_witness",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public lean witness",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-lean-proof-witness-validates-public-lean-witness"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.corpus_readiness_mathlib_absence_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Corpus Readiness Mathlib Absence Gate",
"relation": "depends on",
"url": "paper-module-corpus-readiness-mathlib-absence-gate.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.formal_math_premise_retrieval",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Premise Retrieval",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[25:paper_module.formal_math_premise_retrieval].depends_on",
"url": "paper-module-formal-math-premise-retrieval.html"
},
{
"id": "paper_module.formal_math_readiness_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Readiness Gate",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[21:paper_module.formal_math_readiness_gate].depends_on",
"url": "paper-module-formal-math-readiness-gate.html"
},
{
"id": "paper_module.proof_diagnostic_evidence_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof Diagnostic Evidence Spine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[14:paper_module.proof_diagnostic_evidence_spine].depends_on",
"url": "paper-module-proof-diagnostic-evidence-spine.html"
},
{
"id": "paper_module.verifier_lab_execution_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Execution Spine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[44:paper_module.verifier_lab_execution_spine].depends_on",
"url": "paper-module-verifier-lab-execution-spine.html"
},
{
"id": "paper_module.verifier_lab_kernel",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Kernel",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[0:paper_module.verifier_lab_kernel].depends_on",
"url": "paper-module-verifier-lab-kernel.html"
}
],
"paper_module_id": "paper_module.formal_math_lean_proof_witness",
"reader_packet": {
"abstract": "The public Lean proof witness runs local Lean/Lake over a tiny synthetic project, validates copied public source-module digests and negative cases, and emits redacted result records without claiming general proof authority.",
"authority_ceiling": "Declared public toy Lean/Lake fixture result records only; no Mathlib/Aesop/Batteries authority, general formal-result correctness, private proof import, external model access, benchmark performance, launch-scope decision, publishing-scope decision, source-file changes, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.formal_math_lean_proof_witness.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-formal-math-lean-proof-witness-diagram-1",
"html_url": "docs/paper-module-formal-math-lean-proof-witness.html#paper-module-formal-math-lean-proof-witness-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Fixture input witness_manifest.json + lake_project\"] --> B[\"run() include_negative=true\"] C[\"Exported bundle source_module_manifest.json\"] --> D[\"run_witness_bundle() include_negative=false\"] B --> E[\"_build_result\"] D --> E E --> F[\"_validate_manifest reject proof bodies, non-public source refs\"] E --> G[\"validate_source_module_imports recompute copied digests, check exact_copy vs replacement\"] E --> H[\"_tool_versions lean/lake on path?\"] H --> I[\"_run_command lake build MicrocosmProofWitness\"] I --> J[\"Negative cases: invalid proof rejected, forbidden import blocked\"] F --> K[\"metadata-only result record: return code, line counts, hashes, declaration names\"] G --> K I --> K J --> K",
"source_format": "mermaid",
"source_hash": "sha256:327b2f8cd1b8186e50e5ab1b3c2d7c4ef9720871f3b55afc2f1e3738ee47503a",
"source_line_end": 185,
"source_line_start": 170
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.corpus_readiness_mathlib_absence_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Corpus Readiness Mathlib Absence Gate",
"relation": "depends on",
"url": "paper-module-corpus-readiness-mathlib-absence-gate.html"
}
],
"relation": "depends on"
},
{
"count": 3,
"links": [
{
"id": "formal_math_lean_proof_witness",
"kind": "organ",
"kind_label": "component",
"label": "Formal Math Lean Proof Witness",
"relation": "explains",
"url": "component-formal-math-lean-proof-witness.html"
},
{
"id": "mechanism.formal_math_lean_proof_witness.validates_public_lean_lake_witness",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public lean lake witness",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-lean-proof-witness-validates-public-lean-lake-witness"
},
{
"id": "mechanism.formal_math_lean_proof_witness.validates_public_lean_witness",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public lean witness",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-lean-proof-witness-validates-public-lean-witness"
}
],
"relation": "explains"
},
{
"count": 2,
"links": [
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.formal_math_lean_proof_witness",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 284,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_math_lean_proof_witness.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Lean theorem prover system description",
"section_id": "the-lean-proof-assistant-lineage",
"url": "https://www.microsoft.com/en-us/research/publication/the-lean-theorem-prover-system-description/"
},
{
"kind": "external_reference",
"label": "Lean mathematical library",
"section_id": "the-lean-proof-assistant-lineage",
"url": "https://arxiv.org/abs/1910.09336"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component checks whether the installed Lean toolchain actually compiles a tiny purpose-built project end to end and whether that run can be recorded without copying the proof into the record.** This component answers one narrow question and refuses the larger ones. The question is whether the installed Lean toolchain will actually compile a tiny, purpose-built Lean project end to end, and whether that run can be recorded without copying the proof into the record. It is not a theorem prover, a benchmark, a Mathlib check, or an import of any private proof. It runs lake build over a synthetic fixture and writes down what happened.",
"level": 4,
"order": 1,
"section_id": "component-checks-whether",
"source_line_end": 13,
"source_line_start": 3,
"title": "Component checks whether"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It emits a metadata-only record of tool availability, build return code, output line counts, source hashes, declaration names, and per-file counts, while the proof text and raw tool output never reach it.** What it emits is a metadata-only result record: whether Lean and Lake were found, the build return code, line counts for standard output and standard error, source file hashes, the theorem and definition names it parsed out, and per-file line counts. The proof text and the raw tool output never reach that record. A reader gets evidence that the build ran and what it contained, not a copy of the proof.",
"level": 4,
"order": 2,
"section_id": "emits-body-free-record",
"source_line_end": 24,
"source_line_start": 14,
"title": "Emits metadata-only record"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The point is the discipline around the run: a real subprocess ran, a deliberately broken proof was rejected, and no answer was smuggled in through the inputs.** The point of the component is the discipline around the run. Anyone can print \"Lean says it compiles.\" The harder claim is that a real subprocess ran, that a deliberately broken proof was rejected, and that no answer was smuggled in through the inputs.",
"level": 4,
"order": 3,
"section_id": "point-is-discipline",
"source_line_end": 32,
"source_line_start": 25,
"title": "Point is discipline"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 34,
"source_line_start": 33,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The first failure mode is an integration that reports success without invoking the checker, so the witness runs a real lake build subprocess and separately compiles an invalid file it must reject.** Two failure modes shape the design. The first is an integration that reports success without ever invoking the checker. The witness guards against that by running a real lake build subprocess, recording its exit status, and separately compiling an intentionally invalid Lean file to confirm the toolchain rejects it. A green result that never touched Lean is not reachable through this path.",
"level": 4,
"order": 5,
"section_id": "guarding-against-a-checker-free-pass",
"source_line_end": 45,
"source_line_start": 35,
"title": "Guarding against a checker-free pass"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The second failure mode is a circular pass where the manifest carries the answer, so the component rejects any manifest embedding a proof body, ground truth, provider output, oracle ids, or private body.** The second failure mode is a circular pass, where the input manifest quietly carries the answer. The component rejects manifests that embed a proofbody, a groundtruth_proof, provider output, oracle premise ids, or a private source body. Each of those has a named negative case. A loose witness would accept a manifest that already states the conclusion; this one does not.",
"level": 4,
"order": 6,
"section_id": "rejecting-the-circular-pass",
"source_line_end": 56,
"source_line_start": 46,
"title": "Rejecting the circular pass"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The scope is deliberately small: Mathlib, Aesop, and Batteries imports are rejected before anything runs, so this witnesses a toy theorem under a local toolchain, not library-dependent proof work.** The scope is small on purpose. Imports of Mathlib, Aesop, and Batteries are rejected before anything runs. This is a witness for a toy theorem under a local toolchain, not a claim about library-dependent proof work. That boundary is the deliverable: it shows the record-keeping a larger formal-math component would need, without borrowing authority it has not earned.",
"level": 4,
"order": 7,
"section_id": "the-scope-is-small-on-purpose",
"source_line_end": 67,
"source_line_start": 57,
"title": "The scope is small on purpose"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 69,
"source_line_start": 68,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run drives the first-wave fixture with negative cases on and runwitness_bundle drives the exported bundle with them off, both delegating to buildresult, which runs in a fixed order.** The runtime lives in src/microcosmcore/components/formalmathleanproof_witness.py. Two entry points exist. run drives the first-wave fixture with include_negative=True. runwitnessbundle drives the exported public bundle with include_negative=False, and reuses a fresh prior result record through freshbundle_receipt when the inputs have not changed. Both delegate the real work to buildresult, which executes in this order.",
"level": 4,
"order": 9,
"section_id": "two-entry-points-one-builder",
"source_line_end": 81,
"source_line_start": 70,
"title": "Two entry points, one builder"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "In bundle mode validatesourcemoduleimports recomputes each copied module's SHA-256 against its live source, checks the declared exact-copy or replacement relation, material class, and anchors, and blocks on any mismatch.** First it collects the input file set with inputpaths and, for the bundle mode, validates copied source modules with validatesourcemodule_imports. That function reads sourcemodulemanifest.json, and for each declared module it recomputes the SHA-256 of the copied target, recomputes the SHA-256 of the live source it was copied from, and checks that the declared sourcetotarget_relation is exactcopy (matching digests, sha256match: true) or publicreplacementsourcebody (distinct\u2026",
"level": 4,
"order": 10,
"section_id": "validating-copied-source-modules",
"source_line_end": 98,
"source_line_start": 82,
"title": "Validating copied source modules"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It scans every input and copied artifact for forbidden material, then validatemanifest rejects forbidden keys and non-public refs and requires the positive witness rows, Lean files, and source pattern ids.** Next it scans every input and copied artifact for forbidden material with scan_paths against the public forbidden-class policy, then validates the main witness_manifest.json with validatemanifest. That check walks the manifest for the forbidden keys, rejects non-public source refs, and requires the positive witness rows, declared Lean files, source refs, public replacement refs, projection result record refs, and source pattern ids to all be present.",
"level": 4,
"order": 11,
"section_id": "scanning-inputs-and-the-witness-manifest",
"source_line_end": 110,
"source_line_start": 99,
"title": "Scanning inputs and the witness manifest"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "toolversions records whether lean and lake are on the path, then runcommand runs lake build in a temporary copy, keeping only the return code, output line counts, and timeout flag.** Then it probes the toolchain. toolversions uses shutil.which to record whether lean and lake are on the path, and caches the answer. For the live path, copyprojecttotemp copies the Lake project into a temporary workspace and runcommand runs lake build MicrocosmProofWitness with a 90-second timeout. runcommand captures the subprocess but keeps only the return code, the standard-output and standard-error line counts, and the timeout flag. It never keeps the output body.",
"level": 4,
"order": 12,
"section_id": "probing-the-toolchain-and-running-the-build",
"source_line_end": 122,
"source_line_start": 111,
"title": "Probing the toolchain and running the build"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourcemetadata records each positive Lean file's ref, hash, line count, declaration names, and imports with body_redacted: True, and sums the compiled declaration count across files. For each positive Lean file, source_metadata reads the text and records its public-relative ref, its hash, its line count, the declaration names found by a theorem|lemma|def regular expression, and its imports, with body_redacted: True. The compiled declaration count is the sum across those files.",
"level": 4,
"order": 13,
"section_id": "redacted-per-file-source-metadata",
"source_line_end": 132,
"source_line_start": 123,
"title": "Redacted per-file source metadata"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "With negatives enabled, build_result compiles the invalid proof and forbidden-import files for real and runs validatenegative_manifest over four tampered manifests, recording the expected code for each.** When negative cases are enabled, buildresult runs the rejection checks. It copies invalid_proof.lean into the built project and runs lake env lean on it; a non-zero exit records LEANWITNESSINVALIDPROOFREJECTED. It scans mathlibimportforbidden.lean with forbiddenimports and records LEANWITNESSFORBIDDEN_IMPORT when a blocked import is present. It then runs validatenegative_manifest over four tampered manifests for the private-source-ref, proof-body, provider-output, and oracle-premise-id cas\u2026",
"level": 4,
"order": 14,
"section_id": "running-the-rejection-checks",
"source_line_end": 145,
"source_line_start": 133,
"title": "Running the rejection checks"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "**A result is pass only when Lean and Lake are present, the build returned 0, the manifest and bundle modules validated, no forbidden import appeared, no negative case is missing, and the scan is clean.** Finally the status is computed. A result is pass only when Lean and Lake are available, the Lake build returned 0, the manifest validated, bundle source modules validated when in bundle mode, no forbidden imports appeared in the positive files, no expected negative case is missing, and the non-public-state scan found zero blocking hits. Otherwise the status is blocked. buildresult returns the record with the scope limit, the compiled declaration count, the redacted tool metadata, and the s\u2026",
"level": 4,
"order": 15,
"section_id": "computing-the-final-status",
"source_line_end": 186,
"source_line_start": 146,
"title": "Computing the final status"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "negative-cases",
"source_line_end": 188,
"source_line_start": 187,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The fixture ships six named negative cases, each with an expected error code, defined in EXPECTEDNEGATIVECASES.** The fixture ships six named negative cases, each with an expected error code, defined in EXPECTEDNEGATIVECASES.",
"level": 4,
"order": 17,
"section_id": "fixture-ships-six",
"source_line_end": 194,
"source_line_start": 189,
"title": "Fixture ships six"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A real invalid proof, a blocked import, and four tampered manifests each raise their named error code, and if any expected case is not observed the run is blocked. An invalid Lean proof (invalidproof.lean) is compiled for real and must fail: LEANWITNESSINVALIDPROOFREJECTED. A file importing a blocked library (mathlibimportforbidden.lean) is caught before execution: LEANWITNESSFORBIDDEN_IMPORT. A manifest citing a non-public source refs is refused: LEANWITNESSPRIVATESOURCEREF_FORBIDDEN. A manifest embedding a proof body is refused: LEANWITNESSPROOFBODYINMANIFESTFORBIDDEN. A manifest carrying provider output is refused: LEANWITNESSPROVIDEROUTPUTINMANIFESTFORBIDDEN. A manifest carrying oracl\u2026",
"level": 4,
"order": 18,
"section_id": "six-refusals-each-with-a-code",
"source_line_end": 209,
"source_line_start": 195,
"title": "Six refusals, each with a code"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "prior-art-grounding",
"source_line_end": 211,
"source_line_start": 210,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component sits in the Lean small-kernel lineage, borrowing the discipline of a local toolchain run with source hashes, declaration names, negative cases, and metadata-only records, not Mathlib-dependent authority.** This component sits in the Lean proof-assistant lineage and the small-kernel theorem-proving tradition it belongs to. The Lean theorem prover system description describes the toolchain this witness invokes, and the Lean mathematical library shows why real proof authority depends on explicit imports, declarations, and a checked environment. The witness borrows the discipline, a local toolchain run with source hashes, declaration names, negative cases, and metadata-only reco\u2026",
"level": 4,
"order": 20,
"section_id": "the-lean-proof-assistant-lineage",
"source_line_end": 225,
"source_line_start": 212,
"title": "The Lean proof-assistant lineage"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "validation-receipt-path",
"source_line_end": 227,
"source_line_start": 226,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "durable result or projection files:",
"level": 4,
"order": 22,
"section_id": "validate-reader-projection",
"source_line_end": 239,
"source_line_start": 228,
"title": "Validate reader projection"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "**A pass means the focused test saw a real Lake build for the toy witness, the six negative cases, redacted records, and a generated row reporting edge_count: 8 and sourceauthority: jsoncapsule.** A pass means the focused test saw a successful Lake build for the toy witness when Lean and Lake were available, saw the six negative cases, saw public-relative redacted records, and the generated row reports edgecount: 8, mermaidstatus: availablefromcapsule_edges, sourceauthority: jsoncapsule, and zero unresolved selective relations. To run the component directly:",
"level": 4,
"order": 23,
"section_id": "what-a-pass-means-here",
"source_line_end": 255,
"source_line_start": 240,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 24,
"section_id": "claim-ceiling",
"source_line_end": 257,
"source_line_start": 256,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is that a tiny public Lean fixture compiles in a temporary local workspace and the run is recorded as hashes, declarations, tool metadata, and verdicts without exporting any proof body.** The strongest claim the evidence supports is narrow: a tiny public Lean fixture compiles in a temporary local workspace using the installed Lean and Lake, and the run can be recorded as source hashes, declaration names, tool-return metadata, and negative-case verdicts without exporting a single proof body. That is the whole proof boundary. The positive evidence is one declared toy fixture, one exported public bundle, the copied source-module rows, and the compiled declarations when th\u2026",
"level": 4,
"order": 25,
"section_id": "the-narrow-proof-boundary",
"source_line_end": 269,
"source_line_start": 258,
"title": "The narrow proof boundary"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It proves no arbitrary Lean goals, Mathlib coverage, or general correctness, and excludes external model access, private proof import, benchmark claim, launch, deployment, or source-file changes, even on a green run.** It does not establish arbitrary Lean goals, Mathlib coverage, or general formal-result correctness. It excludes external model access, no private proof import, no benchmark performance claim, no launch, no hosted deployment, no public sharing, no source-file changes, and no whole-system correctness. Records may cite refs, hashes, material classes, declaration names, counts, and tool-return summaries; they may not embed proof bodies, model-output data, oracle answers, non-pu\u2026",
"level": 4,
"order": 26,
"section_id": "what-the-witness-does-not-authorize",
"source_line_end": 283,
"source_line_start": 270,
"title": "What the witness excludes"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/formal_math_lean_proof_witness.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public Lean proof witness runs local Lean/Lake over a tiny synthetic project, validates copied public source-module digests and negative cases, and emits redacted result records without claiming general proof authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "formal_math_lean_proof_witness.py",
"path": "src/microcosm_core/organs/formal_math_lean_proof_witness.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for the public Lean/Lake witness fixture, exported witness-bundle validation, copied source-module digest checks, negative-case enforcement, and scope limits.",
"symbols": [
"run",
"run_witness_bundle",
"validate_source_module_imports",
"_build_result",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING",
"SOURCE_MODULE_MANIFEST_NAME"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/formal_math_lean_proof_witness.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_math_lean_proof_witness.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_math_lean_proof_witness.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Formal Math Lean Proof Witness is the public runtime crossing from readiness metadata into a real local Lean/Lake subprocess witness. It copies a bounded public Lake project, records tool availability, Lake build status, source hashes, declaration names, line counts, source-module manifest verdicts, a Mathlib-blocked debt row, and six leakage/invalid-proof negative cases while keeping proof bodies, provider output bodies, oracle premise ids, and command output bodies out of result records.",
"summary_line": "The public Lean proof witness runs local Lean/Lake over a tiny synthetic project, validates copied public source-module digests and negative cases, and emits redacted result records without claiming general proof authority.",
"title": "Formal Math Lean Proof Witness"
},
{
"anchor": "paper-module-formal-math-verifier-trace-repair-loop",
"authority_ceiling": "Copied Ring2 verifier-trace repair metadata, source-module digests, public fixture result records, and one deterministic public toy-theorem rerun only; no Lean/Lake authority, formal-result correctness, proof body, oracle premise id, external model access, human approval as proof authority, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
Verifier failures as repair signals
\n
This is a source-available replay of one proof-lab pattern where a verifier's failure becomes a repair signal, with failure categories, per-failure repair actions, and a no-promotion-without-re-run rule all inspectable from copied data.formal_math_verifier_trace_repair_loop is a source-available replay of one pattern from a proof lab: a verifier reports why an attempt failed, and that failure is turned into a repair signal instead of being thrown away. The component works from copied, run data, so the failure categories, the repair action tied to each failure, and the rule that nothing is promoted without a fresh re-run are all inspectable.
\n
It is not a theorem prover or Lean runner: it validates copied Ring2 rows against their source digests, replays the verifier attempts and repair curriculum, and runs one small deterministic re-run over public inputs. It is not a theorem prover, a Lean or Lake runner, a provider client, or a private import. It validates a bundle of copied Ring2 benchmark rows, checks that the copies match their source digests, replays the verifier attempts and the repair curriculum against those source bodies, and runs one small deterministic re-run over public inputs. The output is a set of metadata-only result records.
\n
Proof and model-output bodies stay excluded
\n
Proofs, oracle answer keys, and model outputs are kept out on purpose, so a row can name a failure class without exposing the proof or oracle premise that produced it. Proofs, oracle answer keys, and model outputs are kept out on purpose. A row can name a failure class without exposing the proof or the oracle premise that produced it.
\n
Purpose
\n
Bookkeeping is where overclaim enters
\n
A failed attempt is cheap to discard and expensive to learn from, and the narrow question is whether its failure can become a reusable public repair signal without inheriting a real prover's authority. A failed proof attempt is cheap to discard and expensive to learn from. The narrow question here is whether a verifier's failure can become a reusable repair signal on the public side, without that signal quietly inheriting the authority of a real prover. The interesting work in a repair loop is the bookkeeping, and bookkeeping is where overclaim usually enters.
\n
A teaching signal needs five fields
\n
A verifier failure counts as a teaching signal only once it carries a trace-backed grade, a repair action named by failure class, a ledger entry, a curriculum edge, and a cold re-run reference. A loose version would collapse the stages into a single verdict: the attempt failed, we fixed it, done. This component refuses that. A verifier failure counts as a teaching signal only once it carries a trace grade backed by trace events, a repair action named against the verifier failure class it answers, an entry in the failure-mode ledger, a curriculum edge, and a cold re-run reference. Each of those is a separate field, and promotion is blocked until the cold re-run reference is present.
\n
Guarding against stale copied rows
\n
The guarded failure is stale copied rows pretending to be live proof-lab evidence, so the copies are re-checked against their sources on every run and the verdict is downgraded on any drift. The failure mode it guards against is stale copied rows pretending to be live proof-lab evidence. The rows are imported from one real Ring2 benchmark run, so the temptation is to treat the copy as if the run were happening now. The answer is to re-check the copies against their sources on every run and downgrade the verdict when anything drifts.
\n
How it works
\n
Two entrypoints, one builder
\n
run validates a fixture with the eight negative cases and run_loop_bundle validates an exported bundle under the stricter source-module path, both calling _build_result, which passes only with every check green and no negative case missing. There are two entrypoints. run validates a fixture directory with the eight negative-case files included. run_loop_bundle validates an exported bundle: it computes a freshness digest over the inputs, can reuse a current result record, and runs the stricter source-module path that requires every copied body to match its manifest. Both call _build_result, which composes the checks below in order and returns one record. A run is a pass only when every check passes and no negative case is missing.
\n
validate_projection_protocol demands provenance
\n
validate_projection_protocol requires each copied row to keep its source_ref, source_sha256, target and validation refs, and copied-material status, blocking thin or unprovenanced projections.validate_projection_protocol reads the projection manifest. It requires copied-material provenance: each copied row must keep its source_ref, source_sha256, target refs, validation refs, and a copied-material status, and any omitted proof, oracle, or provider material must carry an omission reference. Thin or unprovenanced projections are blocked.
\n
validate_source_module_manifest is the strict gate
\n
validate_source_module_manifest requires a import class, body_in_receipt: false, and a matching digest, line, and byte count per Ring2 source ref, emitting typed mismatch findings on any drift.validate_source_module_manifest is the strict gate, required for the exported bundle. It requires a import class, body_in_receipt: false, one row per declared Ring2 source ref, and for each row a matching digest, line count, and byte count against the copied file on disk. Drift produces typed findings such as VERIFIER_TRACE_SOURCE_MODULE_DIGEST_MISMATCH, VERIFIER_TRACE_SOURCE_MODULE_SIZE_MISMATCH, or VERIFIER_TRACE_SOURCE_MODULE_COUNT_MISMATCH. The exported bundle carries seven source modules and all seven must verify.
\n
validate_verifier_attempts replays each attempt
\n
validate_verifier_attempts checks each attempt's verifier class and trace events, then _validate_attempt_source_replay dereferences the premise, oracle-contrast, and candidate rows, counting attempts, events, actions, and promotions.validate_verifier_attempts checks each attempt for a verifier class and trace-event metadata, then hands the row to _validate_attempt_source_replay, which dereferences the premise-run row, the oracle-repair contrast row, and the graph-update candidate named by the attempt. Mismatches become VERIFIER_TRACE_SOURCE_REPLAY_MISMATCH, VERIFIER_TRACE_ORACLE_REPLAY_MISMATCH, VERIFIER_TRACE_COLD_RERUN_SOURCE_MISMATCH, or VERIFIER_TRACE_CANDIDATE_REPLAY_MISMATCH. It counts attempts, trace events, repair actions, and cold-rerun promotions.
\n
Curriculum and promotion policy checks
\n
validate_repair_curriculum rejects a ledger append accepted after a cold re-run with no reference, and validate_promotion_policy requires cold_rerun_receipt_ref and rejects a policy setting formal_proof_authority true.validate_repair_curriculum checks the failure-mode ledger and the curriculum edges, and rejects a ledger append marked accepted after a cold re-run that has no cold re-run reference. validate_promotion_policy requires cold_rerun_receipt_ref in the policy's promotion requirements and rejects a policy that sets formal_proof_authority to true.
\n
The one thing actually executed
\n
_run_toy_theorem_repair_rerun is the only executed step: over inputs 0, 1, 2, 7 the failing candidate returns constant 0 and the repair returns n, plain integer arithmetic, not a Lean proof._run_toy_theorem_repair_rerun is the one thing actually executed. It takes the toy statement toy_add_zero_identity over the inputs 0, 1, 2, and 7. The failing candidate sets the right side to the constant 0, which fails on three of the four inputs. The repair sets the right side to n, which passes on all four. It is plain integer arithmetic over public inputs through _eval_toy_expr, not a Lean proof.
\n
Grading the run from R1 to R4
\n
_runtime_realness_evidence reaches R4 only for a fully verified bundle with 30-plus clean source-replay checks and a passing toy re-run, dropping to R3 on any mismatch, R2 for fixtures, and R1 otherwise._runtime_realness_evidence grades the run. It reaches R4 only for an exported bundle with all seven source modules verified, at least 30 source-replay checks with zero mismatches, at least three attempts, at least nine trace events, at least three failure modes, and a passing toy re-run. A verified bundle with any source-replay mismatch drops to R3 (mutated trace rejected). A fixture with copied material and observed negative cases sits at R2. Everything else is R1, the metadata floor. _write_receipts then emits the result record, a board, a validation record, and a fixture-pass record, all metadata-only.
\nDiagram of the mechanism (9 steps).Diagram source & refs
Source refs
copied-material provenance
validate_projection_protocol
digest, line, byte match; body_in_receipt false
validate_source_module_manifest
trace events, verifier class, source replay
validate_verifier_attempts
failure-mode ledger, curriculum edges
validate_repair_curriculum
requires cold re-run ref
validate_promotion_policy
flowchart TD Input["Fixture or exported bundle copied Ring2 rows + source manifest"] Protocol["validate_projection_protocol copied-material provenance"] Manifest["validate_source_module_manifest digest, line, byte match; body_in_receipt false"] Attempts["validate_verifier_attempts trace events, verifier class, source replay"] Curriculum["validate_repair_curriculum failure-mode ledger, curriculum edges"] Promotion["validate_promotion_policy requires cold re-run ref"] Toy["_run_toy_theorem_repair_rerun fail on 0, repair to n"] Realness["_runtime_realness_evidence clean replay to R4; any drift downgrades"] Result records["_write_receipts result, board, validation, fixture-pass"] Input --> Protocol Protocol --> Manifest Manifest --> Attempts Attempts --> Curriculum Curriculum --> Promotion Promotion --> Toy Toy --> Realness Realness --> Result records
\n
What the shipped bundle produces
\n
On the shipped exported bundle this produces five attempts, 15 trace events, five repair actions, three cold-rerun promotions, seven verified source modules, 37 clean source-replay checks, and an R4 grade. On the shipped exported bundle this produces five attempts, 15 trace events, five repair actions, three cold-rerun promotions, three failure modes, three curriculum edges, seven verified source modules, 37 source-replay checks with zero mismatches, and an R4 grade.
\n
Negative cases
\n
The fixture path carries eight named negative files, and EXPECTED_NEGATIVE_CASES requires each to be observed before a pass. The fixture path carries eight named negative files, and EXPECTED_NEGATIVE_CASES requires each to be observed before a pass. Each maps to one error code:
A repair trace that sets a theorem-correctness flag is rejected, so the loop can say what to try next without saying the theorem is true, and mutations downgrade the grade rather than passing quietly. The last case is why repair metadata cannot become a truth claim: a repair trace that sets a theorem-correctness flag is rejected, so the loop can say what to try next without saying the theorem is true. The focused tests also perturb an oracle source row, a manifest digest, an attempt source label, and a curriculum source count, and confirm each mutation downgrades the grade rather than passing quietly.
\n
Prior Art Grounding
\n
Proof-attempt feedback loops
\n
The component follows interactive theorem-proving feedback loops like GamePad, HOList, and LeanDojo, borrowing the accounting shape of separate verifier-event, grade, failure-class, repair, and cold-rerun fields without treating advice as correctness. This component follows interactive theorem-proving feedback loops where failed proof attempts become structured training or repair signals. GamePad and HOList expose theorem-proving interaction data for machine-learning experiments, and LeanDojo keeps proof-assistant feedback, retrieval, and proof-state interaction reproducible. What is borrowed is the accounting shape: verifier events, grades, failure classes, repair actions, curriculum deltas, and cold re-run references stay separate fields. Human or provider advice is not treated as formal-result correctness.
\n
Validation Result record Path
\n
Run these from the repo root without mutating durable result record surfaces:
A pass means the fixture and bundle replay hold, the eight negative cases are observed, the copied bodies match their digests, and the row reports edge_count: 17, but no theorem is proved. A pass means the fixture and exported-bundle replay hold, the eight negative cases are observed, the copied bodies match their digests, and the generated row reports edge_count: 17, source_authority: json_capsule, and zero unresolved selective relations. It does not mean any theorem is proved.
\n
Scope boundary
\n
Scope limit
\n
The narrow proof boundary
\n
Copied Ring2 rows can drive a public repair loop that demands trace events before a grade, routes repair by failure class, gates promotion on a cold re-run, and replays each copy against its digest. The strongest claim the evidence supports is narrow: copied, Ring2 verifier rows can drive a public repair loop that requires trace events before a grade, routes a repair action by failure class, gates promotion on a cold re-run reference, replays every copy against its source digest, and emits metadata-only result records. The proof boundary is copied metadata, the source digests, the fixture records, and one deterministic toy re-run over public integers.
\n
What the loop refuses
\n
It runs no Lean or Lake, calls no provider, exposes no proof body or oracle id, treats no advice as proof authority, proves no formal-result correctness, and approves no launch. It refuses more. It does not run Lean or Lake, call a provider, expose a proof body or an oracle premise id, or treat human or provider advice as proof authority. It does not establish formal-result correctness, and a repair trace that claims it is rejected. It does not approve launch or certify whole-system correctness. The scope limit is repair-loop accounting over real run rows, stated once here.
\n",
"components": [
{
"anchor": "component-formal_math_verifier_trace_repair_loop",
"component_id": "formal_math_verifier_trace_repair_loop",
"public_label": "Formal Math Verifier Trace Repair Loop"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 4,
"principle": 5
},
"link_count": 16,
"links": [
{
"id": "formal_math_verifier_trace_repair_loop",
"kind": "organ",
"kind_label": "component",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "explains",
"url": "component-formal-math-verifier-trace-repair-loop.html"
},
{
"id": "mechanism.formal_math_verifier_trace_repair_loop.validates_public_verifier_trace_repair_bundle",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public verifier trace repair bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-verifier-trace-repair-loop-validates-public-verifier-trace-repair-bundle"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.lean_std_premise_index",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Lean/Std Premise Index",
"relation": "depends on",
"url": "paper-module-lean-std-premise-index.html"
},
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depends on",
"url": "paper-module-tactic-portfolio-availability.html"
},
{
"id": "paper_module.target_shape_tactic_routing",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Target Shape Tactic Routing",
"relation": "depends on",
"url": "paper-module-target-shape-tactic-routing.html"
},
{
"id": "paper_module.formal_math_premise_retrieval",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Premise Retrieval",
"relation": "depends on",
"url": "paper-module-formal-math-premise-retrieval.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.formal_evidence_cell_anchor_resolver",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Evidence Cell Anchor Resolver",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[24:paper_module.formal_evidence_cell_anchor_resolver].depends_on",
"url": "paper-module-formal-evidence-cell-anchor-resolver.html"
},
{
"id": "paper_module.proof_diagnostic_evidence_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof Diagnostic Evidence Spine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[14:paper_module.proof_diagnostic_evidence_spine].depends_on",
"url": "paper-module-proof-diagnostic-evidence-spine.html"
},
{
"id": "paper_module.target_shape_tactic_routing",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Target Shape Tactic Routing",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[41:paper_module.target_shape_tactic_routing].depends_on",
"url": "paper-module-target-shape-tactic-routing.html"
},
{
"id": "paper_module.verifier_lab_execution_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Execution Spine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[44:paper_module.verifier_lab_execution_spine].depends_on",
"url": "paper-module-verifier-lab-execution-spine.html"
},
{
"id": "paper_module.verifier_lab_kernel",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Kernel",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[0:paper_module.verifier_lab_kernel].depends_on",
"url": "paper-module-verifier-lab-kernel.html"
}
],
"paper_module_id": "paper_module.formal_math_verifier_trace_repair_loop",
"reader_packet": {
"abstract": "The public verifier-trace repair fixture validates copied Ring2 failure taxonomy, graph-update, oracle-repair contrast rows, source-module digests, negative cases, and one deterministic toy rerun without claiming proof authority.",
"authority_ceiling": "Copied Ring2 verifier-trace repair metadata, source-module digests, public fixture result records, and one deterministic public toy-theorem rerun only; no Lean/Lake authority, formal-result correctness, proof body, oracle premise id, external model access, human approval as proof authority, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.formal_math_verifier_trace_repair_loop.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-formal-math-verifier-trace-repair-loop-diagram-1",
"html_url": "docs/paper-module-formal-math-verifier-trace-repair-loop.html#paper-module-formal-math-verifier-trace-repair-loop-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Input[\"Fixture or exported bundle copied Ring2 rows + source manifest\"] Protocol[\"validate_projection_protocol copied-material provenance\"] Manifest[\"validate_source_module_manifest digest, line, byte match; body_in_receipt false\"] Attempts[\"validate_verifier_attempts trace events, verifier class, source replay\"] Curriculum[\"validate_repair_curriculum failure-mode ledger, curriculum edges\"] Promotion[\"validate_promotion_policy requires cold re-run ref\"] Toy[\"_run_toy_theorem_repair_rerun fail on 0, repair to n\"] Realness[\"_runtime_realness_evidence clean replay to R4; any drift downgrades\"] Result records[\"_write_receipts result, board, validation, fixture-pass\"] Input --> Protocol Protocol --> Manifest Manifest --> Attempts Attempts --> Curriculum Curriculum --> Promotion Promotion --> Toy Toy --> Realness Realness --> Result records",
"source_format": "mermaid",
"source_hash": "sha256:77d4b40456a3344baa8f754aca16d23d24fa4c835624fbdfd81059beb166d595",
"source_line_end": 177,
"source_line_start": 157
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
}
],
"relation": "abides by"
},
{
"count": 4,
"links": [
{
"id": "paper_module.lean_std_premise_index",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Lean/Std Premise Index",
"relation": "depends on",
"url": "paper-module-lean-std-premise-index.html"
},
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depends on",
"url": "paper-module-tactic-portfolio-availability.html"
},
{
"id": "paper_module.target_shape_tactic_routing",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Target Shape Tactic Routing",
"relation": "depends on",
"url": "paper-module-target-shape-tactic-routing.html"
},
{
"id": "paper_module.formal_math_premise_retrieval",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Premise Retrieval",
"relation": "depends on",
"url": "paper-module-formal-math-premise-retrieval.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "formal_math_verifier_trace_repair_loop",
"kind": "organ",
"kind_label": "component",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "explains",
"url": "component-formal-math-verifier-trace-repair-loop.html"
},
{
"id": "mechanism.formal_math_verifier_trace_repair_loop.validates_public_verifier_trace_repair_bundle",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public verifier trace repair bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-verifier-trace-repair-loop-validates-public-verifier-trace-repair-bundle"
}
],
"relation": "explains"
},
{
"count": 6,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.formal_math_verifier_trace_repair_loop",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 277,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_math_verifier_trace_repair_loop.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "GamePad",
"section_id": "proof-attempt-feedback-loops",
"url": "https://arxiv.org/abs/1806.00608"
},
{
"kind": "external_reference",
"label": "HOList",
"section_id": "proof-attempt-feedback-loops",
"url": "https://arxiv.org/abs/1904.03241"
},
{
"kind": "external_reference",
"label": "LeanDojo",
"section_id": "proof-attempt-feedback-loops",
"url": "https://arxiv.org/abs/2306.15626"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This is a source-available replay of one proof-lab pattern where a verifier's failure becomes a repair signal, with failure categories, per-failure repair actions, and a no-promotion-without-re-run rule all inspectable from copied data. formalmathverifiertracerepairloop is a source-available replay of one pattern from a proof lab: a verifier reports why an attempt failed, and that failure is turned into a repair signal instead of being thrown away. The component works from copied, run data, so the failure categories, the repair action tied to each failure, and the rule that nothing is promoted without a fresh re-run are all inspectable.",
"level": 4,
"order": 1,
"section_id": "verifier-failures-as-repair-signals",
"source_line_end": 14,
"source_line_start": 3,
"title": "Verifier failures as repair signals"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is not a theorem prover or Lean runner: it validates copied Ring2 rows against their source digests, replays the verifier attempts and repair curriculum, and runs one small deterministic re-run over public inputs.** It is not a theorem prover, a Lean or Lake runner, a provider client, or a private import. It validates a bundle of copied Ring2 benchmark rows, checks that the copies match their source digests, replays the verifier attempts and the repair curriculum against those source bodies, and runs one small deterministic re-run over public inputs. The output is a set of metadata-only result records.",
"level": 4,
"order": 2,
"section_id": "it-is-not-a-theorem-prover-or-lean-runner",
"source_line_end": 25,
"source_line_start": 15,
"title": "It is not a theorem prover or Lean runner"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Proofs, oracle answer keys, and model outputs are kept out on purpose, so a row can name a failure class without exposing the proof or oracle premise that produced it.** Proofs, oracle answer keys, and model outputs are kept out on purpose. A row can name a failure class without exposing the proof or the oracle premise that produced it.",
"level": 4,
"order": 3,
"section_id": "proof-and-model-output-bodies-stay-excluded",
"source_line_end": 33,
"source_line_start": 26,
"title": "Proof and model-output bodies stay excluded"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 35,
"source_line_start": 34,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A failed attempt is cheap to discard and expensive to learn from, and the narrow question is whether its failure can become a reusable public repair signal without inheriting a real prover's authority.** A failed proof attempt is cheap to discard and expensive to learn from. The narrow question here is whether a verifier's failure can become a reusable repair signal on the public side, without that signal quietly inheriting the authority of a real prover. The interesting work in a repair loop is the bookkeeping, and bookkeeping is where overclaim usually enters.",
"level": 4,
"order": 5,
"section_id": "bookkeeping-is-where-overclaim-enters",
"source_line_end": 45,
"source_line_start": 36,
"title": "Bookkeeping is where overclaim enters"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A verifier failure counts as a teaching signal only once it carries a trace-backed grade, a repair action named by failure class, a ledger entry, a curriculum edge, and a cold re-run reference.** A loose version would collapse the stages into a single verdict: the attempt failed, we fixed it, done. This component refuses that. A verifier failure counts as a teaching signal only once it carries a trace grade backed by trace events, a repair action named against the verifier failure class it answers, an entry in the failure-mode ledger, a curriculum edge, and a cold re-run reference. Each of those is a separate field, and promotion is blocked until the cold re-run reference is present.",
"level": 4,
"order": 6,
"section_id": "a-teaching-signal-needs-five-fields",
"source_line_end": 57,
"source_line_start": 46,
"title": "A teaching signal needs five fields"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The guarded failure is stale copied rows pretending to be live proof-lab evidence, so the copies are re-checked against their sources on every run and the verdict is downgraded on any drift.** The failure mode it guards against is stale copied rows pretending to be live proof-lab evidence. The rows are imported from one real Ring2 benchmark run, so the temptation is to treat the copy as if the run were happening now. The answer is to re-check the copies against their sources on every run and downgrade the verdict when anything drifts.",
"level": 4,
"order": 7,
"section_id": "guarding-against-stale-copied-rows",
"source_line_end": 67,
"source_line_start": 58,
"title": "Guarding against stale copied rows"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 69,
"source_line_start": 68,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run validates a fixture with the eight negative cases and runloop_bundle validates an exported bundle under the stricter source-module path, both calling buildresult, which passes only with every check green and no negative case missing.** There are two entrypoints. run validates a fixture directory with the eight negative-case files included. runloopbundle validates an exported bundle: it computes a freshness digest over the inputs, can reuse a current result record, and runs the stricter source-module path that requires every copied body to match its manifest. Both call buildresult, which composes the checks below in order and returns one record. A run is a pass only when every check pass\u2026",
"level": 4,
"order": 9,
"section_id": "two-entrypoints-one-builder",
"source_line_end": 82,
"source_line_start": 70,
"title": "Two entrypoints, one builder"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateprojectionprotocol requires each copied row to keep its sourceref, source_sha256, target and validation refs, and copied-material status, blocking thin or unprovenanced projections. validateprojection_protocol reads the projection manifest. It requires copied-material provenance: each copied row must keep its sourceref, sourcesha256, target refs, validation refs, and a copied-material status, and any omitted proof, oracle, or provider material must carry an omission reference. Thin or unprovenanced projections are blocked.",
"level": 4,
"order": 10,
"section_id": "validate-projection-protocol-demands-provenance",
"source_line_end": 92,
"source_line_start": 83,
"title": "`validate_projection_protocol` demands provenance"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatesourcemodulemanifest requires a import class, bodyinreceipt: false, and a matching digest, line, and byte count per Ring2 source ref, emitting typed mismatch findings on any drift.** validatesourcemodule_manifest is the strict gate, required for the exported bundle. It requires a import class, bodyinreceipt: false, one row per declared Ring2 source ref, and for each row a matching digest, line count, and byte count against the copied file on disk. Drift produces typed findings such as VERIFIERTRACESOURCEMODULEDIGEST_MISMATCH, VERIFIERTRACESOURCEMODULESIZE_MISMATCH, or VERIFIERTRACESOURCEMODULECOUNT_MISMATCH. The exported bundle carries seven source modules and all seven must verify.",
"level": 4,
"order": 11,
"section_id": "validate-source-module-manifest-is-the-strict-gate",
"source_line_end": 106,
"source_line_start": 93,
"title": "`validate_source_module_manifest` is the strict gate"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateverifier_attempts checks each attempt's verifier class and trace events, then validateattemptsourcereplay dereferences the premise, oracle-contrast, and candidate rows, counting attempts, events, actions, and promotions. validateverifier_attempts checks each attempt for a verifier class and trace-event metadata, then hands the row to validateattemptsourcereplay, which dereferences the premise-run row, the oracle-repair contrast row, and the graph-update candidate named by the attempt. Mismatches become VERIFIERTRACESOURCEREPLAYMISMATCH, VERIFIERTRACEORACLEREPLAYMISMATCH, VERIFIERTRACECOLDRERUNSOURCE_MISMATCH, or VERIFIERTRACECANDIDATEREPLAYMISMATCH. It counts attempts, trace events,\u2026",
"level": 4,
"order": 12,
"section_id": "validate-verifier-attempts-replays-each-attempt",
"source_line_end": 120,
"source_line_start": 107,
"title": "`validate_verifier_attempts` replays each attempt"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validaterepair_curriculum rejects a ledger append accepted after a cold re-run with no reference, and validatepromotionpolicy requires coldrerunreceipt_ref and rejects a policy setting formalproofauthority true.** validaterepaircurriculum checks the failure-mode ledger and the curriculum edges, and rejects a ledger append marked accepted after a cold re-run that has no cold re-run reference. validatepromotionpolicy requires coldrerunreceipt_ref in the policy's promotion requirements and rejects a policy that sets formalproofauthority to true.",
"level": 4,
"order": 13,
"section_id": "curriculum-and-promotion-policy-checks",
"source_line_end": 131,
"source_line_start": 121,
"title": "Curriculum and promotion policy checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "runtoytheoremrepair_rerun is the only executed step: over inputs 0, 1, 2, 7 the failing candidate returns constant 0 and the repair returns n, plain integer arithmetic, not a Lean proof. runtoytheoremrepairrerun is the one thing actually executed. It takes the toy statement toyaddzero_identity over the inputs 0, 1, 2, and 7. The failing candidate sets the right side to the constant 0, which fails on three of the four inputs. The repair sets the right side to n, which passes on all four. It is plain integer arithmetic over public inputs through evaltoy_expr, not a Lean proof.",
"level": 4,
"order": 14,
"section_id": "the-one-thing-actually-executed",
"source_line_end": 142,
"source_line_start": 132,
"title": "The one thing actually executed"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "runtimerealness_evidence reaches R4 only for a fully verified bundle with 30-plus clean source-replay checks and a passing toy re-run, dropping to R3 on any mismatch, R2 for fixtures, and R1 otherwise. runtimerealnessevidence grades the run. It reaches R4 only for an exported bundle with all seven source modules verified, at least 30 source-replay checks with zero mismatches, at least three attempts, at least nine trace events, at least three failure modes, and a passing toy re-run. A verified bundle with any source-replay mismatch drops to R3 (mutated trace rejected). A fixture with copied material and observed negative cases sits at R2. Everything else is R1, the metadata floor. writerece\u2026",
"level": 4,
"order": 15,
"section_id": "grading-the-run-from-r1-to-r4",
"source_line_end": 178,
"source_line_start": 143,
"title": "Grading the run from R1 to R4"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**On the shipped exported bundle this produces five attempts, 15 trace events, five repair actions, three cold-rerun promotions, seven verified source modules, 37 clean source-replay checks, and an R4 grade.** On the shipped exported bundle this produces five attempts, 15 trace events, five repair actions, three cold-rerun promotions, three failure modes, three curriculum edges, seven verified source modules, 37 source-replay checks with zero mismatches, and an R4 grade.",
"level": 4,
"order": 16,
"section_id": "what-the-shipped-bundle-produces",
"source_line_end": 187,
"source_line_start": 179,
"title": "What the shipped bundle produces"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "negative-cases",
"source_line_end": 189,
"source_line_start": 188,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The fixture path carries eight named negative files, and EXPECTEDNEGATIVECASES requires each to be observed before a pass. The fixture path carries eight named negative files, and EXPECTEDNEGATIVECASES requires each to be observed before a pass. Each maps to one error code: attemptwithproofbody produces VERIFIERTRACEPROOFBODY_FORBIDDEN. attemptwithoracleids produces VERIFIERTRACEORACLEIDS_FORBIDDEN. tracegradewithouttrace produces VERIFIERTRACEGRADEWITHOUT_TRACE. repairwithoutverifierclass produces VERIFIERREPAIRWITHOUTVERIFIER_CLASS. promotionwithoutcoldrerun produces VERIFIERPROMOTIONWITHOUTCOLD_RERUN. providerpayloadleakage produces VERIFIERPROVIDERPAYLOAD_FORBIDDEN. humanapprovalasproof\u2026",
"level": 4,
"order": 18,
"section_id": "the-fixture-path-carries-eight-named-negative-files",
"source_line_end": 204,
"source_line_start": 190,
"title": "The fixture path carries eight named negative files"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A repair trace that sets a theorem-correctness flag is rejected, so the loop can say what to try next without saying the theorem is true, and mutations downgrade the grade rather than passing quietly.** The last case is why repair metadata cannot become a truth claim: a repair trace that sets a theorem-correctness flag is rejected, so the loop can say what to try next without saying the theorem is true. The focused tests also perturb an oracle source row, a manifest digest, an attempt source label, and a curriculum source count, and confirm each mutation downgrades the grade rather than passing quietly.",
"level": 4,
"order": 19,
"section_id": "why-repair-metadata-is-not-a-truth-claim",
"source_line_end": 215,
"source_line_start": 205,
"title": "Why repair metadata is not a truth claim"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "prior-art-grounding",
"source_line_end": 217,
"source_line_start": 216,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component follows interactive theorem-proving feedback loops like GamePad, HOList, and LeanDojo, borrowing the accounting shape of separate verifier-event, grade, failure-class, repair, and cold-rerun fields without treating advice as correctness.** This component follows interactive theorem-proving feedback loops where failed proof attempts become structured training or repair signals. GamePad and HOList expose theorem-proving interaction data for machine-learning experiments, and LeanDojo keeps proof-assistant feedback, retrieval, and proof-state interaction reproducible. What is borrowed is the accounting shape: verifier events, grades, failure classes, repair actions, curriculum d\u2026",
"level": 4,
"order": 21,
"section_id": "proof-attempt-feedback-loops",
"source_line_end": 232,
"source_line_start": 218,
"title": "Proof-attempt feedback loops"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run these from the repo root without mutating durable result record surfaces:",
"level": 2,
"order": 22,
"section_id": "validation-receipt-path",
"source_line_end": 242,
"source_line_start": 233,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the fixture and bundle replay hold, the eight negative cases are observed, the copied bodies match their digests, and the row reports edgecount: 17, but no theorem is proved. A pass means the fixture and exported-bundle replay hold, the eight negative cases are observed, the copied bodies match their digests, and the generated row reports edge_count: 17, sourceauthority: jsoncapsule, and zero unresolved selective relations. It does not mean any theorem is proved.",
"level": 4,
"order": 23,
"section_id": "what-a-pass-means-here",
"source_line_end": 252,
"source_line_start": 243,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 24,
"section_id": "claim-ceiling",
"source_line_end": 254,
"source_line_start": 253,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Copied Ring2 rows can drive a public repair loop that demands trace events before a grade, routes repair by failure class, gates promotion on a cold re-run, and replays each copy against its digest.** The strongest claim the evidence supports is narrow: copied, Ring2 verifier rows can drive a public repair loop that requires trace events before a grade, routes a repair action by failure class, gates promotion on a cold re-run reference, replays every copy against its source digest, and emits metadata-only result records. The proof boundary is copied metadata, the source digests, the fixture records, and one deterministic toy re-run over public integers.",
"level": 4,
"order": 25,
"section_id": "the-narrow-proof-boundary",
"source_line_end": 266,
"source_line_start": 255,
"title": "The narrow proof boundary"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It runs no Lean or Lake, calls no provider, exposes no proof body or oracle id, treats no advice as proof authority, proves no formal-result correctness, and approves no launch.** It refuses more. It does not run Lean or Lake, call a provider, expose a proof body or an oracle premise id, or treat human or provider advice as proof authority. It does not establish formal-result correctness, and a repair trace that claims it is rejected. It does not approve launch or certify whole-system correctness. The scope limit is repair-loop accounting over real run rows, stated once here.",
"level": 4,
"order": 26,
"section_id": "what-the-loop-refuses",
"source_line_end": 276,
"source_line_start": 267,
"title": "What the loop refuses"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/formal_math_verifier_trace_repair_loop.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public verifier-trace repair fixture validates copied Ring2 failure taxonomy, graph-update, oracle-repair contrast rows, source-module digests, negative cases, and one deterministic toy rerun without claiming proof authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "formal_math_verifier_trace_repair_loop.py",
"path": "src/microcosm_core/organs/formal_math_verifier_trace_repair_loop.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public verifier-trace repair fixture validation, exported bundle validation, copied source-module manifest checks, negative-case enforcement, toy rerun evidence, and scope limits.",
"symbols": [
"run",
"run_loop_bundle",
"validate_source_module_manifest",
"_build_result",
"_write_receipts",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING",
"SOURCE_MODULE_MANIFEST_REF"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/formal_math_verifier_trace_repair_loop.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_math_verifier_trace_repair_loop.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_math_verifier_trace_repair_loop.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Formal Math Verifier Trace Repair Loop is the public evidence membrane for proof-lab repair mechanics. It checks copied Ring2 run refs and digests, verifier attempts, trace grades, repair actions, promotion gates, source-module manifests, secret exclusion, and eight leakage or overclaim negative cases, then writes metadata-only result records that keep proof bodies, oracle premise ids, model-output data, repair-trace theorem-correctness claims, Lean/Lake execution, and formal proof authority out of scope.",
"summary_line": "The public verifier-trace repair fixture validates copied Ring2 failure taxonomy, graph-update, oracle-repair contrast rows, source-module digests, negative cases, and one deterministic toy rerun without claiming proof authority.",
"title": "Formal Math Verifier Trace Repair Loop"
},
{
"anchor": "paper-module-formal-evidence-cell-anchor-resolver",
"authority_ceiling": "Evidence-cell anchor metadata and source-open runtime result records only; no formal-result correctness, proof-body import, private source-ref authority, human approval as proof authority, Lean/Lake execution, external model access, launch-scope decision, publishing-scope decision, source-file changes, or formal-proof certification.",
"body_html": "
Component asks whether
\n
This component asks whether a page's proof-language claim can be tied to a named piece of public evidence, resolving each claim to an evidence cell and capping its strength at what the cell allows.formal_evidence_cell_anchor_resolver checks the project's proof-flavoured writing against recorded evidence. When a paper page says a claim is proved or certified, this component asks a narrower question: can the words be tied to a named piece of public evidence, and does that evidence stay below formal-result correctness. It answers by resolving each claim to an evidence cell, checking the cell's source anchors point at files that exist on the public path, and capping the claim's strength at what the cell allows.
\n
Unit of evidence
\n
The unit of evidence is a cell, a stable id bundling source-anchor refs, a machine_anchor_class, and allowed claim strengths, and the resolver reports which claims resolved, overreached, or had no backing. The unit of evidence is a cell. A cell is a stable id that stands for a bundle of recorded evidence: a list of source-anchor refs, a machine_anchor_class naming what kind of machine artifact backs it, and the claim strengths the cell is allowed to support. The resolver reads paper claims, an evidence-cell registry, a claim-boundary policy, and an obligation graph, then writes a result record that reports which claims resolved, which overreached, and which had no backing. Proof bodies, non-public source refs, and model-output data stay out of that record.
\n
Is not theorem
\n
It is not a theorem prover: it runs over already-public inputs, emits metadata-only result records, and states its full proof boundary once under Scope limit. It is not a theorem prover. It runs over already-public inputs and emits metadata-only result records. The full proof boundary is stated once, under Scope limit.
\n
Purpose
\n
Where proof-adjacent prose drifts
\n
Proof-adjacent prose drifts easily, and a loose check that only scanned for the word proof would pass any sentence avoiding it and could not bound how strong a backed claim may be. Proof-adjacent prose is the easiest place for a claim to drift. A page can write "this proves the theorem" and a reader has no cheap way to tell whether the words rest on a checked artifact or on nothing. A loose check that only scanned for the word "proof" would pass any sentence that avoided the word, and would have no way to bound how strong a backed claim is allowed to be.
\n
proof_language_requires_machine_anchor
\n
The key policy is proof_language_requires_machine_anchor: proof-language claims must name a resolving cell whose anchors point at existing files, with strength capped and human approval refused as a substitute. The policy that makes the component useful is proof_language_requires_machine_anchor. A claim that uses proof language must name a cell, the cell must resolve in the registry, and the cell's source anchors must point at files that exist. A claim that uses proof language but names no cell, or names a cell absent from the registry, lowers the run to a blocked status instead of passing as green prose. On top of the resolved cell sit two limits: the claim's strength is capped by the cell's allowed_claim_strengths, and human approval is refused as a substitute for a machine anchor.
\n
How it works
\n
Two entrypoints share one core
\n
run validates a fixture directory with negative cases included and run_anchor_bundle validates an exported bundle with them off, both calling _build_result to run the pipeline. Two entrypoints share one core. run validates a fixture directory with the negative cases included. run_anchor_bundle validates an exported bundle directory with negative cases off and the source-module manifest required. Both call _build_result, which runs the pipeline and returns the result record.
\n
_build_result scans, validates, folds
\n
_build_result loads the five inputs, scans every path for secret-equivalent material, runs six validators, and passes only when all pass, every expected negative case was observed, and no scan hit fired._build_result loads the five fixture inputs with _load_payloads, scans every input path for secret-equivalent material with scan_paths, then runs six validators and folds their findings into one status. The run passes only when every validator passes, every expected negative case was observed, and no secret scan hit fired.
\n
The validators, in the order _build_result calls them:
\n
Function
What it checks
validate_projection_protocol
The projection cites enough source refs, pattern ids, and public runtime refs, and its Ring2 anchor refs resolve on the public path.
validate_cell_registry
Every evidence cell has at least one public source anchor, exposes no non-public refs, and records a machine_anchor_class.
validate_claim_boundary_policy
The policy sets proof_language_requires_machine_anchor, declares claim-strength levels and oracle-boundary classes, and excludes theorem-correctness claims or advisory export.
validate_obligation_graph
Declared obligations are met and carry no non-public refs, proof body, or oracle payload.
validate_source_module_manifest
For an exported bundle, the frozen source-module snapshots exist, match their recorded digests, and cover every declared owner ref.
validate_claims
Each claim resolves to a known cell, stays inside the cell's allowed strengths, and trips no leakage or overclaim rule.
\n
validate_claims resolves each claim
\n
validate_claims passes each claim to _inspect_claim_row, which rejects proof bodies, non-public refs, oracle payloads, missing boundary classes, unknown cells, and out-of-range strengths as structured findings.validate_claims is the resolver proper. It builds a cell_by_id map from the registry and passes each claim to _inspect_claim_row. That function rejects a claim when it carries a proof body, a non-public source refs, or a provider or oracle payload; when it omits an oracle-boundary class; when it uses proof language with no cell; when its cell id is unknown; or when its stated strength is not in the cell's allowed list. A theorem-correctness overclaim is caught either from an explicit flag or from the claim text by THEOREM_CORRECTNESS_OVERCLAIM_RE, a regex that matches "proves" or "certifies" near "theorem" and "correctness". Each rejection is a structured finding, not a silent pass.
\n
Resolving source anchors on the public path
\n
_validate_anchor_refs splits each ref, rejects private schemes and escaping paths, resolves the file among public candidates including a source-artifact mirror, and records digest matches without holding file bodies. Source anchors resolve through _validate_anchor_refs. It splits each ref into a path and an optional marker, rejects private schemes and absolute or parent paths through _public_ref_is_forbidden, then looks for the file among the candidates from _public_ref_candidates. A state/runs/... anchor is allowed to resolve through a public source-artifact mirror under examples/*/*/source_artifacts, so the resolver can confirm an anchor exists from an existing public example bundle without copying wider source trees into this component. When a ref carries a digest, _sha256_file recomputes it and the row records whether it matched. The record holds refs, digests, and status, not file bodies.
\n
metadata-only card shaping
\n
result_card, _source_module_summary_card, and _board_from_result shape the result into compact cards, each carrying body_in_receipt: False. The result record is shaped by result_card, _source_module_summary_card, and _board_from_result into compact cards. Every card and finding carries body_in_receipt: False.
\nDiagram of the mechanism (7 steps).Diagram source & refs
flowchart TD Inputs["paper claims, cell registry, boundary policy, obligation graph, projection protocol"] --> Build["_build_result"] Build --> Scan["scan_paths secret-equivalent material"] Build --> Validators["six validators projection, registry, policy, obligations, source modules, claims"] Validators --> Claims["validate_claims _inspect_claim_row"] Claims --> Status["status = pass only if all validators pass and every negative case observed"] Scan --> Status Status --> Record["metadata-only result record refs, digests, findings"]
\n
Negative cases
\n
Thirteen required negative-case classes
\n
EXPECTED_NEGATIVE_CASES declares thirteen negative-case classes the fixture run must observe, and a run fails if any is missing, keeping the refusal coverage from silently eroding.EXPECTED_NEGATIVE_CASES declares thirteen negative-case classes that the fixture run must observe. A run fails if any of them is missing, which keeps the refusal coverage from silently eroding. The classes name the exact rejection each case exercises: unknown_cell_overclaim, missing_source_anchor, proof_language_without_cell, private_source_ref_leakage, proof_body_leakage, theorem_correctness_overclaim, human_approval_as_evidence_cell, oracle_payload_leakage, oracle_boundary_forward_success_overclaim, and four obligation-graph variants for unmet obligations and private-ref, proof-body, or oracle-payload leakage inside an obligation.
\n
Prior Art Grounding
\n
Provenance and proof-certificate lineage
\n
The component follows provenance and proof-certificate work like W3C PROV and Proof-Carrying Code, borrowing anchor resolution so a small checker bounds what cell metadata can assert without granting theorem authority. This component follows provenance and proof-certificate work where a claim must point at checkable evidence rather than untyped narrative. The W3C PROV model links entities, activities, and agents in an evidence graph. Proof-Carrying Code and small-kernel proof assistants motivate keeping a certificate or anchor separate from the trusted checker that bounds its meaning. This resolver borrows the anchor-resolution pattern: proof-language claims must name cell ids, source anchors, and machine-anchor classes, and a small checker bounds what those metadata can assert. It does not turn metadata cells into theorem-correctness authority.
\n
Validation Result record Path
\n
Run from repo
\n
Run these from the repo root. Run these from the repo root. They do not mutate durable result record or generated projection surfaces.
A pass means the fixture run observed all thirteen negative cases, resolved three claims to three cells across eight anchors, validated six frozen snapshots by digest, and kept everything metadata-only. A pass means the fixture run observed all thirteen negative cases, resolved three paper claims to three evidence cells across eight source anchors and three machine anchors, validated the exported bundle against six frozen source-module snapshots by digest, and kept every card and finding metadata-only. It is bounded validator evidence over that fixture, not a statement about the wider formal-math corpus.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
Public evidence-cell metadata can bind proof-language claims to recorded digest-checked cells, cap each claim at the strength its cell allows, and reject claims that leak private material or overreach. The strongest claim the evidence supports: public evidence-cell metadata can bind proof-language claims to recorded, digest-checked cells, cap each claim at the strength its cell allows, and reject claims that leak private material or overreach into formal-result correctness. The counts are the fixture's contract, not a scale claim.
\n
The proof boundary is anchor metadata and source-open runtime records only: this runs no Lean, proves no formal-result correctness, and a digest match confirms only that a snapshot equals its recorded hash. The proof boundary is anchor metadata and source-open runtime records only. This component does not run Lean or Lake, prove formal-result correctness, import proof bodies, expose non-public source refs, carry provider or oracle payloads, treat human approval as proof, or include launch operations or public sharing. A digest match confirms a snapshot equals its recorded hash; it does not certify that the snapshot equals the latest live source, that every source file was imported, or that the copied bodies may be released. AUTHORITY_CEILING records these limits as explicit false flags in every result record.
\n",
"components": [
{
"anchor": "component-formal_evidence_cell_anchor_resolver",
"component_id": "formal_evidence_cell_anchor_resolver",
"public_label": "Formal Evidence Cell Anchor Resolver"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 2,
"organ": 1,
"paper_module": 1,
"principle": 5
},
"link_count": 14,
"links": [
{
"id": "formal_evidence_cell_anchor_resolver",
"kind": "organ",
"kind_label": "component",
"label": "Formal Evidence Cell Anchor Resolver",
"relation": "explains",
"url": "component-formal-evidence-cell-anchor-resolver.html"
},
{
"id": "mechanism.formal_evidence_cell_anchor_resolver.validates_public_evidence_cell_anchor_resolution",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public evidence cell anchor resolution",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-evidence-cell-anchor-resolver-validates-public-evidence-cell-anchor-resolution"
},
{
"id": "mechanism.formal_evidence_cell_anchor_resolver.validates_public_evidence_cell_anchors",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public evidence cell anchors",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-evidence-cell-anchor-resolver-validates-public-evidence-cell-anchors"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depends on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.proof_diagnostic_evidence_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof Diagnostic Evidence Spine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[14:paper_module.proof_diagnostic_evidence_spine].depends_on",
"url": "paper-module-proof-diagnostic-evidence-spine.html"
},
{
"id": "paper_module.target_shape_tactic_routing",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Target Shape Tactic Routing",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[41:paper_module.target_shape_tactic_routing].depends_on",
"url": "paper-module-target-shape-tactic-routing.html"
}
],
"paper_module_id": "paper_module.formal_evidence_cell_anchor_resolver",
"reader_packet": {
"abstract": "The formal evidence cell anchor resolver binds proof-language paper claims to public evidence cells, source anchors, machine-anchor metadata, frozen public source-module snapshots, public source-artifact mirrors, and negative-case result records without claiming formal-result correctness.",
"authority_ceiling": "Evidence-cell anchor metadata and source-open runtime result records only; no formal-result correctness, proof-body import, private source-ref authority, human approval as proof authority, Lean/Lake execution, external model access, launch-scope decision, publishing-scope decision, source-file changes, or formal-proof certification.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.formal_evidence_cell_anchor_resolver.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-formal-evidence-cell-anchor-resolver-diagram-1",
"html_url": "docs/paper-module-formal-evidence-cell-anchor-resolver.html#paper-module-formal-evidence-cell-anchor-resolver-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Inputs[\"paper claims, cell registry, boundary policy, obligation graph, projection protocol\"] --> Build[\"_build_result\"] Build --> Scan[\"scan_paths secret-equivalent material\"] Build --> Validators[\"six validators projection, registry, policy, obligations, source modules, claims\"] Validators --> Claims[\"validate_claims _inspect_claim_row\"] Claims --> Status[\"status = pass only if all validators pass and every negative case observed\"] Scan --> Status Status --> Record[\"metadata-only result record refs, digests, findings\"]",
"source_format": "mermaid",
"source_hash": "sha256:2612e2e7964d91d4a3c0f5b1049b4f7f8e5895ddc39553041d57266ad3a696e1",
"source_line_end": 141,
"source_line_start": 132
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depends on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
}
],
"relation": "depends on"
},
{
"count": 3,
"links": [
{
"id": "formal_evidence_cell_anchor_resolver",
"kind": "organ",
"kind_label": "component",
"label": "Formal Evidence Cell Anchor Resolver",
"relation": "explains",
"url": "component-formal-evidence-cell-anchor-resolver.html"
},
{
"id": "mechanism.formal_evidence_cell_anchor_resolver.validates_public_evidence_cell_anchor_resolution",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public evidence cell anchor resolution",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-evidence-cell-anchor-resolver-validates-public-evidence-cell-anchor-resolution"
},
{
"id": "mechanism.formal_evidence_cell_anchor_resolver.validates_public_evidence_cell_anchors",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public evidence cell anchors",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-evidence-cell-anchor-resolver-validates-public-evidence-cell-anchors"
}
],
"relation": "explains"
},
{
"count": 6,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.formal_evidence_cell_anchor_resolver",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 224,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_evidence_cell_anchor_resolver.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "W3C PROV",
"section_id": "provenance-and-proof-certificate-lineage",
"url": "https://www.w3.org/TR/prov-overview/"
},
{
"kind": "external_reference",
"label": "Proof-Carrying Code",
"section_id": "provenance-and-proof-certificate-lineage",
"url": "https://www.usenix.org/legacy/publications/library/proceedings/osdi96/full_papers/necula/html/node2.html"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component asks whether a page's proof-language claim can be tied to a named piece of public evidence, resolving each claim to an evidence cell and capping its strength at what the cell allows. formalevidencecellanchor_resolver checks the project's proof-flavoured writing against recorded evidence. When a paper page says a claim is proved or certified, this component asks a narrower question: can the words be tied to a named piece of public evidence, and does that evidence stay below formal-result correctness. It answers by resolving each claim to an evidence cell, checking the cell's source anchors point at files that exist on the public path, and capping the claim's strength at what\u2026",
"level": 4,
"order": 1,
"section_id": "component-asks-whether",
"source_line_end": 14,
"source_line_start": 3,
"title": "Component asks whether"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The unit of evidence is a cell, a stable id bundling source-anchor refs, a machineanchorclass, and allowed claim strengths, and the resolver reports which claims resolved, overreached, or had no backing.** The unit of evidence is a cell. A cell is a stable id that stands for a bundle of recorded evidence: a list of source-anchor refs, a machineanchorclass naming what kind of machine artifact backs it, and the claim strengths the cell is allowed to support. The resolver reads paper claims, an evidence-cell registry, a claim-boundary policy, and an obligation graph, then writes a result record that reports which claims resolved, which overreached, and which had no backing. Proof bodies, non\u2026",
"level": 4,
"order": 2,
"section_id": "unit-of-evidence",
"source_line_end": 27,
"source_line_start": 15,
"title": "Unit of evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is not a theorem prover: it runs over already-public inputs, emits metadata-only result records, and states its full proof boundary once under Scope limit.** It is not a theorem prover. It runs over already-public inputs and emits metadata-only result records. The full proof boundary is stated once, under Scope limit.",
"level": 4,
"order": 3,
"section_id": "is-not-theorem",
"source_line_end": 34,
"source_line_start": 28,
"title": "Is not theorem"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 36,
"source_line_start": 35,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Proof-adjacent prose drifts easily, and a loose check that only scanned for the word proof would pass any sentence avoiding it and could not bound how strong a backed claim may be.** Proof-adjacent prose is the easiest place for a claim to drift. A page can write \"this proves the theorem\" and a reader has no cheap way to tell whether the words rest on a checked artifact or on nothing. A loose check that only scanned for the word \"proof\" would pass any sentence that avoided the word, and would have no way to bound how strong a backed claim is allowed to be.",
"level": 4,
"order": 5,
"section_id": "where-proof-adjacent-prose-drifts",
"source_line_end": 46,
"source_line_start": 37,
"title": "Where proof-adjacent prose drifts"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The key policy is prooflanguagerequiresmachine_anchor: proof-language claims must name a resolving cell whose anchors point at existing files, with strength capped and human approval refused as a substitute.** The policy that makes the component useful is prooflanguagerequiresmachineanchor. A claim that uses proof language must name a cell, the cell must resolve in the registry, and the cell's source anchors must point at files that exist. A claim that uses proof language but names no cell, or names a cell absent from the registry, lowers the run to a blocked status instead of passing as green prose. On top of the resolved cell sit two limits: the claim's strength is capped by the cell's al\u2026",
"level": 4,
"order": 6,
"section_id": "proof-language-requires-machine-anchor",
"source_line_end": 60,
"source_line_start": 47,
"title": "`proof_language_requires_machine_anchor`"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 62,
"source_line_start": 61,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**run validates a fixture directory with negative cases included and runanchorbundle validates an exported bundle with them off, both calling buildresult to run the pipeline.** Two entrypoints share one core. run validates a fixture directory with the negative cases included. runanchorbundle validates an exported bundle directory with negative cases off and the source-module manifest required. Both call buildresult, which runs the pipeline and returns the result record.",
"level": 4,
"order": 8,
"section_id": "two-entrypoints-share-one-core",
"source_line_end": 72,
"source_line_start": 63,
"title": "Two entrypoints share one core"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "buildresult loads the five inputs, scans every path for secret-equivalent material, runs six validators, and passes only when all pass, every expected negative case was observed, and no scan hit fired. build_result loads the five fixture inputs with loadpayloads, scans every input path for secret-equivalent material with scan_paths, then runs six validators and folds their findings into one status. The run passes only when every validator passes, every expected negative case was observed, and no secret scan hit fired. The validators, in the order buildresult calls them:",
"level": 4,
"order": 9,
"section_id": "build-result-scans-validates-folds",
"source_line_end": 93,
"source_line_start": 73,
"title": "`_build_result` scans, validates, folds"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateclaims passes each claim to inspectclaim_row, which rejects proof bodies, non-public refs, oracle payloads, missing boundary classes, unknown cells, and out-of-range strengths as structured findings. validateclaims is the resolver proper. It builds a cellbyid map from the registry and passes each claim to inspectclaim_row. That function rejects a claim when it carries a proof body, a non-public source refs, or a provider or oracle payload; when it omits an oracle-boundary class; when it uses proof language with no cell; when its cell id is unknown; or when its stated strength is not in the cell's allowed list. A theorem-correctness overclaim is caught either from an explicit flag or\u2026",
"level": 4,
"order": 10,
"section_id": "validate-claims-resolves-each-claim",
"source_line_end": 108,
"source_line_start": 94,
"title": "`validate_claims` resolves each claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateanchor_refs splits each ref, rejects private schemes and escaping paths, resolves the file among public candidates including a source-artifact mirror, and records digest matches without holding file bodies.** Source anchors resolve through validateanchor_refs. It splits each ref into a path and an optional marker, rejects private schemes and absolute or parent paths through publicrefisforbidden, then looks for the file among the candidates from publicref_candidates. A state/runs/... anchor is allowed to resolve through a public source-artifact mirror under examples///source_artifacts, so the resolver can confirm an anchor exists from an existing public example bundle without copying\u2026",
"level": 4,
"order": 11,
"section_id": "resolving-source-anchors-on-the-public-path",
"source_line_end": 123,
"source_line_start": 109,
"title": "Resolving source anchors on the public path"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "resultcard, sourcemodulesummarycard, and boardfrom_result shape the result into compact cards, each carrying bodyinreceipt: False.** The result record is shaped by resultcard, sourcemodulesummary_card, and boardfrom_result into compact cards. Every card and finding carries bodyinreceipt: False.",
"level": 4,
"order": 12,
"section_id": "body-free-card-shaping",
"source_line_end": 142,
"source_line_start": 124,
"title": "metadata-only card shaping"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 144,
"source_line_start": 143,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVE_CASES declares thirteen negative-case classes the fixture run must observe, and a run fails if any is missing, keeping the refusal coverage from silently eroding. EXPECTEDNEGATIVE_CASES declares thirteen negative-case classes that the fixture run must observe. A run fails if any of them is missing, which keeps the refusal coverage from silently eroding. The classes name the exact rejection each case exercises: unknowncelloverclaim, missingsourceanchor, prooflanguagewithoutcell, privatesourcerefleakage, proofbodyleakage, theoremcorrectnessoverclaim, humanapprovalasevidencecell, oraclepayloadleakage, oracleboundaryforwardsuccessoverclaim, and four obligation-graph variants fo\u2026",
"level": 4,
"order": 14,
"section_id": "thirteen-required-negative-case-classes",
"source_line_end": 158,
"source_line_start": 145,
"title": "Thirteen required negative-case classes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 160,
"source_line_start": 159,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component follows provenance and proof-certificate work like W3C PROV and Proof-Carrying Code, borrowing anchor resolution so a small checker bounds what cell metadata can assert without granting theorem authority.** This component follows provenance and proof-certificate work where a claim must point at checkable evidence rather than untyped narrative. The W3C PROV model links entities, activities, and agents in an evidence graph. Proof-Carrying Code and small-kernel proof assistants motivate keeping a certificate or anchor separate from the trusted checker that bounds its meaning. This resolver borrows the anchor-resolution pattern: proof-language claims must name cell ids, source a\u2026",
"level": 4,
"order": 16,
"section_id": "provenance-and-proof-certificate-lineage",
"source_line_end": 175,
"source_line_start": 161,
"title": "Provenance and proof-certificate lineage"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 177,
"source_line_start": 176,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run these from the repo root. Run these from the repo root. They do not mutate durable result record or generated projection surfaces.",
"level": 4,
"order": 18,
"section_id": "run-from-repo",
"source_line_end": 187,
"source_line_start": 178,
"title": "Run from repo"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the fixture run observed all thirteen negative cases, resolved three claims to three cells across eight anchors, validated six frozen snapshots by digest, and kept everything metadata-only.** A pass means the fixture run observed all thirteen negative cases, resolved three paper claims to three evidence cells across eight source anchors and three machine anchors, validated the exported bundle against six frozen source-module snapshots by digest, and kept every card and finding metadata-only. It is bounded validator evidence over that fixture, not a statement about the wider formal-math corpus.",
"level": 4,
"order": 19,
"section_id": "what-a-pass-means-here",
"source_line_end": 198,
"source_line_start": 188,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 200,
"source_line_start": 199,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Public evidence-cell metadata can bind proof-language claims to recorded digest-checked cells, cap each claim at the strength its cell allows, and reject claims that leak private material or overreach.** The strongest claim the evidence supports: public evidence-cell metadata can bind proof-language claims to recorded, digest-checked cells, cap each claim at the strength its cell allows, and reject claims that leak private material or overreach into formal-result correctness. The counts are the fixture's contract, not a scale claim.",
"level": 4,
"order": 21,
"section_id": "the-strongest-supported-claim",
"source_line_end": 210,
"source_line_start": 201,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The proof boundary is anchor metadata and source-open runtime records only: this runs no Lean, proves no formal-result correctness, and a digest match confirms only that a snapshot equals its recorded hash.** The proof boundary is anchor metadata and source-open runtime records only. This component does not run Lean or Lake, prove formal-result correctness, import proof bodies, expose non-public source refs, carry provider or oracle payloads, treat human approval as proof, or include launch operations or public sharing. A digest match confirms a snapshot equals its recorded hash; it does not certify that the snapshot equals the latest live source, that every source file was imported, or t\u2026",
"level": 4,
"order": 22,
"section_id": "the-proof-boundary-is-anchor-metadata",
"source_line_end": 223,
"source_line_start": 211,
"title": "The proof boundary is anchor metadata"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/formal_evidence_cell_anchor_resolver.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The formal evidence cell anchor resolver binds proof-language paper claims to public evidence cells, source anchors, machine-anchor metadata, frozen public source-module snapshots, public source-artifact mirrors, and negative-case result records without claiming formal-result correctness."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "formal_evidence_cell_anchor_resolver.py",
"path": "src/microcosm_core/organs/formal_evidence_cell_anchor_resolver.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for evidence-cell anchor resolution, exported bundle validation, source-module manifest digest checks, frozen public snapshot checks, public source-artifact mirror lookup, source-module summary cards, negative-case enforcement, and scope limits.",
"symbols": [
"run",
"run_anchor_bundle",
"validate_source_module_manifest",
"_public_ref_candidates",
"_public_source_artifact_mirror_roots",
"_build_result",
"_source_module_summary_card",
"EXPECTED_NEGATIVE_CASES"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/formal_evidence_cell_anchor_resolver.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_evidence_cell_anchor_resolver.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_evidence_cell_anchor_resolver.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Formal Evidence Cell Anchor Resolver is the evidence-legibility membrane for Microcosm's formal math claims. It resolves three paper claims to public evidence-cell ids, checks source-anchor and machine-anchor metadata, anchors the verifier-trace cell to real Ring2 verifier-trace repair result records, validates six frozen source-open body snapshots, resolves public source-artifact mirrors for state/runs anchors, observes seven proof/private/human-approval/theorem-correctness negative cases, and emits metadata-only result records that make proof-language boundaries inspectable without becoming proof authority.",
"summary_line": "The formal evidence cell anchor resolver binds proof-language paper claims to public evidence cells, source anchors, machine-anchor metadata, frozen public source-module snapshots, public source-artifact mirrors, and negative-case result records without claiming formal-result correctness.",
"title": "Formal Evidence Cell Anchor Resolver"
},
{
"anchor": "paper-module-formal-math-premise-retrieval",
"authority_ceiling": "Copied public source retrieval metadata and runtime validation result records only; no formal-result correctness, proof-body import, oracle-needed premise authority, Mathlib authority, Lean/Lake execution, external model access, benchmark claim, launch-scope decision, publishing-scope decision, source-file changes, or general formal-proof authority.",
"body_html": "
Public premise retrieval after source import
\n
This is the first source-available formal-math import slice after the source projection protocol, turning the source prover lab's premise-index, term-scoring, context-budget, and strategy-selection patterns into a runnable component.formal_math_premise_retrieval is the source-available first real formal-math import slice after the source projection protocol. It turns the source prover lab's premise-index, term-scoring, context-budget, and strategy-selection patterns into a runnable Microcosm component.
\n
It is still deliberately below proof authority. It validates:
\n
Lean/Std premise metadata;
query term scoring across public premise ids, namespaces, declaration names, statement excerpts, and retrieval terms;
split eligibility;
context recipe budgets;
public strategy ids;
redacted result records;
negative cases.
\n
It stays deliberately below proof authority, refusing Lean or Lake execution, external model access, proof-body or oracle-id exposure, test-split tuning, correctness claims, and launch. It does not run Lean or Lake, use external model services, expose proof bodies, expose oracle-needed premise ids, tune on test split truth, claim formal-result correctness, or include launch operations.
\n
Purpose
\n
Premise selection narrows the library
\n
Before a model can attempt a proof it must find the right lemmas, and this component is the smallest honest premise-selection step: score every public premise against a query and return a ranked shortlist. Before a model can attempt a formal proof, it has to find the right lemmas. A Lean library holds thousands of theorems and definitions, and the useful ones for a given goal are a handful. Premise selection is the step that narrows that library down to candidates worth putting in front of a prover. This component is the smallest honest version of that step: it takes a query, scores every public premise against it, and returns a ranked shortlist.
\n
Retrieval must work and the boundary must hold
\n
The narrow question is whether transparent term-scoring over copied public Lean/Std metadata returns the premises a query should find without touching a proof, so retrieval must work and the boundary must hold. The single question it answers is narrow and checkable: given a copied catalogue of public Lean/Std premise metadata, does a transparent term-scoring retrieval return the premises a query should find, without ever touching a proof? Both halves matter. The retrieval has to actually work, so each fixture query carries the premise ids it is expected to surface and the run fails if the shortlist misses them. And the boundary has to hold, so the same run refuses any input that smuggles in a proof body, an oracle answer, or test-split truth.
\n
The restraint is the point
\n
The index is plain inspectable term overlap, not a learned embedding or benchmark claims, so the claim is that retrieval runs over real copied Lean metadata and forbids the shortcuts that would flatter it. What is unusual is the restraint. The retrieval index is not a learned embedding model and the scoring is not a benchmark claims. It is plain term overlap over fields that a reader can inspect: premise ids, namespaces, declaration names, statement excerpts, and retrieval terms. The interesting claim is therefore not "this retrieves well" but "this retrieves over real, copied Lean metadata and can be audited end to end, and the design forbids the shortcuts that would make a premise-selection result look better than it is".
\n
Shape
\nDiagram of the mechanism (14 steps).Diagram source & refs
Source refs
JSON source record
paper_module.formal_math_premise_retrieval
Runtime component
formal_math_premise_retrieval.py
flowchart TD bundle["JSON source record paper_module.formal_math_premise_retrieval"] --> instance["Generated paper-module instance 15 relationship edges"] instance --> component["Runtime component formal_math_premise_retrieval.py"] subgraph Inputs["Public inputs"] index["Premise index copied Lean/Std metadata"] queries["Retrieval queries terms, split, strategy, top_k"] recipes["Context recipes byte budgets"] negatives["Negative-case inputs proof body, oracle ids, test-split tuning, budget, strategy"] end component --> index component --> queries component --> recipes component --> negatives index --> split["Split gate skip premises not in allowed_for_split"] queries --> split split --> score["Term-overlap scoring shared tokens + strategy bonus"] score --> shortlist["Ranked top_k shortlist"] shortlist --> recall["Recall check vs expected premise ids"] negatives --> reject["Required rejections five leakage/overclaim guards"] recipes --> reject recall --> result records["metadata-only result records board, validation, sign-off"] reject --> result records result records --> ceiling["Scope limit metadata coherence, no Lean/Lake, no proof"]
\n
Evidence/accounting:
\n
Bundle authority: core/paper_module_capsules.json::paper_modules[25:paper_module.formal_math_premise_retrieval] has source_authority: json_capsule, three subjects, one resolved code_loci[0].path, depends_on naming paper_module.formal_math_lean_proof_witness, and generated projection statuses for Markdown, Mermaid, and Atlas.
Generated instance: paper_modules/formal_math_premise_retrieval.json::paper_module_payload repeats the bundle authority_ceiling, reports Mermaid status available_from_capsule_edges, and derives 15 relationships.edges with relationships.unpopulated_selective_relations: [].
Component atlas: core/organ_atlas.json::organs[9:formal_math_premise_retrieval] classifies the component in family: formal_math_and_proof, cites the runtime locus, and restates that retrieval metadata coherence is not Lean/Lake, provider, theorem-correctness, benchmark, or launch-scope decision.
Mechanism rows: core/mechanism_sources.json::mechanisms[27:mechanism.formal_math_premise_retrieval.validates_public_premise_retrieval_slice] and core/mechanism_sources.json::mechanisms[37:mechanism.formal_math_premise_retrieval.validates_public_premise_retrieval_projection] point at src/microcosm_core/organs/formal_math_premise_retrieval.py and name first-wave, sign-off, and runtime-shell result record refs.
Runtime and tests: src/microcosm_core/organs/formal_math_premise_retrieval.py exposes run, run_retrieval_bundle, EXPECTED_NEGATIVE_CASES, and AUTHORITY_CEILING; tests/test_formal_math_premise_retrieval.py checks 11 premises, 4 queries, 44 considered candidates, five negative cases, metadata-only result records, and compact runtime-shell cards.
Result records: receipts/first_wave/formal_math_premise_retrieval/formal_math_premise_retrieval_result.json records status: pass, 11 premises, 4 queries, 44 considered candidates, five observed negative cases, missing_negative_cases: [], and a secret-exclusion scan with blocking_hit_count: 0; the exported runtime result record at receipts/runtime_shell/demo_project/organs/formal_math_premise_retrieval/exported_premise_retrieval_bundle_validation_result.json records status: pass, the same premise/query/candidate counts, no negative cases, and secret_exclusion_scan.scanned_path_count: 11.
Standard ceiling: standards/std_microcosm_formal_math_premise_retrieval.json::authority_ceiling has status: pass while keeping formal_proof_authority, lean_lake_authority, provider_authority, and release_authority false.
\n
Runtime Surfaces
\n
Component runner: python -m microcosm_core.organs.formal_math_premise_retrieval run --input fixtures/first_wave/formal_math_premise_retrieval/input --out receipts/first_wave/formal_math_premise_retrieval
Microcosm can show a real formal-math retrieval mechanism in miniature:
\n
a source-available Lean/Std premise index;
public field-haystack term-scored queries;
split-aware eligibility;
context recipe ceilings;
strategy gates;
redacted validation result records.
\n
How retrieval scoring works
\n
Five inspectable fields per premise
\n
Each premise contributes five inspectable fields, premise id, namespace, declaration name, statement excerpt, and retrieval terms, while a query carries its own terms, split, optional strategy, recipe, and expected ids. Each premise row contributes five inspectable fields to the haystack: its premise id, namespace, declaration name, statement excerpt, and a list of retrieval terms. A query carries its own terms, a data split, an optional strategy id, a context recipe, and the public premise ids it is expected to return.
\n
Term-overlap scoring with a split gate
\n
Scoring is per-query term overlap over the five fields, with a split gate that structurally excludes premises whose allowed_for_split omits the query split and a single-point bonus for a matching strategy tag. Scoring is term overlap, computed per query. Both the query and each premise are tokenised into lowercase word counts. A premise is only considered if the query's split appears in that premise's allowed_for_split list, which is how test-split leakage is kept out at the structural level rather than by trust. For each eligible premise the score is the summed minimum count of every shared token across the five fields, so a term that appears in both the query and the premise contributes as many points as the smaller of the two counts. A premise that also carries the query's strategy id as a tag gets a single extra point. The ranked list is sorted by score descending, ties broken by premise id, and the top of that list up to the query's top_k is taken as the retrieval.
\n
Grading recall against expected ids
\n
The retrieval is graded against itself: each query declares the ids it should surface, recall is the fraction that landed in the shortlist, and any miss blocks the run. The retrieval is then graded against itself. Each query declares the public premise ids it should surface, and the component computes recall as the fraction of those expected ids that actually landed in the shortlist. A query that declares expectations but misses any of them blocks the run. In the first-wave fixture this is eleven premises and four queries, scoring forty-four considered candidates in total, and every query is expected to reach full recall.
\n
Five refusals keep the signal honest
\n
The five negative-case inputs each encode a shortcut, a shipped proof body, oracle ids, test-split tuning, a budget overflow, and an unknown strategy, and all five rejections must be observed or the fixture blocks. The failure mode this guards against is a premise-selection result that looks good because it cheated. The five negative-case inputs each encode one such shortcut: a premise index that ships a proof body, a query that lists the oracle premise ids it is "meant" to find, a query that tunes on test-split truth, a context recipe that blows past the byte budget, and a query naming a strategy id outside the allowed set. The run is required to observe all five rejections; if any expected rejection is missing, the whole fixture is blocked rather than passed. Recall over copied real metadata is the positive signal; the refusals are what keep that signal honest.
\n
Prior Art Grounding
\n
Premise-selection and retrieval-augmented proving
\n
The component is grounded in premise-selection and retrieval-augmented theorem proving, closest to LeanDojo and motivated by earlier environments like HOList and GamePad that extract premise metadata for learning-assisted proving. This component is grounded in premise-selection and retrieval-augmented theorem proving work. LeanDojo is the closest modern anchor because it couples Lean interaction with retrieval-augmented premise selection. Earlier theorem-proving environments such as HOList and GamePad also motivate extracting proof-state or premise metadata for learning-assisted theorem proving.
\n
Borrowing the retrieval accounting pattern
\n
Microcosm borrows the retrieval accounting pattern, requiring premise ids, excerpts, terms, split eligibility, budgets, and strategy gates to be inspectable before any retrieval claim, without running Lean/Lake or exposing proof bodies. Microcosm borrows the retrieval accounting pattern: premise ids, namespaces, statement excerpts, retrieval terms, split eligibility, context budgets, and strategy gates must be inspectable before premise-retrieval claims are admitted. It does not run Lean/Lake or expose proof bodies.
\n
Negative Cases
\n
premise_index_proof_body_forbidden
query_oracle_ids_forbidden
test_split_tuning_attempt
context_recipe_budget_overflow
unknown_strategy_id
\n
Reader Evidence Routing
\n
Start with the JSON Bundle Binding to identify the source record, generated instance, proof boundary, and scope limit.
Use Structured Lattice Bindings for navigation; the generated JSON row is the authority for relationship counts and dependency state.
Use Runtime Surfaces and Result record Expectations when checking metadata coherence, redaction, leakage checks, and source-available bundle behavior.
Use Negative Cases, Scope limit, and Scope limit together before admitting any formal-math public claim.
The component proves only that public retrieval metadata is internally coherent and leakage-checked, and the deferred formal_math_lean_proof_witness boundary remains unchanged. The component proves only that public retrieval metadata is internally coherent and leakage-checked. The deferred formal_math_lean_proof_witness boundary remains unchanged.
\n
Scope limit
\n
The reader-verifiable claim and its limits
\n
The module supports only the reader-verifiable claim that public premise metadata, retrieval terms, split eligibility, strategy gates, and result records are coherent and leakage-checked, running no Lean and proving no formal-result correctness. This module supports only the reader-verifiable claim that public premise metadata, retrieval terms, split eligibility, strategy gates, and redacted result records are coherent and leakage-checked. It does not run Lean or Lake, prove formal-result correctness, expose proof bodies, authorize oracle-needed premise ids, tune on test split truth, use external model services, approve public sharing, or expand the deferred Lean proof-witness boundary.
\n",
"components": [
{
"anchor": "component-formal_math_premise_retrieval",
"component_id": "formal_math_premise_retrieval",
"public_label": "Formal Math Premise Retrieval"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 2,
"organ": 1,
"paper_module": 1,
"principle": 5
},
"link_count": 14,
"links": [
{
"id": "formal_math_premise_retrieval",
"kind": "organ",
"kind_label": "component",
"label": "Formal Math Premise Retrieval",
"relation": "explains",
"url": "component-formal-math-premise-retrieval.html"
},
{
"id": "mechanism.formal_math_premise_retrieval.validates_public_premise_retrieval_slice",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public premise retrieval slice",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-premise-retrieval-validates-public-premise-retrieval-slice"
},
{
"id": "mechanism.formal_math_premise_retrieval.validates_public_premise_retrieval_projection",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public premise retrieval projection",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-premise-retrieval-validates-public-premise-retrieval-projection"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.formal_math_lean_proof_witness",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Lean Proof Witness",
"relation": "depends on",
"url": "paper-module-formal-math-lean-proof-witness.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[23:paper_module.formal_math_verifier_trace_repair_loop].depends_on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
},
{
"id": "paper_module.lean_std_premise_index",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Lean/Std Premise Index",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[26:paper_module.lean_std_premise_index].depends_on",
"url": "paper-module-lean-std-premise-index.html"
},
{
"id": "paper_module.mathematical_strategy_atlas",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mathematical Strategy Atlas",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[43:paper_module.mathematical_strategy_atlas].depends_on",
"url": "paper-module-mathematical-strategy-atlas.html"
},
{
"id": "paper_module.ring2_premise_precision_recall",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Ring-2 Premise Precision Recall",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[42:paper_module.ring2_premise_precision_recall].depends_on",
"url": "paper-module-ring2-premise-precision-recall.html"
},
{
"id": "paper_module.verifier_lab_kernel",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Kernel",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[0:paper_module.verifier_lab_kernel].depends_on",
"url": "paper-module-verifier-lab-kernel.html"
}
],
"paper_module_id": "paper_module.formal_math_premise_retrieval",
"reader_packet": {
"abstract": "Formal math premise retrieval validates copied public Lean/Std premise metadata, query scoring, context budgets, strategy gates, body-floor provenance, and leakage negative cases without claiming proof authority.",
"authority_ceiling": "Copied public source retrieval metadata and runtime validation result records only; no formal-result correctness, proof-body import, oracle-needed premise authority, Mathlib authority, Lean/Lake execution, external model access, benchmark claim, launch-scope decision, publishing-scope decision, source-file changes, or general formal-proof authority.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.formal_math_premise_retrieval.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-formal-math-premise-retrieval-diagram-1",
"html_url": "docs/paper-module-formal-math-premise-retrieval.html#paper-module-formal-math-premise-retrieval-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD bundle[\"JSON source record paper_module.formal_math_premise_retrieval\"] --> instance[\"Generated paper-module instance 15 relationship edges\"] instance --> component[\"Runtime component formal_math_premise_retrieval.py\"] subgraph Inputs[\"Public inputs\"] index[\"Premise index copied Lean/Std metadata\"] queries[\"Retrieval queries terms, split, strategy, top_k\"] recipes[\"Context recipes byte budgets\"] negatives[\"Negative-case inputs proof body, oracle ids, test-split tuning, budget, strategy\"] end component --> index component --> queries component --> recipes component --> negatives index --> split[\"Split gate skip premises not in allowed_for_split\"] queries --> split split --> score[\"Term-overlap scoring shared tokens + strategy bonus\"] score --> shortlist[\"Ranked top_k shortlist\"] shortlist --> recall[\"Recall check vs expected premise ids\"] negatives --> reject[\"Required rejections five leakage/overclaim guards\"] recipes --> reject recall --> result records[\"metadata-only result records board, validation, sign-off\"] reject --> result records result records --> ceiling[\"Scope limit metadata coherence, no Lean/Lake, no proof\"]",
"source_format": "mermaid",
"source_hash": "sha256:2a537a867822aa6a3b3621163a9ea5be58875acc136a92764c6a5fc6bc4fb1fe",
"source_line_end": 146,
"source_line_start": 117
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.formal_math_lean_proof_witness",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Lean Proof Witness",
"relation": "depends on",
"url": "paper-module-formal-math-lean-proof-witness.html"
}
],
"relation": "depends on"
},
{
"count": 3,
"links": [
{
"id": "formal_math_premise_retrieval",
"kind": "organ",
"kind_label": "component",
"label": "Formal Math Premise Retrieval",
"relation": "explains",
"url": "component-formal-math-premise-retrieval.html"
},
{
"id": "mechanism.formal_math_premise_retrieval.validates_public_premise_retrieval_slice",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public premise retrieval slice",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-premise-retrieval-validates-public-premise-retrieval-slice"
},
{
"id": "mechanism.formal_math_premise_retrieval.validates_public_premise_retrieval_projection",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public premise retrieval projection",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-formal-math-premise-retrieval-validates-public-premise-retrieval-projection"
}
],
"relation": "explains"
},
{
"count": 6,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.formal_math_premise_retrieval",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 70,
"title": "JSON Bundle Binding"
},
{
"reason": "public_projection_internal_section_denylist",
"section_id": "structured-lattice-bindings",
"source_line_start": 90,
"title": "Structured Lattice Bindings"
},
{
"reason": "public_projection_internal_section_denylist",
"section_id": "receipt-expectations",
"source_line_start": 299,
"title": "Result record Expectations"
},
{
"reason": "public_projection_internal_section_denylist",
"section_id": "re-entry-conditions",
"source_line_start": 364,
"title": "Re-Entry Conditions"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_math_premise_retrieval.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "LeanDojo",
"section_id": "premise-selection-and-retrieval-augmented-proving",
"url": "https://arxiv.org/abs/2306.15626"
},
{
"kind": "external_reference",
"label": "HOList",
"section_id": "premise-selection-and-retrieval-augmented-proving",
"url": "https://arxiv.org/abs/1904.03241"
},
{
"kind": "external_reference",
"label": "GamePad",
"section_id": "premise-selection-and-retrieval-augmented-proving",
"url": "https://arxiv.org/abs/1806.00608"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**This is the first source-available formal-math import slice after the source projection protocol, turning the source prover lab's premise-index, term-scoring, context-budget, and strategy-selection patterns into a runnable component.** formalmathpremise_retrieval is the source-available first real formal-math import slice after the source projection protocol. It turns the source prover lab's premise-index, term-scoring, context-budget, and strategy-selection patterns into a runnable Microcosm component. It is still deliberately below proof authority. It validates: Lean/Std premise metadata; query term scoring across public premise ids, namespaces, declaration names, statement excerpts, an\u2026",
"level": 4,
"order": 1,
"section_id": "public-premise-retrieval-after-macro-import",
"source_line_end": 23,
"source_line_start": 3,
"title": "Public premise retrieval after source import"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It stays deliberately below proof authority, refusing Lean or Lake execution, external model access, proof-body or oracle-id exposure, test-split tuning, correctness claims, and launch.** It does not run Lean or Lake, use external model services, expose proof bodies, expose oracle-needed premise ids, tune on test split truth, claim formal-result correctness, or include launch operations.",
"level": 4,
"order": 2,
"section_id": "it-stays-deliberately-below-proof-authority",
"source_line_end": 31,
"source_line_start": 24,
"title": "It stays deliberately below proof authority"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 3,
"section_id": "purpose",
"source_line_end": 33,
"source_line_start": 32,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Before a model can attempt a proof it must find the right lemmas, and this component is the smallest honest premise-selection step: score every public premise against a query and return a ranked shortlist.** Before a model can attempt a formal proof, it has to find the right lemmas. A Lean library holds thousands of theorems and definitions, and the useful ones for a given goal are a handful. Premise selection is the step that narrows that library down to candidates worth putting in front of a prover. This component is the smallest honest version of that step: it takes a query, scores every public premise against it, and returns a ranked shortlist.",
"level": 4,
"order": 4,
"section_id": "premise-selection-narrows-the-library",
"source_line_end": 44,
"source_line_start": 34,
"title": "Premise selection narrows the library"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The narrow question is whether transparent term-scoring over copied public Lean/Std metadata returns the premises a query should find without touching a proof, so retrieval must work and the boundary must hold.** The single question it answers is narrow and checkable: given a copied catalogue of public Lean/Std premise metadata, does a transparent term-scoring retrieval return the premises a query should find, without ever touching a proof? Both halves matter. The retrieval has to actually work, so each fixture query carries the premise ids it is expected to surface and the run fails if the shortlist misses them. And the boundary has to hold, so the same run refuses any input that smuggle\u2026",
"level": 4,
"order": 5,
"section_id": "retrieval-must-work-and-the-boundary-must-hold",
"source_line_end": 57,
"source_line_start": 45,
"title": "Retrieval must work and the boundary must hold"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The index is plain inspectable term overlap, not a learned embedding or benchmark claims, so the claim is that retrieval runs over real copied Lean metadata and forbids the shortcuts that would flatter it.** What is unusual is the restraint. The retrieval index is not a learned embedding model and the scoring is not a benchmark claims. It is plain term overlap over fields that a reader can inspect: premise ids, namespaces, declaration names, statement excerpts, and retrieval terms. The interesting claim is therefore not \"this retrieves well\" but \"this retrieves over real, copied Lean metadata and can be audited end to end, and the design forbids the shortcuts that would make a premise-sel\u2026",
"level": 4,
"order": 6,
"section_id": "the-restraint-is-the-point",
"source_line_end": 69,
"source_line_start": 58,
"title": "The restraint is the point"
},
{
"block_kinds": [
"diagram",
"list",
"paragraph"
],
"excerpt": "Evidence/accounting: Bundle authority: core/papermodulecapsules.json::papermodules[25:papermodule.formalmathpremise_retrieval] has sourceauthority: jsoncapsule, three subjects, one resolved codeloci[0].path, dependson naming papermodule.formalmathleanproof_witness, and generated projection statuses for Markdown, Mermaid, and Atlas. Generated instance: papermodules/formalmathpremiseretrieval.json::papermodulepayload repeats the bundle scope_limit, reports Mermaid status availablefromcapsule_edges, and derives 15 relationships.edges with relationships.unpopulatedselectiverelations: []. Component atlas: core/organatlas.json::components[9:formalmathpremiseretrieval] classifies the component in\u2026",
"level": 2,
"order": 7,
"section_id": "shape",
"source_line_end": 192,
"source_line_start": 115,
"title": "Shape"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "Component runner: python -m microcosmcore.components.formalmathpremiseretrieval run --input fixtures/firstwave/formalmathpremiseretrieval/input --out result records/firstwave/formalmathpremiseretrieval Exported bundle runner: python -m microcosmcore.components.formalmathpremiseretrieval run-retrieval-bundle --input examples/formalmathpremiseretrieval/exportedpremiseretrievalbundle --out result records/runtimeshell/demoproject/components/formalmathpremise_retrieval CLI route: microcosm formal-math-premise-retrieval run-retrieval-bundle Standard: standards/stdmicrocosmformalmathpremise_retrieval.json Fixture manifest: core/fixturemanifests/formalmathpremiseretrieval.fixture_manifest.json",
"level": 2,
"order": 8,
"section_id": "runtime-surfaces",
"source_line_end": 202,
"source_line_start": 193,
"title": "Runtime Surfaces"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "Microcosm can show a real formal-math retrieval mechanism in miniature: a source-available Lean/Std premise index; public field-haystack term-scored queries; split-aware eligibility; context recipe ceilings; strategy gates; redacted validation result records.",
"level": 2,
"order": 9,
"section_id": "public-claim",
"source_line_end": 213,
"source_line_start": 203,
"title": "Public Claim"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 10,
"section_id": "how-retrieval-scoring-works",
"source_line_end": 215,
"source_line_start": 214,
"title": "How retrieval scoring works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Each premise contributes five inspectable fields, premise id, namespace, declaration name, statement excerpt, and retrieval terms, while a query carries its own terms, split, optional strategy, recipe, and expected ids.** Each premise row contributes five inspectable fields to the haystack: its premise id, namespace, declaration name, statement excerpt, and a list of retrieval terms. A query carries its own terms, a data split, an optional strategy id, a context recipe, and the public premise ids it is expected to return.",
"level": 4,
"order": 11,
"section_id": "five-inspectable-fields-per-premise",
"source_line_end": 225,
"source_line_start": 216,
"title": "Five inspectable fields per premise"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Scoring is per-query term overlap over the five fields, with a split gate that structurally excludes premises whose allowedforsplit omits the query split and a single-point bonus for a matching strategy tag.** Scoring is term overlap, computed per query. Both the query and each premise are tokenised into lowercase word counts. A premise is only considered if the query's split appears in that premise's allowedforsplit list, which is how test-split leakage is kept out at the structural level rather than by trust. For each eligible premise the score is the summed minimum count of every shared token across the five fields, so a term that appears in both the query and the premise contributes a\u2026",
"level": 4,
"order": 12,
"section_id": "term-overlap-scoring-with-a-split-gate",
"source_line_end": 241,
"source_line_start": 226,
"title": "Term-overlap scoring with a split gate"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The retrieval is graded against itself: each query declares the ids it should surface, recall is the fraction that landed in the shortlist, and any miss blocks the run.** The retrieval is then graded against itself. Each query declares the public premise ids it should surface, and the component computes recall as the fraction of those expected ids that actually landed in the shortlist. A query that declares expectations but misses any of them blocks the run. In the first-wave fixture this is eleven premises and four queries, scoring forty-four considered candidates in total, and every query is expected to reach full recall.",
"level": 4,
"order": 13,
"section_id": "grading-recall-against-expected-ids",
"source_line_end": 252,
"source_line_start": 242,
"title": "Grading recall against expected ids"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The five negative-case inputs each encode a shortcut, a shipped proof body, oracle ids, test-split tuning, a budget overflow, and an unknown strategy, and all five rejections must be observed or the fixture blocks.** The failure mode this guards against is a premise-selection result that looks good because it cheated. The five negative-case inputs each encode one such shortcut: a premise index that ships a proof body, a query that lists the oracle premise ids it is \"meant\" to find, a query that tunes on test-split truth, a context recipe that blows past the byte budget, and a query naming a strategy id outside the allowed set. The run is required to observe all five rejections; if any exp\u2026",
"level": 4,
"order": 14,
"section_id": "five-refusals-keep-the-signal-honest",
"source_line_end": 266,
"source_line_start": 253,
"title": "Five refusals keep the signal honest"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 268,
"source_line_start": 267,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component is grounded in premise-selection and retrieval-augmented theorem proving, closest to LeanDojo and motivated by earlier environments like HOList and GamePad that extract premise metadata for learning-assisted proving.** This component is grounded in premise-selection and retrieval-augmented theorem proving work. LeanDojo is the closest modern anchor because it couples Lean interaction with retrieval-augmented premise selection. Earlier theorem-proving environments such as HOList and GamePad also motivate extracting proof-state or premise metadata for learning-assisted theorem proving.",
"level": 4,
"order": 16,
"section_id": "premise-selection-and-retrieval-augmented-proving",
"source_line_end": 280,
"source_line_start": 269,
"title": "Premise-selection and retrieval-augmented proving"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Microcosm borrows the retrieval accounting pattern, requiring premise ids, excerpts, terms, split eligibility, budgets, and strategy gates to be inspectable before any retrieval claim, without running Lean/Lake or exposing proof bodies.** Microcosm borrows the retrieval accounting pattern: premise ids, namespaces, statement excerpts, retrieval terms, split eligibility, context budgets, and strategy gates must be inspectable before premise-retrieval claims are admitted. It does not run Lean/Lake or expose proof bodies.",
"level": 4,
"order": 17,
"section_id": "borrowing-the-retrieval-accounting-pattern",
"source_line_end": 290,
"source_line_start": 281,
"title": "Borrowing the retrieval accounting pattern"
},
{
"block_kinds": [
"list"
],
"excerpt": "premiseindexproofbodyforbidden queryoracleids_forbidden testsplittuning_attempt contextrecipebudget_overflow unknownstrategyid",
"level": 2,
"order": 18,
"section_id": "negative-cases",
"source_line_end": 298,
"source_line_start": 291,
"title": "Negative Cases"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 318,
"source_line_start": 317,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "result record or generated projection surfaces:",
"level": 4,
"order": 20,
"section_id": "reader-validation-without-durable-writes",
"source_line_end": 329,
"source_line_start": 319,
"title": "Reader validation without durable writes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "authority-ceiling",
"source_line_end": 331,
"source_line_start": 330,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component proves only that public retrieval metadata is internally coherent and leakage-checked, and the deferred formalmathleanproofwitness boundary remains unchanged.** The component proves only that public retrieval metadata is internally coherent and leakage-checked. The deferred formalmathleanproofwitness boundary remains unchanged.",
"level": 4,
"order": 22,
"section_id": "coherence-only-boundary-unchanged",
"source_line_end": 339,
"source_line_start": 332,
"title": "Coherence only, boundary unchanged"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 23,
"section_id": "claim-ceiling",
"source_line_end": 341,
"source_line_start": 340,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The module supports only the reader-verifiable claim that public premise metadata, retrieval terms, split eligibility, strategy gates, and result records are coherent and leakage-checked, running no Lean and proving no formal-result correctness.** This module supports only the reader-verifiable claim that public premise metadata, retrieval terms, split eligibility, strategy gates, and redacted result records are coherent and leakage-checked. It does not run Lean or Lake, prove formal-result correctness, expose proof bodies, authorize oracle-needed premise ids, tune on test split truth, use external model services, approve public sharing, or expand the deferred Lean proof-witness boundary.",
"level": 4,
"order": 24,
"section_id": "the-reader-verifiable-claim-and-its-limits",
"source_line_end": 352,
"source_line_start": 342,
"title": "The reader-verifiable claim and its limits"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "Start with the JSON Bundle Binding to identify the source record, generated instance, proof boundary, and scope limit. Use Structured Lattice Bindings for navigation; the generated JSON row is the authority for relationship counts and dependency state. Use Runtime Surfaces and Result record Expectations when checking metadata coherence, redaction, leakage checks, and source-available bundle behavior. Use Negative Cases, Scope limit, and Scope limit together before admitting any formal-math public claim.",
"level": 2,
"order": 25,
"section_id": "reader-evidence-routing",
"source_line_end": 363,
"source_line_start": 353,
"title": "Reader Evidence Routing"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/formal_math_premise_retrieval.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Formal math premise retrieval validates copied public Lean/Std premise metadata, query scoring, context budgets, strategy gates, body-floor provenance, and leakage negative cases without claiming proof authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "formal_math_premise_retrieval.py",
"path": "src/microcosm_core/organs/formal_math_premise_retrieval.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for first-wave premise-retrieval fixture validation, exported bundle validation, source-module/body-floor scanning, negative-case enforcement, card freshness, and scope limits.",
"symbols": [
"run",
"run_retrieval_bundle",
"_build_result",
"_scan_input_paths",
"_retrieval_bundle_freshness_basis",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING",
"BUNDLE_RESULT_NAME"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/formal_math_premise_retrieval.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_math_premise_retrieval.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/formal_math_premise_retrieval.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Formal Math Premise Retrieval is the source-backed retrieval slice between the Lean/Std premise catalog and proof-witness boundary. It validates eleven public premise descriptors, four retrieval queries, forty-four considered candidates, exact and source-faithful source body imports, context recipe byte budgets, strategy ids, card freshness, and five leakage/overclaim negative cases while keeping proof bodies, oracle premise ids, model-output data, Lean/Lake execution, and launch claims out of result records.",
"summary_line": "Formal math premise retrieval validates copied public Lean/Std premise metadata, query scoring, context budgets, strategy gates, body-floor provenance, and leakage negative cases without claiming proof authority.",
"title": "Formal Math Premise Retrieval"
},
{
"anchor": "paper-module-lean-std-premise-index",
"authority_ceiling": "Copied public Lean/Std descriptor index and Ring2 premise-retrieval source result records only; no Lean/Lake execution, Mathlib authority, proof-body import, oracle-needed premise authority, external model access, benchmark claim, launch-scope decision, publishing-scope decision, source-file changes, or theorem-correctness claim.",
"body_html": "
A premise catalogue before proof search
\n
A premise index is the catalogue a theorem-proving system reads before it tries to prove anything: the named lemmas and definitions it may cite, with enough metadata to retrieve the relevant ones. A premise index is the catalogue a theorem-proving system reads before it tries to prove anything: the named lemmas and definitions it may cite, with enough metadata to retrieve the relevant ones. lean_std_premise_index does not build such an index and does not run a prover. It takes a small public catalogue that was copied out of a private Ring2 premise-retrieval run and checks that the copy is faithful and metadata-only.
\n
The component consumes a fixture of JSON inputs, opens the source artifact the catalogue claims to copy, and emits metadata-only result records. The component consumes a fixture of JSON inputs, opens the source artifact the catalogue claims to copy, and emits metadata-only result records. A pass means the public rows match the opened source row for row after a single allowed path rewrite, the six copied body files match their declared digests and line and byte counts, and none of the five things that would turn the catalogue into an answer key slipped in.
\n
It reads metadata only. It never imports Mathlib, never runs Lean or Lake, and never prints a proof body. It reads metadata only. It never imports Mathlib, never runs Lean or Lake, and never prints a proof body.
\n
Purpose
\n
The value of a premise index for a benchmark sits one edit away from its danger. The same rows that tell a reader which standard-library facts are on the table would, with a proof body or an oracle-needed id attached, tell a solver exactly which lemmas it needs. The value of a premise index for a benchmark sits one edit away from its danger. The same rows that tell a reader which standard-library facts are on the table would, with a proof body or an oracle-needed id attached, tell a solver exactly which lemmas it needs. A loose copy is worse than no copy: it looks authoritative while leaking the answer.
\n
So the question this component answers is narrow. Given that a real premise index already exists inside a private run, can a cold reader see its public shape and be sure the public shape is a faithful copy and not a hand-written stand-in? So the question this component answers is narrow. Given that a real premise index already exists inside a private run, can a cold reader see its public shape and be sure the public shape is a faithful copy and not a hand-written stand-in? The answer rests on opening the declared source and comparing, not on trusting a hand-maintained mirror.
\n
How it works
\n
run validates the first-wave fixture with negative cases included. run_index_bundle validates the exported bundle with the source-module manifest required.run validates the first-wave fixture with negative cases included. run_index_bundle validates the exported bundle with the source-module manifest required. Both call _build_result, which runs the checks below in order and sets status to pass only when there are no positive findings, no missing negative case, and no blocking secret-scan hit.
\n
_validate_entries walks the eleven premise rows in the positive fixture. Each row must cite an Init/ Lean toolchain source ref, carry retrieval terms, and name a nonempty split drawn from train, dev, and test._validate_entries walks the eleven premise rows in the positive fixture. Each row must cite an Init/ Lean toolchain source ref, carry retrieval terms, and name a nonempty split drawn from train, dev, and test. The index as a whole must carry at least ten rows and cover the namespaces Nat, Bool, List, and Iff. A row is rejected if it carries a Mathlib ref, a proof body, an oracle_needed_premise_ids field, or a duplicate premise id.
\n
_validate_source_artifact is the check that makes the copy claim real. It reads the declared source_ref, recomputes the file's SHA-256, and compares it to the declared digest._validate_source_artifact is the check that makes the copy claim real. It reads the declared source_ref, recomputes the file's SHA-256, and compares it to the declared digest. It then matches every public row to a source row by premise_id and compares signatures. The one permitted difference is a path rewrite: _normalize_source_ref turns a raw Lean toolchain path into a public lean-toolchain://.../Init/... reference so the reader sees where a lemma lives without seeing a private filesystem. A digest mismatch, a row-count mismatch, or a signature mismatch blocks the record.
\n
_source_module_manifest_result checks the six copied body files: one normalized Lean/Std descriptor index plus five Ring2 body files._source_module_manifest_result checks the six copied body files: one normalized Lean/Std descriptor index plus five Ring2 body files. For each it recomputes the target digest, line count, and byte count against the manifest, confirms the declared source file opens and matches its own digest, and requires body_copied true with body_in_receipt false. Exact copies must still equal their opened source. The body text stays out of every record.
\n
The five expected negative cases are declared in EXPECTED_NEGATIVE_CASES. _build_result runs each negative input through the same entry and policy checks, collects the codes it actually observed, and blocks if any expected case was not raised. The five expected negative cases are declared in EXPECTED_NEGATIVE_CASES. _build_result runs each negative input through the same entry and policy checks, collects the codes it actually observed, and blocks if any expected case was not raised. A silent secret-exclusion scan over the inputs is the last gate.
\n
Function
Role
run
Validate the first-wave fixture, negative cases included
run_index_bundle
Validate the exported bundle, source-module manifest required
_build_result
Order the checks and decide pass or blocked
_validate_entries
Per-row and whole-index checks over the premise rows
_validate_source_artifact
Open the declared source, rehash it, compare rows by id
_source_module_manifest_result
Verify the six copied body files by digest, lines, bytes
_normalize_source_ref
Rewrite a raw toolchain path to a public lean-toolchain:// ref
\nDiagram of the mechanism (6 steps).Diagram source & refs
flowchart TD inputs["fixture inputs premise_index, projection_protocol, index_policy, 5 negative cases, source_module_manifest"] entries["_validate_entries 11 rows, Init/ refs, namespaces, splits"] source["_validate_source_artifact reopen source, rehash SHA-256, compare rows by premise_id"] manifest["_source_module_manifest_result 6 copied body files by digest, line, byte"] gate["_build_result no findings + all negatives seen + clean secret scan"] record["metadata-only result record"] inputs --> entries inputs --> source inputs --> manifest entries --> gate source --> gate manifest --> gate gate --> record
\n
Negative cases
\n
The fixture ships five inputs that must be rejected, each mapped to its error code in EXPECTED_NEGATIVE_CASES: The fixture ships five inputs that must be rejected, each mapped to its error code in EXPECTED_NEGATIVE_CASES:
If any expected code is not observed, the run is blocked. These stay stable because the index is meant to be useful without becoming proof authority. If any expected code is not observed, the run is blocked. These stay stable because the index is meant to be useful without becoming proof authority.
\n
Prior Art Grounding
\n
This component borrows the closed-index discipline from formal-library indexing and premise-selection work. The Lean mathematical library anchors the library-as-corpus side. This component borrows the closed-index discipline from formal-library indexing and premise-selection work. The Lean mathematical library anchors the library-as-corpus side. LeanDojo and HOList anchor the need for premise metadata, retrieval splits, and theorem-proving environments a learning system can inspect. Premise ids, declaration names, namespaces, source refs, retrieval terms, split eligibility, and copied-body digests are public metadata here; proof bodies and oracle-needed ids stay outside the boundary. It does not import Mathlib or prove theorems.
\n
Validation Result record Path
\n
Run these from the public clone root without mutating durable result record or generated projection surfaces: Run these from the public clone root without mutating durable result record or generated projection surfaces:
The focused test drives the runner over the fixture and the exported bundle: it asserts all five negative cases are observed, opens the real Ring2 source artifact, and rejects digest, row-count, row-signature, source-ref, source-module digest, and rehash-body-swap mutations. The focused test drives the runner over the fixture and the exported bundle: it asserts all five negative cases are observed, opens the real Ring2 source artifact, and rejects digest, row-count, row-signature, source-ref, source-module digest, and rehash-body-swap mutations. A pass means the copy was faithful and metadata-only on this input, not that any theorem is correct.
\n
Scope boundary
\n
Scope limit
\n
The bounded premise-index claim
\n
The strongest claim the evidence supports is bounded: public Lean/Std premise metadata, source refs, retrieval terms, split eligibility, and copied source-module digests can be indexed and shown to match their opened source without exposing proof bodies or oracle-needed ids. The strongest claim the evidence supports is bounded: public Lean/Std premise metadata, source refs, retrieval terms, split eligibility, and copied source-module digests can be indexed and shown to match their opened source without exposing proof bodies or oracle-needed ids. That is the proof boundary, and it is also the scope limit. A reader should not treat this page, the generated projections, or the validation result records as Lean or Lake execution, Mathlib authority, a proof-body import, oracle-needed id authority, a external model access, benchmark correctness, publishing-scope decision, or launch-scope decision. It checks a copy. It does not certify a proof.
\n",
"components": [
{
"anchor": "component-lean_std_premise_index",
"component_id": "lean_std_premise_index",
"public_label": "Lean Std Premise Index"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 2,
"organ": 1,
"paper_module": 1,
"principle": 5
},
"link_count": 14,
"links": [
{
"id": "lean_std_premise_index",
"kind": "organ",
"kind_label": "component",
"label": "Lean Std Premise Index",
"relation": "explains",
"url": "component-lean-std-premise-index.html"
},
{
"id": "mechanism.lean_std_premise_index.validates_public_lean_std_premise_catalog",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public lean std premise catalog",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-lean-std-premise-index-validates-public-lean-std-premise-catalog"
},
{
"id": "mechanism.lean_std_premise_index.validates_public_lean_std_premise_index",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public lean std premise index",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-lean-std-premise-index-validates-public-lean-std-premise-index"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.formal_math_premise_retrieval",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Premise Retrieval",
"relation": "depends on",
"url": "paper-module-formal-math-premise-retrieval.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[23:paper_module.formal_math_verifier_trace_repair_loop].depends_on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
},
{
"id": "paper_module.mathematical_strategy_atlas",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mathematical Strategy Atlas",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[43:paper_module.mathematical_strategy_atlas].depends_on",
"url": "paper-module-mathematical-strategy-atlas.html"
},
{
"id": "paper_module.undeclared_library_prior_classifier",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Undeclared Library Prior Classifier",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[55:paper_module.undeclared_library_prior_classifier].depends_on",
"url": "paper-module-undeclared-library-prior-classifier.html"
},
{
"id": "paper_module.verifier_lab_kernel",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Kernel",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[0:paper_module.verifier_lab_kernel].depends_on",
"url": "paper-module-verifier-lab-kernel.html"
}
],
"paper_module_id": "paper_module.lean_std_premise_index",
"reader_packet": {
"abstract": "The Lean/Std premise index validates a copied public Lean/Std descriptor catalog plus Ring2 premise-retrieval source bodies without claiming proof, Mathlib, Lean/Lake, provider, launch, or theorem-correctness authority.",
"authority_ceiling": "Copied public Lean/Std descriptor index and Ring2 premise-retrieval source result records only; no Lean/Lake execution, Mathlib authority, proof-body import, oracle-needed premise authority, external model access, benchmark claim, launch-scope decision, publishing-scope decision, source-file changes, or theorem-correctness claim.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.lean_std_premise_index.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-lean-std-premise-index-diagram-1",
"html_url": "docs/paper-module-lean-std-premise-index.html#paper-module-lean-std-premise-index-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD inputs[\"fixture inputs premise_index, projection_protocol, index_policy, 5 negative cases, source_module_manifest\"] entries[\"_validate_entries 11 rows, Init/ refs, namespaces, splits\"] source[\"_validate_source_artifact reopen source, rehash SHA-256, compare rows by premise_id\"] manifest[\"_source_module_manifest_result 6 copied body files by digest, line, byte\"] gate[\"_build_result no findings + all negatives seen + clean secret scan\"] record[\"metadata-only result record\"] inputs --> entries inputs --> source inputs --> manifest entries --> gate source --> gate manifest --> gate gate --> record",
"source_format": "mermaid",
"source_hash": "sha256:c69a029d7281ecba6ccff6ef641a081c7e75d71f9f8750e6257b6d475d0c4128",
"source_line_end": 126,
"source_line_start": 110
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.formal_math_premise_retrieval",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Premise Retrieval",
"relation": "depends on",
"url": "paper-module-formal-math-premise-retrieval.html"
}
],
"relation": "depends on"
},
{
"count": 3,
"links": [
{
"id": "lean_std_premise_index",
"kind": "organ",
"kind_label": "component",
"label": "Lean Std Premise Index",
"relation": "explains",
"url": "component-lean-std-premise-index.html"
},
{
"id": "mechanism.lean_std_premise_index.validates_public_lean_std_premise_catalog",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public lean std premise catalog",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-lean-std-premise-index-validates-public-lean-std-premise-catalog"
},
{
"id": "mechanism.lean_std_premise_index.validates_public_lean_std_premise_index",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public lean std premise index",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-lean-std-premise-index-validates-public-lean-std-premise-index"
}
],
"relation": "explains"
},
{
"count": 6,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.lean_std_premise_index",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 201,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/lean_std_premise_index.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Lean mathematical library",
"section_id": "this-component-borrows-the-closed-index-discipline-from-formal-library-indexing-and-premise-selection-work",
"url": "https://arxiv.org/abs/1910.09336"
},
{
"kind": "external_reference",
"label": "LeanDojo",
"section_id": "this-component-borrows-the-closed-index-discipline-from-formal-library-indexing-and-premise-selection-work",
"url": "https://arxiv.org/abs/2306.15626"
},
{
"kind": "external_reference",
"label": "HOList",
"section_id": "this-component-borrows-the-closed-index-discipline-from-formal-library-indexing-and-premise-selection-work",
"url": "https://arxiv.org/abs/1904.03241"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "A premise index is the catalogue a theorem-proving system reads before it tries to prove anything: the named lemmas and definitions it may cite, with enough metadata to retrieve the relevant ones. A premise index is the catalogue a theorem-proving system reads before it tries to prove anything: the named lemmas and definitions it may cite, with enough metadata to retrieve the relevant ones. leanstdpremise_index does not build such an index and does not run a prover. It takes a small public catalogue that was copied out of a private Ring2 premise-retrieval run and checks that the copy is faithful and metadata-only.",
"level": 4,
"order": 1,
"section_id": "a-premise-catalogue-before-proof-search",
"source_line_end": 12,
"source_line_start": 3,
"title": "A premise catalogue before proof search"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component consumes a fixture of JSON inputs, opens the source artifact the catalogue claims to copy, and emits metadata-only result records. The component consumes a fixture of JSON inputs, opens the source artifact the catalogue claims to copy, and emits metadata-only result records. A pass means the public rows match the opened source row for row after a single allowed path rewrite, the six copied body files match their declared digests and line and byte counts, and none of the five things that would turn the catalogue into an answer key slipped in.",
"level": 4,
"order": 2,
"section_id": "the-component-consumes-a-fixture-of-json-inputs",
"source_line_end": 22,
"source_line_start": 13,
"title": "The component consumes a fixture of JSON inputs"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It reads metadata only. It never imports Mathlib, never runs Lean or Lake, and never prints a proof body. It reads metadata only. It never imports Mathlib, never runs Lean or Lake, and never prints a proof body.",
"level": 4,
"order": 3,
"section_id": "it-reads-metadata-only",
"source_line_end": 28,
"source_line_start": 23,
"title": "It reads metadata only"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 30,
"source_line_start": 29,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The value of a premise index for a benchmark sits one edit away from its danger. The same rows that tell a reader which standard-library facts are on the table would, with a proof body or an oracle-needed id attached, tell a solver exactly which lemmas it needs. The value of a premise index for a benchmark sits one edit away from its danger. The same rows that tell a reader which standard-library facts are on the table would, with a proof body or an oracle-needed id attached, tell a solver exactly which lemmas it needs. A loose copy is worse than no copy: it looks authoritative while leaking the answer.",
"level": 4,
"order": 5,
"section_id": "the-value-of-a-premise-index-for-a-benchmark-sits-one-edit-away-from-its-danger",
"source_line_end": 39,
"source_line_start": 31,
"title": "The value of a premise index for a benchmark sits one edit away from its danger"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "So the question this component answers is narrow. Given that a real premise index already exists inside a private run, can a cold reader see its public shape and be sure the public shape is a faithful copy and not a hand-written stand-in? So the question this component answers is narrow. Given that a real premise index already exists inside a private run, can a cold reader see its public shape and be sure the public shape is a faithful copy and not a hand-written stand-in? The answer rests on opening the declared source and comparing, not on trusting a hand-maintained mirror.",
"level": 4,
"order": 6,
"section_id": "so-the-question-this-component-answers-is-narrow",
"source_line_end": 48,
"source_line_start": 40,
"title": "So the question this component answers is narrow"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 50,
"source_line_start": 49,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run validates the first-wave fixture with negative cases included. runindexbundle validates the exported bundle with the source-module manifest required. run validates the first-wave fixture with negative cases included. runindexbundle validates the exported bundle with the source-module manifest required. Both call buildresult, which runs the checks below in order and sets status to pass only when there are no positive findings, no missing negative case, and no blocking secret-scan hit.",
"level": 4,
"order": 8,
"section_id": "run-validates-the-first-wave-fixture-with-negative-cases-included",
"source_line_end": 59,
"source_line_start": 51,
"title": "`run` validates the first-wave fixture with negative cases included"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateentries walks the eleven premise rows in the positive fixture. Each row must cite an Init/ Lean toolchain source ref, carry retrieval terms, and name a nonempty split drawn from train, dev, and test.** validateentries walks the eleven premise rows in the positive fixture. Each row must cite an Init/ Lean toolchain source ref, carry retrieval terms, and name a nonempty split drawn from train, dev, and test. The index as a whole must carry at least ten rows and cover the namespaces Nat, Bool, List, and Iff. A row is rejected if it carries a Mathlib ref, a proof body, an oracleneededpremise_ids field, or a duplicate premise id.",
"level": 4,
"order": 9,
"section_id": "validate-entries-walks-the-eleven-premise-rows-in-the-positive-fixture",
"source_line_end": 69,
"source_line_start": 60,
"title": "`_validate_entries` walks the eleven premise rows in the positive fixture"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatesourceartifact is the check that makes the copy claim real. It reads the declared sourceref, recomputes the file's SHA-256, and compares it to the declared digest.** validatesource_artifact is the check that makes the copy claim real. It reads the declared source_ref, recomputes the file's SHA-256, and compares it to the declared digest. It then matches every public row to a source row by premise_id and compares signatures. The one permitted difference is a path rewrite: normalizesource_ref turns a raw Lean toolchain path into a public lean-toolchain://.../Init/... reference so the reader sees where a lemma lives without seeing a private filesystem. A digest mismatch, a row-count mi\u2026",
"level": 4,
"order": 10,
"section_id": "validate-source-artifact-is-the-check-that-makes-the-copy-claim-real",
"source_line_end": 81,
"source_line_start": 70,
"title": "`_validate_source_artifact` is the check that makes the copy claim real"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourcemodulemanifestresult checks the six copied body files: one normalized Lean/Std descriptor index plus five Ring2 body files.** sourcemodulemanifestresult checks the six copied body files: one normalized Lean/Std descriptor index plus five Ring2 body files. For each it recomputes the target digest, line count, and byte count against the manifest, confirms the declared source file opens and matches its own digest, and requires bodycopied true with bodyin_receipt false. Exact copies must still equal their opened source. The body text stays out of every record.",
"level": 4,
"order": 11,
"section_id": "source-module-manifest-result-checks-the-six-copied-body-files",
"source_line_end": 91,
"source_line_start": 82,
"title": "`_source_module_manifest_result` checks the six copied body files"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "The five expected negative cases are declared in EXPECTEDNEGATIVECASES. buildresult runs each negative input through the same entry and policy checks, collects the codes it actually observed, and blocks if any expected case was not raised. The five expected negative cases are declared in EXPECTEDNEGATIVECASES. buildresult runs each negative input through the same entry and policy checks, collects the codes it actually observed, and blocks if any expected case was not raised. A silent secret-exclusion scan over the inputs is the last gate.",
"level": 4,
"order": 12,
"section_id": "the-five-expected-negative-cases-are-declared-in-expected-negative-cases",
"source_line_end": 127,
"source_line_start": 92,
"title": "The five expected negative cases are declared in `EXPECTED_NEGATIVE_CASES`"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 129,
"source_line_start": 128,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The fixture ships five inputs that must be rejected, each mapped to its error code in EXPECTEDNEGATIVECASES: The fixture ships five inputs that must be rejected, each mapped to its error code in EXPECTEDNEGATIVECASES: mathlibpremiseforbidden raises LEANSTDINDEXMATHLIBFORBIDDEN; proofbodyleakage raises LEANSTDINDEXPROOFBODY_FORBIDDEN; oracleneededidsleakage raises LEANSTDINDEXORACLEIDSFORBIDDEN; testsplittuningattempt raises LEANSTDINDEXTESTSPLITTUNING_FORBIDDEN; namespacewithoutsourceref raises LEANSTDINDEXSOURCEREFREQUIRED.",
"level": 4,
"order": 14,
"section_id": "the-fixture-ships-five-inputs-that-must-be-rejected",
"source_line_end": 141,
"source_line_start": 130,
"title": "The fixture ships five inputs that must be rejected"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "If any expected code is not observed, the run is blocked. These stay stable because the index is meant to be useful without becoming proof authority. If any expected code is not observed, the run is blocked. These stay stable because the index is meant to be useful without becoming proof authority.",
"level": 4,
"order": 15,
"section_id": "if-any-expected-code-is-not-observed-the-run-is-blocked",
"source_line_end": 147,
"source_line_start": 142,
"title": "If any expected code is not observed, the run is blocked"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 149,
"source_line_start": 148,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "This component borrows the closed-index discipline from formal-library indexing and premise-selection work. The Lean mathematical library anchors the library-as-corpus side. This component borrows the closed-index discipline from formal-library indexing and premise-selection work. The [Lean mathematical library](https://arxiv.org/abs/1910.09336) anchors the library-as-corpus side. LeanDojo and HOList anchor the need for premise metadata, retrieval splits, and theorem-proving environments a learning system can inspect. Premise ids, declaration names, namespaces, source refs, retrieval terms, split eligibility, and copied-body digests are public metadata here; proof bodies and oracle-needed i\u2026",
"level": 4,
"order": 17,
"section_id": "this-component-borrows-the-closed-index-discipline-from-formal-library-indexing-and-premise-selection-work",
"source_line_end": 163,
"source_line_start": 150,
"title": "This component borrows the closed-index discipline from formal-library indexing and premise-selection work"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 165,
"source_line_start": 164,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run these from the public clone root without mutating durable result record or generated projection surfaces: Run these from the public clone root without mutating durable result record or generated projection surfaces:",
"level": 4,
"order": 19,
"section_id": "run-these-from-the-public-clone-root-without-mutating-durable-receipt-or-generated-projection",
"source_line_end": 176,
"source_line_start": 166,
"title": "Run these from the public clone root without mutating durable result record or generated projection"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The focused test drives the runner over the fixture and the exported bundle: it asserts all five negative cases are observed, opens the real Ring2 source artifact, and rejects digest, row-count, row-signature, source-ref, source-module digest, and rehash-body-swap mutations. The focused test drives the runner over the fixture and the exported bundle: it asserts all five negative cases are observed, opens the real Ring2 source artifact, and rejects digest, row-count, row-signature, source-ref, source-module digest, and rehash-body-swap mutations. A pass means the copy was faithful and metadata-only on this input, not that any theorem is correct.",
"level": 4,
"order": 20,
"section_id": "the-focused-test-drives-the-runner-over-the-fixture-and-the-exported-bundle",
"source_line_end": 185,
"source_line_start": 177,
"title": "The focused test drives the runner over the fixture and the exported bundle"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 187,
"source_line_start": 186,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim the evidence supports is bounded: public Lean/Std premise metadata, source refs, retrieval terms, split eligibility, and copied source-module digests can be indexed and shown to match their opened source without exposing proof bodies or oracle-needed ids. The strongest claim the evidence supports is bounded: public Lean/Std premise metadata, source refs, retrieval terms, split eligibility, and copied source-module digests can be indexed and shown to match their opened source without exposing proof bodies or oracle-needed ids. That is the proof boundary, and it is also the scope limit. A reader should not treat this page, the generated projections, or the validation resul\u2026",
"level": 4,
"order": 22,
"section_id": "the-bounded-premise-index-claim",
"source_line_end": 200,
"source_line_start": 188,
"title": "The bounded premise-index claim"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/lean_std_premise_index.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The Lean/Std premise index validates a copied public Lean/Std descriptor catalog plus Ring2 premise-retrieval source bodies without claiming proof, Mathlib, Lean/Lake, provider, launch, or theorem-correctness authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "lean_std_premise_index.py",
"path": "src/microcosm_core/organs/lean_std_premise_index.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for first-wave Lean/Std premise-index fixture validation, exported bundle validation, source-module manifest checks, negative-case enforcement, and scope limits.",
"symbols": [
"run",
"run_index_bundle",
"_source_module_manifest_result",
"_build_result",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING",
"SOURCE_MODULE_MANIFEST_NAME"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/lean_std_premise_index.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/lean_std_premise_index.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/lean_std_premise_index.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Lean/Std Premise Index is the source-open formal-math catalog component for Microcosm. It imports a premise descriptor index, validates eleven Lean/Std premise rows across Nat, Bool, List, and Iff namespaces, checks six copied body modules through a source-module manifest, observes Mathlib/proof-body/oracle/test-split/source-ref negative cases, and writes metadata-only result records that make premise system inspectable without turning metadata into proof authority.",
"summary_line": "The Lean/Std premise index validates a copied public Lean/Std descriptor catalog plus Ring2 premise-retrieval source bodies without claiming proof, Mathlib, Lean/Lake, provider, launch, or theorem-correctness authority.",
"title": "Lean/Std Premise Index"
},
{
"anchor": "paper-module-world-model-projection-drift-control-room",
"authority_ceiling": "Public metadata-only runtime result record and copied source-module evidence only; no private runtime body inspection, source authority, source-file changes, live route repair, automatic doctrine changes, model-output data export, launch-scope decision, publishing-scope decision, or whole-system correctness claim.",
"body_html": "
A projected read model is a copy that can drift from its source and quietly read as truth, so this component checks public drift rows still trace back and refuses to let a copy speak for its source. A projected read model is a copy. A dashboard row, a generated structured source record, a route card, or a public runtime result record can stay useful for navigation long after it has stopped matching the source it was copied from. When that happens quietly, the copy starts to read as truth. This component checks that public drift rows still trace back to the source they describe, and it refuses to let a copy speak for its source.
\n
Recomputing drift instead of trusting declarations
\n
The runtime recomputes drift rows from a public result record rather than trusting the supplied file, diffs each against a real source-state artifact, and writes metadata-only records. The runtime is src/microcosm_core/organs/world_model_projection_drift_control_room.py. It takes a small bundle of public JSON inputs, recomputes the drift rows from a public runtime result record rather than trusting the supplied file, diffs each row against a real source-state artifact, and writes metadata-only result records. Every record carries paths, hashes, counts, and verdicts. None of them carries a private body.
\n
Source-linked guards for each drift row
\n
A pass means each named drift row is guarded by a source, target, repair route, validation reference, and fact-authority record and was witnessed, not that any source was repaired or made authoritative. A pass here means the named public drift rows are guarded: each names its source, its target, a repair route, a validation reference, and a fact-authority record, and each was witnessed in the runtime result record. It does not mean any source was repaired, mutated, or made authoritative.
\n
Purpose
\n
Drift between a world model and projection
\n
The failure mode is drift between a world model and its projection, where trusting the result record as source lets a stale or edited row pass a shape check while matching nothing real. The failure mode is drift between a world model and its projection. The reducer world_model.py and the /api/drift endpoint in main.py produce a live view. A public result record then projects that view for readers. If the result record is trusted as the source, a stale or edited row can pass a shape check while no longer matching anything real.
\n
Recompute, never trust the supplied file
\n
Rather than a shape check, _build_result recomputes rows from the result record, treats the supplied file as an expected snapshot only, and blocks any row it cannot re-derive from source. A shape check alone fails here, so the design does not do one. The supplied drift_rows.json is never treated as input. _build_result recomputes the rows from the public runtime result record, then treats the supplied file only as an expected snapshot whose role is recorded as expected_snapshot_not_source_authority. If the snapshot disagrees with the recomputed rows, that is staleness, not fact. Each recomputed row is then diffed against a real source-state artifact. A row that cannot be re-derived from source, or whose guard reference or derivation path has changed, moves the verdict to blocked.
\n
A repair route stays a label
\n
A named repair route stays a label, and _row_policy_findings rejects any row authorising live repair, source-file changes, doctrine changes, or launch, so a row names a fix but cannot act. The same limit holds the other way. A row may name a repair route, but the route stays a label. _row_policy_findings rejects any row that authorises live repair, source-file changes, automatic doctrine changes, or launch. A row here can say what is wrong and where to go next. It cannot act, and it cannot stand in for its source.
\n
How it works
\n
Two entry points over _build_result
\n
run validates the fixture with negative cases and run_drift_control_bundle validates the exported bundle without them, both flowing through _build_result into _write_receipts.run validates the accepted fixture with negative cases included. run_drift_control_bundle validates the exported public bundle without them. Both call _build_result, which runs the stages below in order and then hands the payload to _write_receipts.
\n
Function
Role
_projection_recompute_result
A selected id absent from the result record blocks.
_supplied_drift_rows_snapshot_result
Compares the supplied drift_rows.json against the recomputed rows. Records the supplied file as an expected snapshot only. Any missing, extra, changed, or metadata-mismatched row is flagged as stale.
_source_ref_evidence_result
Checks that each row source_ref resolves to an entry in SUPPORTED_DRIFT_SOURCE_REFS. An unknown reference blocks the row.
_source_state_diff_result
Diffs recomputed rows against real source-state artifacts: a row in the extracted-pattern ledger, and the view-quality action-map lens whose action_summary is re-derived from its own action_rows. The two evidence classes are extracted_pattern_ledger_row_diff and view_quality_action_map_summary_diff.
_source_module_manifest_result
Reads source_module_manifest.json, and for each module checks body_copied is true, body_in_receipt is false, the copied body's sha256 matches both declared digests, and the declared anchors appear in the body. A missing target, digest mismatch, or missing anchor blocks.
_source_open_body_import_summary
Folds the manifest result into a metadata-only summary: material status, copied count, and a reader action that points at the bundle rather than at any result record body.
_view_quality_geometry_grade_result
Imports the copied view_quality_census.py, grades view_quality_geometry_probe.json through its own _geometry_vector_from_summary and _geometry_calibration_review, and requires the verdict to be calibrated_pass.
_runtime_receipt_witness_result
Requires every recomputed row id to appear in the public runtime result record. An unwitnessed row blocks.
_row_policy_findings
Per-row gate: required fields, a source reference, a repair route and validation reference, a target reference, body_in_receipt false, no private needles, no model-output data, and a complete fact-authority record whose authority_ref equals the source, whose guard_ref equals the validation reference, and whose residual_route equals the repair route.
\n
_build_result gates the overall status
\n
_build_result passes only when every stage and the secret scan pass with no findings, and the accepted fixture yields eight witnessed drift rows with all authority counts at zero._build_result collects the findings from every stage. The overall status is PASS only when there are drift rows, no findings, the recomputed rows are metadata-only, all forbidden booleans are false, the runtime witness, recompute, snapshot, source-ref, source-state, source-module, and geometry stages all pass, and the secret-exclusion scan passes. Anything else is blocked. The accepted fixture produces eight drift rows, all eight witnessed in the runtime result record, with source_authority_claim_count, live_repair_authorized_count, source_mutation_authorized_count, and automatic_doctrine_promotion_count at zero.
\n
_write_receipts normalises and pins authority
\n
_write_receipts writes four records with host-temporary paths normalised away and body_in_receipt, release_authorized, and AUTHORITY_CEILING pinned on every one._write_receipts then writes the result record, a board, a validation record, and a fixture_acceptance record. Paths are normalised so host-temporary paths never surface. Every record sets body_in_receipt and release_authorized to false and repeats AUTHORITY_CEILING.
\nDiagram of the mechanism (8 steps).Diagram source & refs
Source refs
Public runtime result record
public_projection_drift_control_lens.json
expected snapshot, source-linked only
Supplied drift_rows.json
Geometry grade
copied view_quality_census.py
flowchart TD Result record["Public runtime result record public_projection_drift_control_lens.json"] Recompute["Recompute rows selected_pattern_ids + result record drift_rows"] Snapshot["Supplied drift_rows.json expected snapshot, source-linked only"] SourceDiff["Source-state diff pattern ledger + view-quality action map"] Geometry["Geometry grade copied view_quality_census.py"] Witness["Runtime witness every row appears in the result record"] Policy["Per-row policy gate refs, fact-authority, forbidden booleans"] Records["metadata-only records result, board, validation, fixture_acceptance"] Result record --> Recompute Recompute --> Snapshot Recompute --> SourceDiff Recompute --> Geometry Recompute --> Witness Recompute --> Policy Snapshot --> Records SourceDiff --> Records Geometry --> Records Witness --> Records Policy --> Records
\n
Four copied source bodies in the bundle
\n
The exported bundle copies four public source bodies, the reducer, the /api/drift endpoint, the action-map builder, and its test, so a reader inspects the implementation without private state. The exported bundle copies four public source bodies so a reader can inspect the implementation without receiving private state: the world-model reducer system/server/world_model.py, the /api/drift endpoint in system/server/main.py, the action-map builder tools/meta/observability/view_quality_census.py, and its regression test system/server/tests/test_view_quality_census.py. The bodies live in the bundle. The records carry only their paths, hashes, and verdicts.
\n
Negative cases
\n
Nine negative inputs the run requires
\n
run loads nine named negative inputs defined in EXPECTED_NEGATIVE_CASES and requires each to produce its rejection code, blocking the run if any expected code is missing.run loads nine named negative inputs and requires each to produce its expected rejection code, defined in EXPECTED_NEGATIVE_CASES. A missing expected code blocks the run.
\n
The nine cases and added source-drift cases
\n
The nine cases each map a malformed or overclaiming row to a stable code, and the source-drift stages add cases where a coherent row still loses authority if its source evidence changes. The nine cases are: drift_row_without_source_ref to DRIFT_SOURCE_REF_REQUIRED; repair_route_without_validation_ref to DRIFT_VALIDATION_REF_REQUIRED; drift_row_without_fact_authority to DRIFT_FACT_AUTHORITY_REQUIRED; projection_claiming_source_authority to DRIFT_SOURCE_AUTHORITY_FORBIDDEN; live_repair_action_authorized to DRIFT_LIVE_REPAIR_FORBIDDEN; private_runtime_data_export to DRIFT_PRIVATE_RUNTIME_EXPORT_FORBIDDEN; provider_payload_export to DRIFT_PROVIDER_PAYLOAD_FORBIDDEN; automatic_doctrine_promotion to DRIFT_AUTOMATIC_DOCTRINE_PROMOTION_FORBIDDEN; and release_from_drift_projection to DRIFT_RELEASE_AUTHORITY_FORBIDDEN. The source-drift stages add further cases for unwitnessed rows, stale supplied snapshots, missing source-ledger rows, mutated view-quality summaries, and internally consistent but unsupported source references. A row can be internally coherent and still lose authority if its source evidence, guard, or derivation path changes.
\n
Prior Art Grounding
\n
Concept-drift monitoring as signal
\n
It borrows the model-monitoring and concept-drift idea of a growing predicted-versus-observed gap as an operational signal, yielding fixture-bound monitoring evidence, source-linked only. This component watches a projection for drift between what the model expects and what the runtime reports. It draws on the model-monitoring and concept-drift literature, which treats a growing gap between predicted and observed behaviour as an operational signal. The borrowed shape is drift-as-signal over metadata-only records. The result is fixture-bound monitoring evidence, source-linked only or whole-system correctness.
A pass means a fixture run produced the described records and rejected every named negative case. A pass means the component ran over the public fixture, produced the records described here, and rejected every named negative case.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
Over public synthetic inputs the named drift rows are guarded by references, per-row fact-authority, copied source-body evidence, and negative-case rejection, all reproducing under the validation commands. The strongest claim this evidence supports is narrow: over public synthetic inputs, the named drift rows are guarded by source references, target references, repair routes, validation references, a fact-authority record per row, copied source-body evidence, metadata-only records, and negative-case rejection, and all of this reproduces under the validation commands above.
\n
The proof boundary and scope limit
\n
The proof boundary is the metadata-only record, copied manifest, fixture rows, and checks on this page, and the ceiling is public metadata-only evidence only, with no source authority, mutation, or launch. It refuses everything past that. The proof boundary is the metadata-only runtime record, the copied source-module manifest, the fixture rows, the source-state and geometry checks, and the validation result records named on this page. The scope limit is public metadata-only evidence only: no private runtime body inspection, no source authority, no source-file changes, no live route repair, no automatic doctrine changes, no model-output data export, no launch-scope decision, no publishing-scope decision, and no whole-system correctness claim.
\n",
"components": [
{
"anchor": "component-world_model_projection_drift_control_room",
"component_id": "world_model_projection_drift_control_room",
"public_label": "Projection-Drift Contract Validator"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 9
},
"link_count": 19,
"links": [
{
"id": "world_model_projection_drift_control_room",
"kind": "organ",
"kind_label": "component",
"label": "Projection-Drift Contract Validator",
"relation": "explains",
"url": "component-world-model-projection-drift-control-room.html"
},
{
"id": "mechanism.world_model_projection_drift_control_room.validates_public_projection_drift_control_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public projection drift control boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-world-model-projection-drift-control-room-validates-public-projection-drift-control-boundary"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch12_market_dashboard_read_model_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 12 Market Dashboard Read-Model Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[64:paper_module.batch12_market_dashboard_read_model_capsule].depends_on",
"url": "paper-module-batch12-market-dashboard-read-model-capsule.html"
},
{
"id": "paper_module.batch8_structural_theses_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Structural Theses Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[61:paper_module.batch8_structural_theses_capsule].depends_on",
"url": "paper-module-batch8-structural-theses-capsule.html"
},
{
"id": "paper_module.materials_chemistry_closed_loop_lab_safety_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Materials Chemistry Closed-Loop Lab-Safety Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[6:paper_module.materials_chemistry_closed_loop_lab_safety_replay].depends_on",
"url": "paper-module-materials-chemistry-closed-loop-lab-safety-replay.html"
},
{
"id": "paper_module.mechanistic_interpretability_circuit_attribution_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mechanistic Interpretability Circuit Attribution Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[51:paper_module.mechanistic_interpretability_circuit_attribution_replay].depends_on",
"url": "paper-module-mechanistic-interpretability-circuit-attribution-replay.html"
},
{
"id": "paper_module.prediction_oracle_reconciliation",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Prediction Oracle Reconciliation",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[53:paper_module.prediction_oracle_reconciliation].depends_on",
"url": "paper-module-prediction-oracle-reconciliation.html"
},
{
"id": "paper_module.spatial_world_model_counterfactual_simulation_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Spatial World Model Counterfactual Simulation Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[52:paper_module.spatial_world_model_counterfactual_simulation_replay].depends_on",
"url": "paper-module-spatial-world-model-counterfactual-simulation-replay.html"
},
{
"id": "paper_module.tool_server_pressure_inventory",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tool Server Pressure Inventory",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[50:paper_module.tool_server_pressure_inventory].depends_on",
"url": "paper-module-tool-server-pressure-inventory.html"
}
],
"paper_module_id": "paper_module.world_model_projection_drift_control_room",
"reader_packet": {
"abstract": "The world-model projection drift control room validates public metadata-only projection-drift rows and copied source-module bodies without treating projections as source authority or repair authority.",
"authority_ceiling": "Public metadata-only runtime result record and copied source-module evidence only; no private runtime body inspection, source authority, source-file changes, live route repair, automatic doctrine changes, model-output data export, launch-scope decision, publishing-scope decision, or whole-system correctness claim.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.world_model_projection_drift_control_room.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-world-model-projection-drift-control-room-diagram-1",
"html_url": "docs/paper-module-world-model-projection-drift-control-room.html#paper-module-world-model-projection-drift-control-room-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Result record[\"Public runtime result record public_projection_drift_control_lens.json\"] Recompute[\"Recompute rows selected_pattern_ids + result record drift_rows\"] Snapshot[\"Supplied drift_rows.json expected snapshot, source-linked only\"] SourceDiff[\"Source-state diff pattern ledger + view-quality action map\"] Geometry[\"Geometry grade copied view_quality_census.py\"] Witness[\"Runtime witness every row appears in the result record\"] Policy[\"Per-row policy gate refs, fact-authority, forbidden booleans\"] Records[\"metadata-only records result, board, validation, fixture_acceptance\"] Result record --> Recompute Recompute --> Snapshot Recompute --> SourceDiff Recompute --> Geometry Recompute --> Witness Recompute --> Policy Snapshot --> Records SourceDiff --> Records Geometry --> Records Witness --> Records Policy --> Records",
"source_format": "mermaid",
"source_hash": "sha256:8feaa484d1644f839f898c8c3d321841a9217db349ca7e40bfa780e76d82195d",
"source_line_end": 137,
"source_line_start": 115
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "world_model_projection_drift_control_room",
"kind": "organ",
"kind_label": "component",
"label": "Projection-Drift Contract Validator",
"relation": "explains",
"url": "component-world-model-projection-drift-control-room.html"
},
{
"id": "mechanism.world_model_projection_drift_control_room.validates_public_projection_drift_control_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public projection drift control boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-world-model-projection-drift-control-room-validates-public-projection-drift-control-boundary"
}
],
"relation": "explains"
},
{
"count": 10,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.world_model_projection_drift_control_room",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 247,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/world_model_projection_drift_control_room.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "concept-drift",
"section_id": "concept-drift-monitoring-as-signal",
"url": "https://en.wikipedia.org/wiki/Concept_drift"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A projected read model is a copy that can drift from its source and quietly read as truth, so this component checks public drift rows still trace back and refuses to let a copy speak for its source.** A projected read model is a copy. A dashboard row, a generated structured source record, a route card, or a public runtime result record can stay useful for navigation long after it has stopped matching the source it was copied from. When that happens quietly, the copy starts to read as truth. This component checks that public drift rows still trace back to the source they describe, and it refuses to let a copy speak for its source.",
"level": 4,
"order": 1,
"section_id": "a-projected-read-model-is-a-copy-that-can-drift-from-its-source",
"source_line_end": 13,
"source_line_start": 3,
"title": "A projected read model is a copy that can drift from its source"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The runtime recomputes drift rows from a public result record rather than trusting the supplied file, diffs each against a real source-state artifact, and writes metadata-only records. The runtime is src/microcosmcore/components/worldmodelprojectiondriftcontrol_room.py. It takes a small bundle of public JSON inputs, recomputes the drift rows from a public runtime result record rather than trusting the supplied file, diffs each row against a real source-state artifact, and writes metadata-only result records. Every record carries paths, hashes, counts, and verdicts. None of them carries a private body.",
"level": 4,
"order": 2,
"section_id": "recomputing-drift-instead-of-trusting-declarations",
"source_line_end": 24,
"source_line_start": 14,
"title": "Recomputing drift instead of trusting declarations"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means each named drift row is guarded by a source, target, repair route, validation reference, and fact-authority record and was witnessed, not that any source was repaired or made authoritative.** A pass here means the named public drift rows are guarded: each names its source, its target, a repair route, a validation reference, and a fact-authority record, and each was witnessed in the runtime result record. It does not mean any source was repaired, mutated, or made authoritative.",
"level": 4,
"order": 3,
"section_id": "source-linked-guards-for-each-drift-row",
"source_line_end": 33,
"source_line_start": 25,
"title": "Source-linked guards for each drift row"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 35,
"source_line_start": 34,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The failure mode is drift between a world model and its projection, where trusting the result record as source lets a stale or edited row pass a shape check while matching nothing real.** The failure mode is drift between a world model and its projection. The reducer world_model.py and the /api/drift endpoint in main.py produce a live view. A public result record then projects that view for readers. If the result record is trusted as the source, a stale or edited row can pass a shape check while no longer matching anything real.",
"level": 4,
"order": 5,
"section_id": "drift-between-a-world-model-and-projection",
"source_line_end": 44,
"source_line_start": 36,
"title": "Drift between a world model and projection"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Rather than a shape check, build_result recomputes rows from the result record, treats the supplied file as an expected snapshot only, and blocks any row it cannot re-derive from source.** A shape check alone fails here, so the design does not do one. The supplied driftrows.json is never treated as input. build_result recomputes the rows from the public runtime result record, then treats the supplied file only as an expected snapshot whose role is recorded as expectedsnapshotnotsourceauthority. If the snapshot disagrees with the recomputed rows, that is staleness, not fact. Each recomputed row is then diffed against a real source-state artifact. A row that cannot be re-derived from source,\u2026",
"level": 4,
"order": 6,
"section_id": "recompute-never-trust-the-supplied-file",
"source_line_end": 58,
"source_line_start": 45,
"title": "Recompute, never trust the supplied file"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A named repair route stays a label, and rowpolicyfindings rejects any row authorising live repair, source-file changes, doctrine changes, or launch, so a row names a fix but cannot act.** The same limit holds the other way. A row may name a repair route, but the route stays a label. rowpolicy_findings rejects any row that authorises live repair, source-file changes, automatic doctrine changes, or launch. A row here can say what is wrong and where to go next. It cannot act, and it cannot stand in for its source.",
"level": 4,
"order": 7,
"section_id": "a-repair-route-stays-a-label",
"source_line_end": 68,
"source_line_start": 59,
"title": "A repair route stays a label"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 70,
"source_line_start": 69,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "run validates the fixture with negative cases and rundriftcontrolbundle validates the exported bundle without them, both flowing through buildresult into writereceipts.** run validates the accepted fixture with negative cases included. rundriftcontrol_bundle validates the exported public bundle without them. Both call buildresult, which runs the stages below in order and then hands the payload to writereceipts.",
"level": 4,
"order": 9,
"section_id": "two-entry-points-over-build-result",
"source_line_end": 91,
"source_line_start": 71,
"title": "Two entry points over `_build_result`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult passes only when every stage and the secret scan pass with no findings, and the accepted fixture yields eight witnessed drift rows with all authority counts at zero. build_result collects the findings from every stage. The overall status is PASS only when there are drift rows, no findings, the recomputed rows are metadata-only, all forbidden booleans are false, the runtime witness, recompute, snapshot, source-ref, source-state, source-module, and geometry stages all pass, and the secret-exclusion scan passes. Anything else is blocked. The accepted fixture produces eight drift rows, all eight witnessed in the runtime result record, with sourceauthorityclaimcount, liverepairauthor\u2026",
"level": 4,
"order": 10,
"section_id": "build-result-gates-the-overall-status",
"source_line_end": 105,
"source_line_start": 92,
"title": "`_build_result` gates the overall status"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "writereceipts writes four records with host-temporary paths normalised away and bodyinreceipt, releaseauthorized, and AUTHORITYCEILING pinned on every one. write_receipts then writes the result record, a board, a validation record, and a fixture_acceptance record. Paths are normalised so host-temporary paths never surface. Every record sets bodyinreceipt and release_authorized to false and repeats scope_limit.",
"level": 4,
"order": 11,
"section_id": "write-receipts-normalises-and-pins-authority",
"source_line_end": 138,
"source_line_start": 106,
"title": "`_write_receipts` normalises and pins authority"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The exported bundle copies four public source bodies, the reducer, the /api/drift endpoint, the action-map builder, and its test, so a reader inspects the implementation without private state.** The exported bundle copies four public source bodies so a reader can inspect the implementation without receiving private state: the world-model reducer system/server/world_model.py, the /api/drift endpoint in system/server/main.py, the action-map builder tools/meta/observability/viewqualitycensus.py, and its regression test system/server/tests/testviewquality_census.py. The bodies live in the bundle. The records carry only their paths, hashes, and verdicts.",
"level": 4,
"order": 12,
"section_id": "four-copied-source-bodies-in-the-bundle",
"source_line_end": 150,
"source_line_start": 139,
"title": "Four copied source bodies in the bundle"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 152,
"source_line_start": 151,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run loads nine named negative inputs defined in EXPECTEDNEGATIVE_CASES and requires each to produce its rejection code, blocking the run if any expected code is missing.** run loads nine named negative inputs and requires each to produce its expected rejection code, defined in EXPECTEDNEGATIVECASES. A missing expected code blocks the run.",
"level": 4,
"order": 14,
"section_id": "nine-negative-inputs-the-run-requires",
"source_line_end": 160,
"source_line_start": 153,
"title": "Nine negative inputs the run requires"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The nine cases each map a malformed or overclaiming row to a stable code, and the source-drift stages add cases where a coherent row still loses authority if its source evidence changes. The nine cases are: driftrowwithoutsource_ref to DRIFTSOURCEREFREQUIRED; repairroutewithoutvalidation_ref to DRIFTVALIDATIONREFREQUIRED; driftrowwithoutfact_authority to DRIFTFACTAUTHORITYREQUIRED; projectionclaimingsourceauthority to DRIFTSOURCEAUTHORITYFORBIDDEN; liverepairactionauthorized to DRIFTLIVEREPAIRFORBIDDEN; privateruntimedataexport to DRIFTPRIVATERUNTIMEEXPORTFORBIDDEN; providerpayloadexport to DRIFTPROVIDERPAYLOADFORBIDDEN; automaticdoctrine_promotion to DRIFTAUTOMATICDOCTRINEPROMOTIONFORBID\u2026",
"level": 4,
"order": 15,
"section_id": "the-nine-cases-and-added-source-drift-cases",
"source_line_end": 180,
"source_line_start": 161,
"title": "The nine cases and added source-drift cases"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 182,
"source_line_start": 181,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It borrows the model-monitoring and concept-drift idea of a growing predicted-versus-observed gap as an operational signal, yielding fixture-bound monitoring evidence, source-linked only.** This component watches a projection for drift between what the model expects and what the runtime reports. It draws on the model-monitoring and concept-drift literature, which treats a growing gap between predicted and observed behaviour as an operational signal. The borrowed shape is drift-as-signal over metadata-only records. The result is fixture-bound monitoring evidence, source-linked only or whole-system correctness.",
"level": 4,
"order": 17,
"section_id": "concept-drift-monitoring-as-signal",
"source_line_end": 194,
"source_line_start": 183,
"title": "Concept-drift monitoring as signal"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Focused runtime validation: Paper-module corpus validation: Paper-module index validation from the repo root:",
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 216,
"source_line_start": 195,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means a fixture run produced the described records and rejected every named negative case.** A pass means the component ran over the public fixture, produced the records described here, and rejected every named negative case.",
"level": 4,
"order": 19,
"section_id": "what-a-pass-means-here",
"source_line_end": 222,
"source_line_start": 217,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 224,
"source_line_start": 223,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Over public synthetic inputs the named drift rows are guarded by references, per-row fact-authority, copied source-body evidence, and negative-case rejection, all reproducing under the validation commands.** The strongest claim this evidence supports is narrow: over public synthetic inputs, the named drift rows are guarded by source references, target references, repair routes, validation references, a fact-authority record per row, copied source-body evidence, metadata-only records, and negative-case rejection, and all of this reproduces under the validation commands above.",
"level": 4,
"order": 21,
"section_id": "the-strongest-supported-claim",
"source_line_end": 234,
"source_line_start": 225,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The proof boundary is the metadata-only record, copied manifest, fixture rows, and checks on this page, and the ceiling is public metadata-only evidence only, with no source authority, mutation, or launch.** It refuses everything past that. The proof boundary is the metadata-only runtime record, the copied source-module manifest, the fixture rows, the source-state and geometry checks, and the validation result records named on this page. The scope limit is public metadata-only evidence only: no private runtime body inspection, no source authority, no source-file changes, no live route repair, no automatic doctrine changes, no model-output data export, no launch-scope decision, no public s\u2026",
"level": 4,
"order": 22,
"section_id": "the-proof-boundary-and-authority-ceiling",
"source_line_end": 246,
"source_line_start": 235,
"title": "The proof boundary and scope limit"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/world_model_projection_drift_control_room.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The world-model projection drift control room validates public metadata-only projection-drift rows and copied source-module bodies without treating projections as source authority or repair authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "world_model_projection_drift_control_room.py",
"path": "src/microcosm_core/organs/world_model_projection_drift_control_room.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for projection-drift fixture validation, exported bundle validation, source-module digest checks, negative-case enforcement, metadata-only result records, and scope limits.",
"symbols": [
"run",
"run_drift_control_bundle",
"_source_module_manifest_result",
"_source_open_body_import_summary",
"_build_result",
"_write_receipts",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/world_model_projection_drift_control_room.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/world_model_projection_drift_control_room.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/world_model_projection_drift_control_room.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "World-Model Projection Drift Control Room is the public projection-drift boundary for Microcosm. It validates eight drift rows, source refs, repair routes, validation refs, target refs, source-module digest evidence, copied world-model and view-quality source bodies, secret-exclusion policy, and eight negative cases while keeping private runtime bodies, model-output data, live repair, source-file changes, automatic doctrine changes, launch, public sharing, and source-authority claims out of scope.",
"summary_line": "The world-model projection drift control room validates public metadata-only projection-drift rows and copied source-module bodies without treating projections as source authority or repair authority.",
"title": "World-Model Projection Drift Control Room"
},
{
"anchor": "paper-module-public-reveal-walkthrough",
"authority_ceiling": "Public reveal fixture and exported-bundle result records only; no launch, hosted deployment, public sharing, recipient work, external model access, secret export, whole-system equivalence, Lean/Lake execution, source-file changes, or whole-system correctness.",
"body_html": "
Executable bindings for the ten-minute tour
\n
public_reveal_walkthrough makes the first-run ten-minute tour executable by binding each step to evidence in the public clone, blocking if any step points at nothing the system runs.public_reveal_walkthrough checks the short guided tour Microcosm shows a first-time reader, and it binds that tour to evidence that exists in the public clone. The reader is promised a ten-minute path: run a command, look at local state, follow one route through patterns, work, events, and evidence, then read the result record and the scope limit behind it. This component makes that promise executable. It blocks if any step on the path points at nothing the system actually runs.
\n
It reads two inputs, a first-wave fixture carrying the walkthrough, evidence map, claim floor, and four rejection cases, and an exported bundle of copied public source bodies, writing metadata-only records from both. It reads two inputs. The first-wave fixture under fixtures/first_wave/public_reveal_walkthrough/input/ carries the declared walkthrough, an evidence map, a claim floor, and four rejection cases. The exported bundle under examples/public_reveal_walkthrough/exported_public_reveal_bundle/ carries copied public source bodies with a manifest. The component runs both and writes metadata-only result records: public_reveal_walkthrough_result.json, ten_minute_reveal_board.json, public_reveal_validation_receipt.json, and public_reveal_walkthrough_fixture_acceptance.json.
\n
A pass is bounded validator evidence rather than a launch, with the scope limit stated once under Scope limit. A pass is bounded validator evidence, not a launch. The scope limit is stated once, under Scope limit.
\n
Purpose
\n
Accounting, not a pitch
\n
A walkthrough drifts into a pitch as soon as commands and evidence refs are dropped, so this component treats the first-run path as accounting where each step points at real state, work, and records. A walkthrough drifts toward a pitch the moment it is allowed to. Dropping the commands and the evidence refs is the easiest way to make a first run look more impressive and prove less. This component removes that option. It treats the first-run path as an accounting exercise rather than a narrative: a command produces local state, a route points at work and events, an evidence ref points at a record, and a scope limit sits beside any impressive-looking total.
\n
The real-lane witness
\n
The fixture does not pass on its own paperwork; it passes only when the exported bundle runs underneath it with copied bodies digest-verified, so the reveal cannot describe a runnable path that is broken. The stricter part is the real-lane witness. The fixture does not pass on its own paperwork. It passes only when the exported bundle actually runs underneath it, with its copied source bodies present and digest-verified. So the reveal cannot describe a runnable path while the runnable path is broken. That is the quiet failure mode of every quick-start that says more than it can execute.
\n
How it works
\n
The two entrypoints share one core builder, _build_result, and differ in what they require. The two entrypoints share one core builder, _build_result, and differ in what they require.
\n
The fixture lane and its three validators
\n
run calls _build_result, which scans inputs and runs three validators for the density floor, the evidence map, and the claim floor, passing only when all pass with no missing rejection and a clean scan.run is the fixture lane. It calls _build_result with include_negative set true, then gates the outcome on a witness, then writes the record set. _build_result loads the fixture payloads, runs the secret-exclusion scan over every input path with scan_paths, and calls three validators. validate_walkthrough enforces the density floor: at least five steps, at least four distinct commands, and at least four evidence refs, otherwise it emits PUBLIC_REVEAL_DENSITY_FLOOR_MISSING. validate_evidence_map checks the evidence map. validate_claim_floor requires the declared product-loop phrases to appear in the public claim and rejects the overclaim cases. The builder sets status to pass only when no expected rejection case is missing, the secret scan has zero blocking hits, and all three validators pass.
\n
Gating the fixture on the real-lane witness
\n
_fixture_real_lane_witness reruns the exported bundle with the manifest required, and _attach_real_lane_witness forces the fixture result to blocked if that witness did not pass. After the builder, run calls _fixture_real_lane_witness, which runs the exported bundle through _build_result with the source-module manifest required, so the copied bodies are digest-checked for real. _attach_real_lane_witness reads that witness: if it did not pass, the fixture result is forced to blocked, real_runtime_receipt is set false, and a PUBLIC_REVEAL_REAL_LANE_WITNESS_REQUIRED or ...BLOCKED finding is added. Then write_receipts emits the records.
\n
The bundle lane digest-verifies each module
\n
run_reveal_bundle makes the source-module manifest mandatory, and _source_module_manifest_result checks each module's class, copied-and-metadata-only flag, existence, three digest fields, and required anchors, blocking on any mismatch.run_reveal_bundle is the exported-bundle lane. It calls _build_result with include_negative false, which makes the source-module manifest mandatory. _source_module_manifest_result walks each declared module: the material class must be one of PUBLIC_SAFE_SOURCE_BODY_CLASSES, the row must mark the body copied and kept out of the record, the target file must exist, all three digest fields must match the recomputed _sha256, and every declared anchor must be present in the copied text. Any mismatch becomes a finding and the manifest status goes to blocked. _source_open_body_import_summary folds the verified modules into a count and class list for the record.
\n
Function
Role
run
Fixture entrypoint: builds the result, attaches the real-lane witness, writes the record set.
run_reveal_bundle
Exported-bundle entrypoint: requires and digest-verifies the copied source bodies, writes the bundle record.
_build_result
Loads inputs, runs the secret scan, calls the three validators plus the source-body check, sets pass or blocked.
validate_walkthrough
Enforces the five-step, four-command, four-evidence-ref density floor and drives the missing-evidence rejection.
validate_claim_floor
Requires the product-loop phrases in the public claim and rejects launch, private-equivalence, and marketing overclaims.
_source_module_manifest_result
Digest-verifies each copied source body against the manifest and keeps record fields metadata-only.
_fixture_real_lane_witness
Reruns the exported bundle so the fixture pass is gated on the real path, not on its own summary.
\nDiagram of the mechanism (6 steps).Diagram source & refs
The fixture ships four inputs that must stay rejected, and if any stops producing its finding the reveal is no longer bounded enough for a cold reader. The fixture ships four inputs that must stay rejected. If any of these stops producing its finding, the reveal is no longer bounded enough for a cold reader.
\n
Rejected input
Finding code
What it catches
release_or_hosting_overclaim
PUBLIC_REVEAL_RELEASE_OVERCLAIM
A step claiming launch, hosting, public sharing, or recipient-work authority.
private_equivalence_overclaim
PUBLIC_REVEAL_PRIVATE_EQUIVALENCE_OVERCLAIM
A claim of non-public data equivalence or whole-system correctness.
missing_evidence_route
PUBLIC_REVEAL_STEP_EVIDENCE_MISSING
A step with no evidence ref or no inspect ref behind it.
marketing_without_runtime
PUBLIC_REVEAL_RUNTIME_COMMAND_MISSING
Marketing copy with no command behind it.
\n
A missing case blocks the run
\n
The four expected cases live in EXPECTED_NEGATIVE_CASES, and a run is blocked if any is absent from the observed set. The four expected cases are declared in EXPECTED_NEGATIVE_CASES; a run is blocked if any of them is absent from the observed set.
\n
Prior Art Grounding
\n
First-run CLI, progressive disclosure, provenance
\n
The path follows first-run CLI and progressive-disclosure practice for one runnable command and a bounded first route, and borrows provenance and tracing work for walking from artifact to source, record, and linked work. The path follows first-run CLI and progressive-disclosure practice. The Command Line Interface Guidelines motivate a single runnable command, worked examples, discoverable next steps, and machine-readable output. Nielsen Norman Group's progressive disclosure motivates showing the bounded first route before expanding into full drilldowns. The evidence walk borrows from provenance and tracing work: W3C PROV for moving from artifact to source and record, and OpenTelemetry traces for representing a causal chain as inspectable linked work.
The focused suite covers the fixture path, exported bundle, digest checks, and four negative cases, and a pass means the density floor was met, bodies verified, overclaims rejected, and records metadata-only for this fixture and bundle only. The first command runs the focused suite over the fixture path, the exported bundle, digest checks, and the four negative cases. The second confirms the paper-module corpus still resolves. A pass means the walkthrough met the density floor, the copied bodies verified, the overclaims stayed rejected, and the records stayed metadata-only for this fixture and this bundle only.
\n
Scope boundary
\n
Scope limit
\n
The narrow claim supported
\n
The strongest claim is that a cold reader can run a local walkthrough from command to route to result record with every step backed by something the system runs, over the fixture, bundle, manifest, tests, and records. The strongest claim the evidence supports is narrow: a cold reader can run a local walkthrough from a command to a route explanation to a result record and its source boundary, and every step on that path is backed by something the system runs. The proof boundary is the fixture, the exported bundle, the source-module manifest, the focused tests, and the metadata-only records named above.
\n
What it does not claim
\n
The component claims no launch, hosting, public sharing, recipient work, external model service, or whole-system correctness, keeping authority with the source, manifest, validators, and records rather than any board or count layered over them. It does not claim launch-scope decision, hosted deployment, publishing-scope decision, recipient work, external model service, secret export, whole-system equivalence, Lean or Lake execution, source-file changes, or whole-system correctness. The scope limit stays with the source, the manifest, the validators, and the records, not with any count, board, or generated card layered over them. Visual boards, screenshots, and copied-body counts are presentation over the record path, not evidence of maturity.
\n",
"components": [
{
"anchor": "component-public_reveal_walkthrough",
"component_id": "public_reveal_walkthrough",
"public_label": "Public Reveal Walkthrough"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 2,
"organ": 1,
"paper_module": 1,
"principle": 9
},
"link_count": 20,
"links": [
{
"id": "public_reveal_walkthrough",
"kind": "organ",
"kind_label": "component",
"label": "Public Reveal Walkthrough",
"relation": "explains",
"url": "component-public-reveal-walkthrough.html"
},
{
"id": "mechanism.public_reveal_walkthrough.validates_public_reveal_walkthrough",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public reveal walkthrough",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-public-reveal-walkthrough-validates-public-reveal-walkthrough"
},
{
"id": "mechanism.first_screen_composition_root.validates_public_first_screen_composition_root",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public first screen composition root",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-first-screen-composition-root-validates-public-first-screen-composition-root"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.entry_and_reveal_route_readiness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "entry and reveal routes as bounded cold-reader readiness evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-entry-and-reveal-route-readiness-bundle"
},
{
"id": "paper_module.first_screen_composition_root",
"kind": "paper_module",
"kind_label": "paper module",
"label": "First-Screen Composition Root",
"relation": "depends on",
"url": "paper-module-first-screen-composition-root.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch12_release_claim_language_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 12 launch claim-Language Gate",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[66:paper_module.batch12_release_claim_language_gate].depends_on",
"url": "paper-module-batch12-release-claim-language-gate.html"
},
{
"id": "paper_module.batch8_audio_level_rms_port",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Audio Level RMS Port",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[58:paper_module.batch8_audio_level_rms_port].depends_on",
"url": "paper-module-batch8-audio-level-rms-port.html"
},
{
"id": "paper_module.batch8_validator_checker_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Validator Checker Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[63:paper_module.batch8_validator_checker_capsule].depends_on",
"url": "paper-module-batch8-validator-checker-capsule.html"
},
{
"id": "paper_module.mechanistic_interpretability_circuit_attribution_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mechanistic Interpretability Circuit Attribution Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[51:paper_module.mechanistic_interpretability_circuit_attribution_replay].depends_on",
"url": "paper-module-mechanistic-interpretability-circuit-attribution-replay.html"
}
],
"paper_module_id": "paper_module.public_reveal_walkthrough",
"reader_packet": {
"abstract": "The public reveal walkthrough validates a ten-minute cold-reader path through commands, routes, evidence refs, source-open body imports, negative cases, and scope limits without claiming launch-scope decision or whole-system equivalence.",
"authority_ceiling": "Public reveal fixture and exported-bundle result records only; no launch, hosted deployment, public sharing, recipient work, external model access, secret export, whole-system equivalence, Lean/Lake execution, source-file changes, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.public_reveal_walkthrough.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-public-reveal-walkthrough-diagram-1",
"html_url": "docs/paper-module-public-reveal-walkthrough.html#paper-module-public-reveal-walkthrough-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Fixture[\"First-wave fixture walkthrough + claim floor + 4 negatives\"] Bundle[\"Exported bundle 5 copied source bodies + manifest\"] Build[\"_build_result load, scan, validate\"] Witness[\"_fixture_real_lane_witness rerun bundle, digest-check bodies\"] Records[\"metadata-only records result, board, validation, fixture\"] Ceiling[\"Scope limit no launch, hosting, or private equivalence\"] Fixture --> Build Bundle --> Witness Build --> Witness Witness --> Records Records --> Ceiling",
"source_format": "mermaid",
"source_hash": "sha256:85ecf9264ca5f2d6252038f9a56a16ee06b132de9cfde39db8cbcda254c2d348",
"source_line_end": 140,
"source_line_start": 126
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.first_screen_composition_root",
"kind": "paper_module",
"kind_label": "paper module",
"label": "First-Screen Composition Root",
"relation": "depends on",
"url": "paper-module-first-screen-composition-root.html"
}
],
"relation": "depends on"
},
{
"count": 3,
"links": [
{
"id": "public_reveal_walkthrough",
"kind": "organ",
"kind_label": "component",
"label": "Public Reveal Walkthrough",
"relation": "explains",
"url": "component-public-reveal-walkthrough.html"
},
{
"id": "mechanism.public_reveal_walkthrough.validates_public_reveal_walkthrough",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public reveal walkthrough",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-public-reveal-walkthrough-validates-public-reveal-walkthrough"
},
{
"id": "mechanism.first_screen_composition_root.validates_public_first_screen_composition_root",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public first screen composition root",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-first-screen-composition-root-validates-public-first-screen-composition-root"
}
],
"relation": "explains"
},
{
"count": 10,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.entry_and_reveal_route_readiness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "entry and reveal routes as bounded cold-reader readiness evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-entry-and-reveal-route-readiness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.public_reveal_walkthrough",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 238,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/public_reveal_walkthrough.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Command Line Interface Guidelines",
"section_id": "first-run-cli-progressive-disclosure-provenance",
"url": "https://clig.dev/"
},
{
"kind": "external_reference",
"label": "progressive disclosure",
"section_id": "first-run-cli-progressive-disclosure-provenance",
"url": "https://www.nngroup.com/articles/progressive-disclosure/"
},
{
"kind": "external_reference",
"label": "W3C PROV",
"section_id": "first-run-cli-progressive-disclosure-provenance",
"url": "https://www.w3.org/TR/prov-overview/"
},
{
"kind": "external_reference",
"label": "OpenTelemetry traces",
"section_id": "first-run-cli-progressive-disclosure-provenance",
"url": "https://opentelemetry.io/docs/concepts/signals/traces/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "publicreveal_walkthrough makes the first-run ten-minute tour executable by binding each step to evidence in the public clone, blocking if any step points at nothing the system runs.** publicrevealwalkthrough checks the short guided tour Microcosm shows a first-time reader, and it binds that tour to evidence that exists in the public clone. The reader is promised a ten-minute path: run a command, look at local state, follow one route through patterns, work, events, and evidence, then read the result record and the scope limit behind it. This component makes that promise executable. It blocks if any step on the path points at nothing the system actually runs.",
"level": 4,
"order": 1,
"section_id": "executable-bindings-for-the-ten-minute-tour",
"source_line_end": 15,
"source_line_start": 3,
"title": "Executable bindings for the ten-minute tour"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It reads two inputs, a first-wave fixture carrying the walkthrough, evidence map, claim floor, and four rejection cases, and an exported bundle of copied public source bodies, writing metadata-only records from both.** It reads two inputs. The first-wave fixture under fixtures/firstwave/publicreveal_walkthrough/input/ carries the declared walkthrough, an evidence map, a claim floor, and four rejection cases. The exported bundle under examples/publicrevealwalkthrough/exportedpublicreveal_bundle/ carries copied public source bodies with a manifest. The component runs both and writes metadata-only result records: publicrevealwalkthrough_result.json, tenminuterevealboard.json, publicrevealval\u2026",
"level": 4,
"order": 2,
"section_id": "it-reads-two-inputs",
"source_line_end": 30,
"source_line_start": 16,
"title": "It reads two inputs"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass is bounded validator evidence rather than a launch, with the scope limit stated once under Scope limit.** A pass is bounded validator evidence, not a launch. The scope limit is stated once, under Scope limit.",
"level": 4,
"order": 3,
"section_id": "a-pass-is-bounded-validator-evidence-rather-than-a-release",
"source_line_end": 37,
"source_line_start": 31,
"title": "A pass is bounded validator evidence rather than a launch"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 39,
"source_line_start": 38,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A walkthrough drifts into a pitch as soon as commands and evidence refs are dropped, so this component treats the first-run path as accounting where each step points at real state, work, and records.** A walkthrough drifts toward a pitch the moment it is allowed to. Dropping the commands and the evidence refs is the easiest way to make a first run look more impressive and prove less. This component removes that option. It treats the first-run path as an accounting exercise rather than a narrative: a command produces local state, a route points at work and events, an evidence ref points at a record, and a scope limit sits beside any impressive-looking total.",
"level": 4,
"order": 5,
"section_id": "accounting-not-a-pitch",
"source_line_end": 51,
"source_line_start": 40,
"title": "Accounting, not a pitch"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The fixture does not pass on its own paperwork; it passes only when the exported bundle runs underneath it with copied bodies digest-verified, so the reveal cannot describe a runnable path that is broken.** The stricter part is the real-lane witness. The fixture does not pass on its own paperwork. It passes only when the exported bundle actually runs underneath it, with its copied source bodies present and digest-verified. So the reveal cannot describe a runnable path while the runnable path is broken. That is the quiet failure mode of every quick-start that says more than it can execute.",
"level": 4,
"order": 6,
"section_id": "the-real-lane-witness",
"source_line_end": 62,
"source_line_start": 52,
"title": "The real-lane witness"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 64,
"source_line_start": 63,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The two entrypoints share one core builder, buildresult, and differ in what they require. The two entrypoints share one core builder, buildresult, and differ in what they require.",
"level": 4,
"order": 8,
"section_id": "the-two-entrypoints-share-one-core-builder",
"source_line_end": 70,
"source_line_start": 65,
"title": "The two entrypoints share one core builder"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run calls build_result, which scans inputs and runs three validators for the density floor, the evidence map, and the claim floor, passing only when all pass with no missing rejection and a clean scan.** run is the fixture lane. It calls buildresult with include_negative set true, then gates the outcome on a witness, then writes the record set. buildresult loads the fixture payloads, runs the secret-exclusion scan over every input path with scanpaths, and calls three validators. validatewalkthrough enforces the density floor: at least five steps, at least four distinct commands, and at least four evidence refs, otherwise it emits PUBLICREVEALDENSITYFLOORMISSING. validateevidencemap checks t\u2026",
"level": 4,
"order": 9,
"section_id": "the-fixture-lane-and-its-three-validators",
"source_line_end": 87,
"source_line_start": 71,
"title": "The fixture lane and its three validators"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "fixturereallanewitness reruns the exported bundle with the manifest required, and attachreallanewitness forces the fixture result to blocked if that witness did not pass.** After the builder, run calls fixturereallanewitness, which runs the exported bundle through buildresult with the source-module manifest required, so the copied bodies are digest-checked for real. attachreallanewitness reads that witness: if it did not pass, the fixture result is forced to blocked, realruntimereceipt is set false, and a PUBLICREVEALREALLANEWITNESS_REQUIRED or ...BLOCKED finding is added. Then write_receipts emits the records.",
"level": 4,
"order": 10,
"section_id": "gating-the-fixture-on-the-real-lane-witness",
"source_line_end": 99,
"source_line_start": 88,
"title": "Gating the fixture on the real-lane witness"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "runreveal_bundle makes the source-module manifest mandatory, and sourcemodulemanifestresult checks each module's class, copied-and-metadata-only flag, existence, three digest fields, and required anchors, blocking on any mismatch.** runrevealbundle is the exported-bundle lane. It calls buildresult with include_negative false, which makes the source-module manifest mandatory. sourcemodulemanifestresult walks each declared module: the material class must be one of PUBLICSAFESOURCEBODYCLASSES, the row must mark the body copied and kept out of the record, the target file must exist, all three digest fields must match the recomputed _sha256, and every declared anchor must be present in the copie\u2026",
"level": 4,
"order": 11,
"section_id": "the-bundle-lane-digest-verifies-each-module",
"source_line_end": 141,
"source_line_start": 100,
"title": "The bundle lane digest-verifies each module"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 143,
"source_line_start": 142,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "**The fixture ships four inputs that must stay rejected, and if any stops producing its finding the reveal is no longer bounded enough for a cold reader.** The fixture ships four inputs that must stay rejected. If any of these stops producing its finding, the reveal is no longer bounded enough for a cold reader.",
"level": 4,
"order": 13,
"section_id": "four-inputs-that-must-stay-rejected",
"source_line_end": 158,
"source_line_start": 144,
"title": "Four inputs that must stay rejected"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The four expected cases live in EXPECTEDNEGATIVE_CASES, and a run is blocked if any is absent from the observed set.** The four expected cases are declared in EXPECTEDNEGATIVECASES; a run is blocked if any of them is absent from the observed set.",
"level": 4,
"order": 14,
"section_id": "a-missing-case-blocks-the-run",
"source_line_end": 165,
"source_line_start": 159,
"title": "A missing case blocks the run"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 167,
"source_line_start": 166,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The path follows first-run CLI and progressive-disclosure practice for one runnable command and a bounded first route, and borrows provenance and tracing work for walking from artifact to source, record, and linked work.** The path follows first-run CLI and progressive-disclosure practice. The Command Line Interface Guidelines motivate a single runnable command, worked examples, discoverable next steps, and machine-readable output. Nielsen Norman Group's progressive disclosure motivates showing the bounded first route before expanding into full drilldowns. The evidence walk borrows from provenance and tracing work: W3C PROV for moving from artifact to source and record, and OpenTelemetry\u2026",
"level": 4,
"order": 16,
"section_id": "first-run-cli-progressive-disclosure-provenance",
"source_line_end": 184,
"source_line_start": 168,
"title": "First-run CLI, progressive disclosure, provenance"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 186,
"source_line_start": 185,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "record or generated projection surfaces:",
"level": 4,
"order": 18,
"section_id": "reader-validation-without-durable-writes",
"source_line_end": 197,
"source_line_start": 187,
"title": "Reader validation without durable writes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The focused suite covers the fixture path, exported bundle, digest checks, and four negative cases, and a pass means the density floor was met, bodies verified, overclaims rejected, and records metadata-only for this fixture and bundle only.** The first command runs the focused suite over the fixture path, the exported bundle, digest checks, and the four negative cases. The second confirms the paper-module corpus still resolves. A pass means the walkthrough met the density floor, the copied bodies verified, the overclaims stayed rejected, and the records stayed metadata-only for this fixture and this bundle only.",
"level": 4,
"order": 19,
"section_id": "what-the-two-commands-check",
"source_line_end": 209,
"source_line_start": 198,
"title": "What the two commands check"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 211,
"source_line_start": 210,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is that a cold reader can run a local walkthrough from command to route to result record with every step backed by something the system runs, over the fixture, bundle, manifest, tests, and records.** The strongest claim the evidence supports is narrow: a cold reader can run a local walkthrough from a command to a route explanation to a result record and its source boundary, and every step on that path is backed by something the system runs. The proof boundary is the fixture, the exported bundle, the source-module manifest, the focused tests, and the metadata-only records named above.",
"level": 4,
"order": 21,
"section_id": "the-narrow-claim-supported",
"source_line_end": 223,
"source_line_start": 212,
"title": "The narrow claim supported"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component claims no launch, hosting, public sharing, recipient work, external model service, or whole-system correctness, keeping authority with the source, manifest, validators, and records rather than any board or count layered over them.** It does not claim launch-scope decision, hosted deployment, public sharing decision, recipient work, external model service, secret export, whole-system equivalence, Lean or Lake execution, source-file changes, or whole-system correctness. The scope limit stays with the source, the manifest, the validators, and the records, not with any count, board, or generated card layered over them. Visual boards, screenshots, and copied-body counts are prese\u2026",
"level": 4,
"order": 22,
"section_id": "what-it-does-not-claim",
"source_line_end": 237,
"source_line_start": 224,
"title": "What it does not claim"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/public_reveal_walkthrough.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The public reveal walkthrough validates a ten-minute cold-reader path through commands, routes, evidence refs, source-open body imports, negative cases, and scope limits without claiming launch-scope decision or whole-system equivalence."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "public_reveal_walkthrough.py",
"path": "src/microcosm_core/organs/public_reveal_walkthrough.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public reveal fixture validation, exported reveal bundle validation, source-module digest checks, negative-case enforcement, cards, and scope limits.",
"symbols": [
"run",
"run_reveal_bundle",
"_source_module_manifest_result",
"_source_open_body_import_summary",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING",
"PUBLIC_SAFE_SOURCE_BODY_CLASSES"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/public_reveal_walkthrough.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/public_reveal_walkthrough.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/public_reveal_walkthrough.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Public Reveal Walkthrough is the source-backed public-entry membrane for Microcosm. It checks the declared reveal steps, runnable command set, evidence refs, claim-floor phrases, secret-exclusion scan, source-open body import manifest, runtime-bundle shape, and four overclaim negative cases while keeping copied bodies out of result records and launch/provider/private-equivalence claims out of scope.",
"summary_line": "The public reveal walkthrough validates a ten-minute cold-reader path through commands, routes, evidence refs, source-open body imports, negative cases, and scope limits without claiming launch-scope decision or whole-system equivalence.",
"title": "Public Reveal Walkthrough"
},
{
"anchor": "paper-module-standards-meta-diagnostics",
"authority_ceiling": "Projection-only diagnostic; never source authority for core/standards_registry.json, no source-file changes, no provider/launch-scope decision, no whole-system correctness.",
"body_html": "
It confirms every accepted component still ties to a standard, a documented way to run it, and a saved result record, emitting a coded finding when a tie is missing.standards_meta_diagnostics is a coverage checker. It reads the project's public catalogs and confirms that every accepted component is still tied to a written standard, a documented way to run it, and a saved result record. When one of those ties is missing, it emits a structured finding with a named error code instead of a paragraph of prose.
\n
Silent drift is the failure: attachments fall out of step one component at a time with nothing breaking, so a gap sits unnoticed until a reader hits the hole. The problem it solves is silent drift. A spine of accepted components stays coherent only while each one keeps its four attachments: a standard that describes it, a runtime contract that runs it, a result record that holds its last verdict, and a stated limit on what it is allowed to claim. As the spine grows, those attachments fall out of step one component at a time, and nothing breaks when they do. A new component can be accepted while its standard file, registry row, or result-record ref is never added. The gap just sits there until a reader trusts the spine and finds a hole.
\n
Its output is a metadata-only record of refs, counts, hashes, and verdicts, readable in the open without exporting private source. The check emits a metadata-only result record: refs, counts, hashes, and verdicts, never the source bodies they describe. So a coverage report can be read in the open without exporting private source.
\n
Purpose
\n
Coverage projected from the live registry
\n
Positive coverage is projected from the live public registry rather than stored, so four public surfaces must agree on one component set or a mismatch is a finding. A loose check fails here in a specific way. If the diagnostic held its own checked-in list of what the spine should contain, that list would itself go stale, and a stale example would quietly become the thing the spine is measured against. So the positive coverage is projected from the live public registry rather than stored. The accepted-component list, the standard rows, the runtime-contract rows, and the result-record refs then have to agree on exactly the same set of components. Any component that appears in one surface but not another is a finding.
\n
Five negative fixtures keep it falsifiable
\n
Five negative fixtures must each surface their expected failure, and the run blocks if one that should fail passes. The check is also falsifiable. It carries five negative fixtures that each must surface their expected failure. If a fixture that should fail passes, the run is blocked, so the checker cannot silently stop catching a class of gap.
\n
How it works
\n
Two entry points drive the same validator over different inputs.
\n
run drives the fixture path
\n
run calls _build_result with live positive projection and all five negatives evaluated, then attaches a freshness basis and writes result records.run is the fixture path. It calls _build_result with include_negative=True and project_positive_from_live=True, so the positive rows come from live registry state and all five negative fixtures are evaluated. It then attaches a freshness basis and calls _write_receipts.
\n
run_diagnostics_bundle reads the exported bundle
\n
run_diagnostics_bundle reads the three exported JSON inputs directly and can return a cached result when the input digests are unchanged.run_diagnostics_bundle is the exported-bundle path. It calls _build_result with include_negative=False and project_positive_from_live=False, reading the three exported JSON inputs directly. With reuse_fresh_receipt it can return a cached result when _fresh_bundle_receipt confirms the input digests have not changed.
\n
_build_result is the core. In order, it:
\n
loads the payloads through _load_payloads, which either reads standards_inventory.json, organ_runtime_contracts.json, and diagnostic_policy.json, or projects those rows from the live registry via build_diagnostics_input_payloads;
validates any exported source-module manifest with _source_module_manifest_result, which digest-checks each copied body;
runs scan_paths over every input with the forbidden-class policy loaded by load_forbidden_classes;
computes _positive_findings over the accepted-component set and _negative_findings over the five negative fixtures;
assembles the status.
\n
When status is pass versus blocked
\n
Status is pass only with no positive findings, no missing negatives, no blocking secret hits, and a passing or absent manifest; anything else is blocked. The status is pass only when there are no positive findings, no missing negative cases, no blocking secret hits, and the source-module manifest is pass or absent. Anything else is blocked.
\n
_positive_findings checks each accepted component
\n
_positive_findings records a gap when any accepted component's inventory row, refs, standard identity, result records, or runtime step is missing, and it rejects overclaim flags._positive_findings does the coverage work. For each component id in the policy's accepted_organ_ids, it looks up the inventory row and the runtime row, and it derives the expected standard identity std_microcosm_<id> and the expected file standards/std_microcosm_<id>.json. It records a finding when the inventory row is absent, when standard_id, standard_ref, or registry_row_ref is empty, when the standard identity does not match the live file, when receipt_refs is empty, or when the runtime row lacks a cli_command or a runtime_step. It also rejects a policy that sets any overclaim flag such as release_authorized.
\n
Negative findings, result records, and the compact card
\n
_negative_findings records each fixture's triggered code, _write_receipts writes the result, board, and validation records, and result_card projects a card that omits the detailed lists._negative_findings walks the five negative fixtures and records the error code each one is built to trigger. _write_receipts writes the result, board, and validation records with write_json_atomic, plus an sign-off record when asked. result_card projects a compact card that omits the covered-component list, findings, and source refs from its payload while keeping them in the full result.
\nDiagram of the mechanism (10 steps).Diagram source & refs
flowchart TD inventory["standards_inventory"] contracts["runtime contract inventory"] policy["diagnostic_policy"] negatives["five negative fixtures"] build["_build_result"] positive["_positive_findings: coverage per component"] negative["_negative_findings: expected failures"] scan["scan_paths: secret exclusion"] status["status: pass or blocked"] result records["_write_receipts: metadata-only records"] inventory --> build contracts --> build policy --> build negatives --> build build --> positive build --> negative build --> scan positive --> status negative --> status scan --> status status --> result records
\n
Negative cases
\n
Five negative fixtures each map to one expected error code, and a run that fails to surface any of them is blocked: Five negative fixtures each map to one expected error code, and a run that fails to surface any of them is blocked:
On the current fixture the sign-off record reports status: pass with 88 accepted components fully mapped, all five error codes present, and zero blocking secret hits. On the current fixture the sign-off record reports status: pass with accepted_organ_count: 88, standard_mapping_count: 88, runtime_contract_count: 88, all five expected error codes present, and secret_exclusion_scan.blocking_hit_count: 0.
\n
Prior Art Grounding
\n
Grounded in schema and contract validation
\n
The design follows JSON Schema and OpenAPI validation practice, applying it to the launch boundary as bounded evidence about consistency rather than a claim of authority. This component is grounded in schema and contract validation practice rather than in a claim that a diagnostic creates authority. JSON Schema treats a schema as a machine-readable vocabulary for validating structured JSON data, and OpenAPI uses interface descriptions so a consumer can understand an API without reading its source. The check applies that pattern to the launch boundary: standards, adapter contracts, result records, and scope limits are validated as public projections, and the check stays bounded evidence about consistency.
\n
JSON Schema validation and structured-data constraints: https://json-schema.org/
OpenAPI interface descriptions and conformance expectations: https://spec.openapis.org/oas/latest.html
A pass means positive inputs cover the accepted set, the five negatives surface their named codes, and the paper-module corpus is in parity, as validator evidence about consistency. A pass means the positive inputs cover the accepted-component set, the five negative cases surface their named error codes, and the paper-module corpus is in parity. It is validator evidence about consistency.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim and its ceiling
\n
The evidence supports only that public inputs project consistently into one inspectable metadata-only record; the check mutates nothing, exports no source, and excludes launch. The strongest claim the evidence supports is that the public standards inventory, runtime contracts, accepted-component refs, result-record refs, diagnostic policy, and secret-exclusion scan are projected consistently into one metadata-only result record a reader can inspect. The proof boundary is those public inputs and the sign-off and validation records built from them. The scope limit stops there: the check does not mutate the standards registry, become source authority, export private source, authorize providers, include launch operations or public sharing, or prove whole-system correctness.
\n",
"components": [
{
"anchor": "component-standards_meta_diagnostics",
"component_id": "standards_meta_diagnostics",
"public_label": "Standards Meta Diagnostics"
}
],
"doctrine_feed": {
"counts": {
"axiom": 3,
"concept": 2,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 3
},
"link_count": 11,
"links": [
{
"id": "standards_meta_diagnostics",
"kind": "organ",
"kind_label": "component",
"label": "Standards Meta Diagnostics",
"relation": "explains",
"url": "component-standards-meta-diagnostics.html"
},
{
"id": "mechanism.standards_meta_diagnostics.validates_public_standards_meta_diagnostics",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public standards meta diagnostics",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-standards-meta-diagnostics-validates-public-standards-meta-diagnostics"
},
{
"id": "P-7",
"kind": "principle",
"kind_label": "principle",
"label": "Track known unknowns without claiming the unknown is mapped",
"relation": "governed by",
"url": "doctrine.html#dcard-p-7"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
},
{
"id": "concept.standards_meta_diagnostics_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "component-standard-runtime-result record coverage as projection-only evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-standards-meta-diagnostics-bundle"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.microcosm_axiom_substrate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Microcosm Axiom System",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[92:paper_module.microcosm_axiom_substrate].depends_on",
"url": "paper-module-microcosm-axiom-substrate.html"
}
],
"paper_module_id": "paper_module.standards_meta_diagnostics",
"reader_packet": {
"abstract": "Terminal public coverage diagnostic: verifies every accepted component stays mapped to a standard, runtime contract, result record, and scope limit.",
"authority_ceiling": "Projection-only diagnostic; never source authority for core/standards_registry.json, no source-file changes, no provider/launch-scope decision, no whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.standards_meta_diagnostics.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-standards-meta-diagnostics-diagram-1",
"html_url": "docs/paper-module-standards-meta-diagnostics.html#paper-module-standards-meta-diagnostics-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD inventory[\"standards_inventory\"] contracts[\"runtime contract inventory\"] policy[\"diagnostic_policy\"] negatives[\"five negative fixtures\"] build[\"_build_result\"] positive[\"_positive_findings: coverage per component\"] negative[\"_negative_findings: expected failures\"] scan[\"scan_paths: secret exclusion\"] status[\"status: pass or blocked\"] result records[\"_write_receipts: metadata-only records\"] inventory --> build contracts --> build policy --> build negatives --> build build --> positive build --> negative build --> scan positive --> status negative --> status scan --> status status --> result records",
"source_format": "mermaid",
"source_hash": "sha256:e87ef37414e7362014a2ed852b163e107f3435ed63f8eadf377760406fc74a53",
"source_line_end": 153,
"source_line_start": 129
}
],
"grouped_relation_feeds": [
{
"count": 3,
"links": [
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "standards_meta_diagnostics",
"kind": "organ",
"kind_label": "component",
"label": "Standards Meta Diagnostics",
"relation": "explains",
"url": "component-standards-meta-diagnostics.html"
},
{
"id": "mechanism.standards_meta_diagnostics.validates_public_standards_meta_diagnostics",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public standards meta diagnostics",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-standards-meta-diagnostics-validates-public-standards-meta-diagnostics"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-7",
"kind": "principle",
"kind_label": "principle",
"label": "Track known unknowns without claiming the unknown is mapped",
"relation": "governed by",
"url": "doctrine.html#dcard-p-7"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.standards_meta_diagnostics_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "component-standard-runtime-result record coverage as projection-only evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-standards-meta-diagnostics-bundle"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.standards_meta_diagnostics",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 235,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/standards_meta_diagnostics.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It confirms every accepted component still ties to a standard, a documented way to run it, and a saved result record, emitting a coded finding when a tie is missing. standardsmeta_diagnostics is a coverage checker. It reads the project's public catalogs and confirms that every accepted component is still tied to a written standard, a documented way to run it, and a saved result record. When one of those ties is missing, it emits a structured finding with a named error code instead of a paragraph of prose.",
"level": 4,
"order": 1,
"section_id": "it-confirms-every-accepted-component-still-ties-to-a-standard",
"source_line_end": 12,
"source_line_start": 3,
"title": "It confirms every accepted component still ties to a standard"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Silent drift is the failure: attachments fall out of step one component at a time with nothing breaking, so a gap sits unnoticed until a reader hits the hole.** The problem it solves is silent drift. A spine of accepted components stays coherent only while each one keeps its four attachments: a standard that describes it, a runtime contract that runs it, a result record that holds its last verdict, and a stated limit on what it is allowed to claim. As the spine grows, those attachments fall out of step one component at a time, and nothing breaks when they do. A new component can be accepted while its standard file, registry row, or result-record ref is never added. The gap just sits there\u2026",
"level": 4,
"order": 2,
"section_id": "silent-drift-is-the-failure",
"source_line_end": 25,
"source_line_start": 13,
"title": "Silent drift is the failure"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Its output is a metadata-only record of refs, counts, hashes, and verdicts, readable in the open without exporting private source.** The check emits a metadata-only result record: refs, counts, hashes, and verdicts, never the source bodies they describe. So a coverage report can be read in the open without exporting private source.",
"level": 4,
"order": 3,
"section_id": "its-output-is-a-body-free-record-of-refs",
"source_line_end": 33,
"source_line_start": 26,
"title": "Its output is a metadata-only record of refs"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 35,
"source_line_start": 34,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Positive coverage is projected from the live public registry rather than stored, so four public surfaces must agree on one component set or a mismatch is a finding.** A loose check fails here in a specific way. If the diagnostic held its own checked-in list of what the spine should contain, that list would itself go stale, and a stale example would quietly become the thing the spine is measured against. So the positive coverage is projected from the live public registry rather than stored. The accepted-component list, the standard rows, the runtime-contract rows, and the result-record refs then have to agree on exactly the same set of components. Any component that appears in one surface\u2026",
"level": 4,
"order": 5,
"section_id": "coverage-projected-from-the-live-registry",
"source_line_end": 48,
"source_line_start": 36,
"title": "Coverage projected from the live registry"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Five negative fixtures must each surface their expected failure, and the run blocks if one that should fail passes.** The check is also falsifiable. It carries five negative fixtures that each must surface their expected failure. If a fixture that should fail passes, the run is blocked, so the checker cannot silently stop catching a class of gap.",
"level": 4,
"order": 6,
"section_id": "five-negative-fixtures-keep-it-falsifiable",
"source_line_end": 56,
"source_line_start": 49,
"title": "Five negative fixtures keep it falsifiable"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Two entry points drive the same validator over different inputs.",
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 60,
"source_line_start": 57,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run calls build_result with live positive projection and all five negatives evaluated, then attaches a freshness basis and writes result records.** run is the fixture path. It calls buildresult with include_negative=True and projectpositivefrom_live=True, so the positive rows come from live registry state and all five negative fixtures are evaluated. It then attaches a freshness basis and calls writereceipts.",
"level": 4,
"order": 8,
"section_id": "run-drives-the-fixture-path",
"source_line_end": 69,
"source_line_start": 61,
"title": "`run` drives the fixture path"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "rundiagnostics_bundle reads the three exported JSON inputs directly and can return a cached result when the input digests are unchanged.** rundiagnosticsbundle is the exported-bundle path. It calls buildresult with includenegative=False and projectpositivefromlive=False, reading the three exported JSON inputs directly. With reusefreshreceipt it can return a cached result when freshbundle_receipt confirms the input digests have not changed. buildresult is the core. In order, it: loads the payloads through loadpayloads, which either reads standardsinventory.json, organruntime_contracts.json, and diagnostic_policy.json, or projects those rows from the live registry via builddiagnosticsinput_pa\u2026",
"level": 4,
"order": 9,
"section_id": "run-diagnostics-bundle-reads-the-exported-bundle",
"source_line_end": 93,
"source_line_start": 70,
"title": "`run_diagnostics_bundle` reads the exported bundle"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Status is pass only with no positive findings, no missing negatives, no blocking secret hits, and a passing or absent manifest; anything else is blocked.** The status is pass only when there are no positive findings, no missing negative cases, no blocking secret hits, and the source-module manifest is pass or absent. Anything else is blocked.",
"level": 4,
"order": 10,
"section_id": "when-status-is-pass-versus-blocked",
"source_line_end": 101,
"source_line_start": 94,
"title": "When status is pass versus blocked"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "positivefindings records a gap when any accepted component's inventory row, refs, standard identity, result records, or runtime step is missing, and it rejects overclaim flags. positive_findings does the coverage work. For each component id in the policy's acceptedorganids, it looks up the inventory row and the runtime row, and it derives the expected standard identity stdmicrocosm and the expected file standards/stdmicrocosm.json. It records a finding when the inventory row is absent, when standardid, standardref, or registryrowref is empty, when the standard identity does not match the live file, when receipt_refs is empty, or when the runtime row lacks a clicommand or a runtimest\u2026",
"level": 4,
"order": 11,
"section_id": "positive-findings-checks-each-accepted-component",
"source_line_end": 116,
"source_line_start": 102,
"title": "`_positive_findings` checks each accepted component"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "negativefindings records each fixture's triggered code, writereceipts writes the result, board, and validation records, and result_card projects a card that omits the detailed lists. negative_findings walks the five negative fixtures and records the error code each one is built to trigger. writereceipts writes the result, board, and validation records with writejsonatomic, plus an sign-off record when asked. result_card projects a compact card that omits the covered-component list, findings, and source refs from its payload while keeping them in the full result.",
"level": 4,
"order": 12,
"section_id": "negative-findings-receipts-and-the-compact-card",
"source_line_end": 154,
"source_line_start": 117,
"title": "Negative findings, result records, and the compact card"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 156,
"source_line_start": 155,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "Five negative fixtures each map to one expected error code, and a run that fails to surface any of them is blocked: Five negative fixtures each map to one expected error code, and a run that fails to surface any of them is blocked: missingstandardref yields STANDARDSMETAMISSINGSTANDARDREF unmappedacceptedorgan yields STANDARDSMETAUNMAPPEDACCEPTEDORGAN missingreceiptref yields STANDARDSMETAMISSINGRECEIPTREF releaseoverclaim yields STANDARDSMETAAUTHORITYOVERCLAIM privatesourceleakage yields STANDARDSMETAPRIVATESOURCEFORBIDDEN",
"level": 4,
"order": 14,
"section_id": "five-negative-fixtures-each-map-to-one-expected-error-code",
"source_line_end": 168,
"source_line_start": 157,
"title": "Five negative fixtures each map to one expected error code"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**On the current fixture the sign-off record reports status: pass with 88 accepted components fully mapped, all five error codes present, and zero blocking secret hits.** On the current fixture the sign-off record reports status: pass with acceptedorgancount: 88, standardmappingcount: 88, runtimecontractcount: 88, all five expected error codes present, and secretexclusionscan.blockinghitcount: 0.",
"level": 4,
"order": 15,
"section_id": "what-the-current-fixture-reports",
"source_line_end": 177,
"source_line_start": 169,
"title": "What the current fixture reports"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 179,
"source_line_start": 178,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**The design follows JSON Schema and OpenAPI validation practice, applying it to the launch boundary as bounded evidence about consistency rather than a claim of authority.** This component is grounded in schema and contract validation practice rather than in a claim that a diagnostic creates authority. JSON Schema treats a schema as a machine-readable vocabulary for validating structured JSON data, and OpenAPI uses interface descriptions so a consumer can understand an API without reading its source. The check applies that pattern to the launch boundary: standards, adapter contracts, result records, and scope limits are validated as public projections, and the check stays bounded evidence\u2026",
"level": 4,
"order": 17,
"section_id": "grounded-in-schema-and-contract-validation",
"source_line_end": 196,
"source_line_start": 180,
"title": "Grounded in schema and contract validation"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 198,
"source_line_start": 197,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "result-record or generated-projection surfaces:",
"level": 4,
"order": 19,
"section_id": "reader-validation-without-durable-writes",
"source_line_end": 211,
"source_line_start": 199,
"title": "Reader validation without durable writes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means positive inputs cover the accepted set, the five negatives surface their named codes, and the paper-module corpus is in parity, as validator evidence about consistency.** A pass means the positive inputs cover the accepted-component set, the five negative cases surface their named error codes, and the paper-module corpus is in parity. It is validator evidence about consistency.",
"level": 4,
"order": 20,
"section_id": "what-a-pass-certifies",
"source_line_end": 219,
"source_line_start": 212,
"title": "What a pass certifies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 221,
"source_line_start": 220,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The evidence supports only that public inputs project consistently into one inspectable metadata-only record; the check mutates nothing, exports no source, and excludes launch.** The strongest claim the evidence supports is that the public standards inventory, runtime contracts, accepted-component refs, result-record refs, diagnostic policy, and secret-exclusion scan are projected consistently into one metadata-only result record a reader can inspect. The proof boundary is those public inputs and the sign-off and validation records built from them. The scope limit stops there: the check does not mutate the standards registry, become source authority, export private source, authorize provi\u2026",
"level": 4,
"order": 22,
"section_id": "the-strongest-supported-claim-and-its-ceiling",
"source_line_end": 234,
"source_line_start": 222,
"title": "The strongest supported claim and its ceiling"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/standards_meta_diagnostics.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Terminal public coverage diagnostic: verifies every accepted component stays mapped to a standard, runtime contract, result record, and scope limit."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "standards_meta_diagnostics.py",
"path": "src/microcosm_core/organs/standards_meta_diagnostics.py",
"resolution": "resolved",
"role": "Runtime + result record-writing locus for the public coverage diagnostic.",
"symbols": [
"run",
"run_diagnostics_bundle",
"_build_result",
"_write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/standards_meta_diagnostics.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/standards_meta_diagnostics.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/standards_meta_diagnostics.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Checks accepted adapter-backed components against standards_inventory/organ_runtime_contracts/diagnostic_policy; rejects 5 boundary failures (missing standard_id/standard_ref, missing inventory row, missing result record ref, launch/provider/public sharing overclaim, private-body leak); secret_exclusion_scan with body_in_receipt:false and synthetic_receipt_standin_allowed:false.",
"summary_line": "Terminal public coverage diagnostic: verifies every accepted component stays mapped to a standard, runtime contract, result record, and scope limit.",
"title": "Standards Meta Diagnostics"
},
{
"anchor": "paper-module-finance-forecast-evaluation-spine",
"authority_ceiling": "Synthetic market-shaped fixtures only; not investment or trading decisions, no live market data, no track record or performance claim, mutates no optimizer, SciPy absence is a typed HLN refusal.",
"body_html": "
Component runs forecast-evaluation
\n
This component runs forecast-evaluation statistics over synthetic, market-shaped fixtures, records both the numbers computed and the cases refused, and writes a metadata-only result record from the copied econometric code.finance_forecast_evaluation_spine runs forecast-evaluation statistics over synthetic, market-shaped fixtures and records both the numbers it computed and the cases it refused. It is a public component with runnable source. It reads a synthetic loss matrix and paired loss series, checks the request is admissible, runs the copied econometric code, and writes a metadata-only result record.
\n
Value is evaluation
\n
The value is the evaluation posture, not a claim about any asset. The value is the evaluation posture, not a claim about any asset.
\n
Result record carries
\n
The result record carries computed statistics when the fixture supports them and typed refusals when it does not, and both are treated as valid outcomes. The result record carries computed statistics when the fixture supports them and typed refusals when it does not. Both are treated as valid outcomes.
\n
Purpose
\n
Why model comparison is hard
\n
A lower average loss does not establish better forecasting because losses are autocorrelated, samples are short, and a careless split lets a model peek past its forecast point. Comparing two forecasting models is harder than it looks. A lower average loss does not establish one model predicts better. Losses are autocorrelated, samples are short, and a careless time split lets a model peek past the point it was supposed to forecast from. A "loose" comparison, one that reports a mean and stops, produces a number that reads like a result but is not one.
\n
The econometric machinery it carries
\n
It carries the careful machinery economists use, Diebold-Mariano with a Bartlett HAC variance, the HLN correction, Hansen's SPA test, a model confidence set, and a stationary bootstrap, run on synthetic data only. This component carries the machinery economists use to answer the question carefully: the Diebold-Mariano loss-differential statistic with a Bartlett HAC long-run variance, the Harvey-Leybourne-Newbold small-sample correction, Hansen's test for superior predictive ability, a model confidence set, and a Politis-Romano stationary bootstrap. It runs them on synthetic data only, and it refuses when the input is inadmissible instead of returning a number that could be misread.
\n
How it works
\n
A thin runtime over SPEC
\n
The component is a thin runtime over a CrownJewelSpec that fixes the four required inputs, names the three negative cases, binds the source manifest, and restricts reads to anchors in four copied modules. The component is a thin runtime over a CrownJewelSpec named SPEC. The spec fixes the four required inputs (family_loss_matrix.json, paired_loss_series.json, finance_boundary_policy.json, projection_protocol.json), names the three negative cases, binds the source manifest, and restricts source-open reads to required anchors in four copied modules: model_selection_stats.py, spa_statistics.py, loss_differentials.py, and family_loss_matrix.py.
\n
Two entrypoints, one bundle difference
\n
run and run_finance_forecast_bundle both delegate to run_crown_jewel_organ, differing only in that the bundle entrypoint sets input_mode to the exported-bundle mode.run and run_finance_forecast_bundle both delegate to run_crown_jewel_organ with evaluate and evaluate_negative_case. The bundle entrypoint sets input_mode to the exported-bundle mode; that is the only difference.
\n
evaluate loads inputs, order matters
\n
evaluate loads the loss matrix, paired loss series, and boundary policy, then calls _evaluate_payloads, where the order of operations matters.evaluate loads the loss matrix, the paired loss series, and the boundary policy, then calls _evaluate_payloads. That function is where the order of operations matters.
\n
Guards run before any computation
\n
_policy_findings and _lookahead_split_findings run first, blocking on any non-False advice flag or forecast-date leakage, so an inadmissible request never reaches a statistic. The guards run first. _policy_findings walks FALSE_BOUNDARY_FLAGS (investment_advice_authorized, trading_advice_authorized, live_market_data_authorized, track_record_claim_authorized, performance_claim_authorized, optimizer_mutation_authorized) and raises FINANCE_NO_ADVICE_OVERCLAIM for any flag that is not exactly False. _lookahead_split_findings parses the row dates with _parse_iso_date and raises FINANCE_LOOKAHEAD_SPLIT_FORBIDDEN when a row's subject_as_of reaches its event_start, or when an event window ends before it begins. If either guard produces a finding, or the source manifest did not pass, _evaluate_payloads returns a blocked record and no statistics run. An inadmissible request never reaches a computation.
\n
Branching to standalone or subprocess stats
\n
Only after the guards pass does _is_exported_finance_bundle branch to a declared, unexecuted contract for the exported bundle or a real child-interpreter subprocess for the first-wave fixture with scipy blocked. Only after the guards pass does the branch happen. _is_exported_finance_bundle decides the path. For the exported public bundle, _standalone_exported_statistics_contract returns a declared contract: the statistics are named and marked executed: false, with not_a_live_run and real_runtime_receipt: false, so the standalone bundle does not depend on a live source subprocess. For the first-wave fixture, _run_stats_subprocess (memoized by _run_stats_subprocess_cached) shells out to a child interpreter that imports the copied tools.finance modules. It runs reality_check_summary, spa_summary, and model_confidence_set_summary with min_sample 10, bootstrap_reps 40, and seed 1729, paired_loss_summary with min_paired 10, and stationary_bootstrap_indices with reps 5 and seed 1729. The subprocess deliberately blocks scipy from importing before calling harvey_leybourne_newbold_correction, which forces a typed refusal instead of a fabricated t-distribution p-value.
\n
Checking the subprocess payload
\n
Back in the parent, _evaluate_payloads checks each statistic field against its expected status, blocks on any mismatch, and keeps source bodies out with body_in_receipt: false and hashed stdout and stderr. Back in the parent, _evaluate_payloads reads the subprocess payload and checks each field: Reality Check and SPA must be computed_bootstrap, the model confidence set must report implemented, Diebold-Mariano must be computed_hac_normal_approximation, and the HLN entry must be a refusal with reason scipy_unavailable_for_t_distribution. Any mismatch becomes a finding and the record is blocked. Source bodies never enter the record; every emitted stats object keeps body_in_receipt: false, and the subprocess is summarized by stdout_sha256 and stderr_sha256 from _sha256_text.
\n
Negative cases reuse the same path
\n
evaluate_negative_case mutates one input to reproduce a specific fault, re-runs _evaluate_payloads, and confirms the expected stable error code appears.evaluate_negative_case reuses the same path. It loads the fixture, mutates one input to reproduce a specific fault (moving a row's subject_as_of onto its event_start for the leakage case, or setting investment_advice_authorized true for the overclaim case), re-runs _evaluate_payloads, and confirms the expected stable error code appears.
\n
Function
Role
run, run_finance_forecast_bundle
Entrypoints; forward the spec and evaluators to run_crown_jewel_organ
evaluate
Loads the three fixture files, then calls _evaluate_payloads
_evaluate_payloads
Runs guards, blocks on failure, else branches to standalone or subprocess stats
_policy_findings
Rejects any advice/live-data/performance flag that is not False
_lookahead_split_findings
Rejects rows where the forecast date reaches the event window
_run_stats_subprocess
Runs the copied tools.finance code in a child interpreter with scipy blocked
_standalone_exported_statistics_contract
Declares, without recomputing, the statistics for the exported bundle
evaluate_negative_case
Mutates one input per case and confirms the expected error code
\nDiagram of the mechanism (8 steps).Diagram source & refs
flowchart TD Inputs["Synthetic fixture family_loss_matrix, paired_loss_series, finance_boundary_policy"] Evaluate["evaluate loads inputs"] Guards["_policy_findings and _lookahead_split_findings"] Blocked["Blocked record no statistics run"] Branch{"Exported bundle?"} Standalone["_standalone_exported_statistics_contract declared, not recomputed"] Subprocess["_run_stats_subprocess DM/HAC, SPA, MCS, bootstrap, HLN refusal"] Record["metadata-only result record refs, digests, verdicts, scope boundaries"] Inputs --> Evaluate Evaluate --> Guards Guards -->|"finding or bad manifest"| Blocked Guards -->|"admissible"| Branch Branch -->|"yes"| Standalone Branch -->|"no"| Subprocess Standalone --> Record Subprocess --> Record Blocked --> Record
\n
Negative cases
\n
Fixture declares three
\n
The fixture declares three negative cases, each bound to a stable code in EXPECTED_NEGATIVE_CASES. The fixture declares three negative cases, each bound to a stable code in EXPECTED_NEGATIVE_CASES.
\n
Three faults, three stable codes
\n
A moved forecast cutoff raises FINANCE_LOOKAHEAD_SPLIT_FORBIDDEN, a flipped advice flag raises FINANCE_NO_ADVICE_OVERCLAIM, and an unavailable scipy yields a typed HLN refusal instead of a guessed p-value.finance_leakage_lookahead_split moves a forecast cutoff onto the event it forecasts and must raise FINANCE_LOOKAHEAD_SPLIT_FORBIDDEN. finance_no_advice_overclaim flips a boundary flag to authorize advice and must raise FINANCE_NO_ADVICE_OVERCLAIM. finance_hln_dependency_refusal is the case where scipy is unavailable for the small-sample t-distribution; the correct behavior is a typed refusal carrying FINANCE_HLN_TYPED_REFUSAL_REQUIRED, not a guessed p-value.
\n
Prior Art Grounding
\n
Standard forecast-evaluation econometrics
\n
The statistics are standard forecast-evaluation econometrics anchored by Diebold-Mariano, the HLN correction, Hansen's SPA test, and proper scoring rules, with a lookahead-split guard against temporal leakage. The statistics are standard forecast-evaluation econometrics, not a trading system. The core anchors are the Diebold-Mariano test for comparing predictive accuracy, the Harvey-Leybourne-Newbold small-sample correction (DOI reference), Hansen's test for superior predictive ability, and proper-scoring-rule work such as Gneiting and Raftery. The lookahead-split guard follows the financial machine-learning concern that temporal leakage makes backtests look stronger than they are.
\n
Validation Result record Path
\n
Run from repo
\n
Run from the repo root, without mutating durable result or projection surfaces. Run from the repo root, without mutating durable result or projection surfaces:
To exercise the component directly and inspect a result record, run the CLI and read the emitted JSON. To exercise the component directly and inspect a result record, run the CLI and read the emitted JSON:
\n
microcosm finance-forecast-evaluation-spine run --input fixtures/first_wave/finance_forecast_evaluation_spine/input --out receipts/first_wave/finance_forecast_evaluation_spine
\n
What a pass means here
\n
A green run computed the positive-fixture statistics, raised all three stable negative codes, kept the advice and live-data flags false, and admitted no source body. A pass means the positive fixture computed the statistics, the three negative cases raised their stable codes, the advice and live-data flags stayed false, and no source body entered the record.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
On synthetic, market-shaped fixtures it computes admissible forecast-evaluation statistics, returns typed refusals for inadmissible ones, and keeps copied source bodies out of every record. The strongest thing this component supports: on synthetic, market-shaped fixtures it computes admissible forecast-evaluation statistics and returns typed refusals for the inadmissible ones, while keeping copied source bodies out of every record. That is the proof boundary, and the validation result records above are what confirm it.
\n
The narrow scope limit
\n
This is not investment or trading decisions, uses no live market data, proves no track record or performance claim, mutates no optimizer, and certifies no strategy. The scope limit is narrow. This is not investment or trading decisions. It uses no live market data. It proves no track record and no performance claim. It mutates no optimizer and certifies no strategy. Absence of scipy is a typed HLN refusal, nothing stronger. The generated diagram and Atlas entry are navigation projections; they help a reader find the module and do not widen what it proves.
\n",
"components": [
{
"anchor": "component-finance_forecast_evaluation_spine",
"component_id": "finance_forecast_evaluation_spine",
"public_label": "Finance Forecast Evaluation Spine"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 1
},
"link_count": 6,
"links": [
{
"id": "finance_forecast_evaluation_spine",
"kind": "organ",
"kind_label": "component",
"label": "Finance Forecast Evaluation Spine",
"relation": "explains",
"url": "component-finance-forecast-evaluation-spine.html"
},
{
"id": "mechanism.finance_forecast_evaluation_spine.validates_public_finance_forecast_evaluation_spine",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public finance forecast evaluation spine",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-finance-forecast-evaluation-spine-validates-public-finance-forecast-evaluation-spine"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch12_prediction_market_board_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 12 Prediction Market Board Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[65:paper_module.batch12_prediction_market_board_capsule].depends_on",
"url": "paper-module-batch12-prediction-market-board-capsule.html"
},
{
"id": "paper_module.prediction_oracle_reconciliation",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Prediction Oracle Reconciliation",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[53:paper_module.prediction_oracle_reconciliation].depends_on",
"url": "paper-module-prediction-oracle-reconciliation.html"
}
],
"paper_module_id": "paper_module.finance_forecast_evaluation_spine",
"reader_packet": {
"abstract": "Forecast-evaluation component: Diebold-Mariano / Hansen-SPA / stationary-bootstrap stats over synthetic fixtures with typed refusal discipline; no market authority.",
"authority_ceiling": "Synthetic market-shaped fixtures only; not investment or trading decisions, no live market data, no track record or performance claim, mutates no optimizer, SciPy absence is a typed HLN refusal.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.finance_forecast_evaluation_spine.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-finance-forecast-evaluation-spine-diagram-1",
"html_url": "docs/paper-module-finance-forecast-evaluation-spine.html#paper-module-finance-forecast-evaluation-spine-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Inputs[\"Synthetic fixture family_loss_matrix, paired_loss_series, finance_boundary_policy\"] Evaluate[\"evaluate loads inputs\"] Guards[\"_policy_findings and _lookahead_split_findings\"] Blocked[\"Blocked record no statistics run\"] Branch{\"Exported bundle?\"} Standalone[\"_standalone_exported_statistics_contract declared, not recomputed\"] Subprocess[\"_run_stats_subprocess DM/HAC, SPA, MCS, bootstrap, HLN refusal\"] Record[\"metadata-only result record refs, digests, verdicts, scope boundaries\"] Inputs --> Evaluate Evaluate --> Guards Guards -->|\"finding or bad manifest\"| Blocked Guards -->|\"admissible\"| Branch Branch -->|\"yes\"| Standalone Branch -->|\"no\"| Subprocess Standalone --> Record Subprocess --> Record Blocked --> Record",
"source_format": "mermaid",
"source_hash": "sha256:859ac7b11920981c32f9c058e96d1df4bcb9293702d403eb2db854b202be9698",
"source_line_end": 169,
"source_line_start": 149
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "finance_forecast_evaluation_spine",
"kind": "organ",
"kind_label": "component",
"label": "Finance Forecast Evaluation Spine",
"relation": "explains",
"url": "component-finance-forecast-evaluation-spine.html"
},
{
"id": "mechanism.finance_forecast_evaluation_spine.validates_public_finance_forecast_evaluation_spine",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public finance forecast evaluation spine",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-finance-forecast-evaluation-spine-validates-public-finance-forecast-evaluation-spine"
}
],
"relation": "explains"
},
{
"count": 2,
"links": [
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.finance_forecast_evaluation_spine",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 259,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/finance_forecast_evaluation_spine.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Diebold-Mariano test for comparing predictive accuracy",
"section_id": "standard-forecast-evaluation-econometrics",
"url": "https://www.nber.org/papers/t0169"
},
{
"kind": "external_reference",
"label": "DOI reference",
"section_id": "standard-forecast-evaluation-econometrics",
"url": "https://doi.org/10.1016/S0169-2070%2896%2900719-4"
},
{
"kind": "external_reference",
"label": "test for superior predictive ability",
"section_id": "standard-forecast-evaluation-econometrics",
"url": "https://ideas.repec.org/a/bes/jnlbes/v23y2005p365-380.html"
},
{
"kind": "external_reference",
"label": "Gneiting and Raftery",
"section_id": "standard-forecast-evaluation-econometrics",
"url": "https://sites.stat.washington.edu/people/raftery/Research/PDF/Gneiting2007jasa.pdf"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component runs forecast-evaluation statistics over synthetic, market-shaped fixtures, records both the numbers computed and the cases refused, and writes a metadata-only result record from the copied econometric code.** financeforecastevaluation_spine runs forecast-evaluation statistics over synthetic, market-shaped fixtures and records both the numbers it computed and the cases it refused. It is a public component with runnable source. It reads a synthetic loss matrix and paired loss series, checks the request is admissible, runs the copied econometric code, and writes a metadata-only result record.",
"level": 4,
"order": 1,
"section_id": "component-runs-forecast-evaluation",
"source_line_end": 13,
"source_line_start": 3,
"title": "Component runs forecast-evaluation"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The value is the evaluation posture, not a claim about any asset. The value is the evaluation posture, not a claim about any asset.",
"level": 4,
"order": 2,
"section_id": "value-is-evaluation",
"source_line_end": 18,
"source_line_start": 14,
"title": "Value is evaluation"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The result record carries computed statistics when the fixture supports them and typed refusals when it does not, and both are treated as valid outcomes.** The result record carries computed statistics when the fixture supports them and typed refusals when it does not. Both are treated as valid outcomes.",
"level": 4,
"order": 3,
"section_id": "result-record-carries",
"source_line_end": 25,
"source_line_start": 19,
"title": "Result record carries"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 27,
"source_line_start": 26,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A lower average loss does not establish better forecasting because losses are autocorrelated, samples are short, and a careless split lets a model peek past its forecast point.** Comparing two forecasting models is harder than it looks. A lower average loss does not establish one model predicts better. Losses are autocorrelated, samples are short, and a careless time split lets a model peek past the point it was supposed to forecast from. A \"loose\" comparison, one that reports a mean and stops, produces a number that reads like a result but is not one.",
"level": 4,
"order": 5,
"section_id": "why-model-comparison-is-hard",
"source_line_end": 37,
"source_line_start": 28,
"title": "Why model comparison is hard"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It carries the careful machinery economists use, Diebold-Mariano with a Bartlett HAC variance, the HLN correction, Hansen's SPA test, a model confidence set, and a stationary bootstrap, run on synthetic data only.** This component carries the machinery economists use to answer the question carefully: the Diebold-Mariano loss-differential statistic with a Bartlett HAC long-run variance, the Harvey-Leybourne-Newbold small-sample correction, Hansen's test for superior predictive ability, a model confidence set, and a Politis-Romano stationary bootstrap. It runs them on synthetic data only, and it refuses when the input is inadmissible instead of returning a number that could be misread.",
"level": 4,
"order": 6,
"section_id": "the-econometric-machinery-it-carries",
"source_line_end": 49,
"source_line_start": 38,
"title": "The econometric machinery it carries"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 51,
"source_line_start": 50,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component is a thin runtime over a CrownJewelSpec that fixes the four required inputs, names the three negative cases, binds the source manifest, and restricts reads to anchors in four copied modules.** The component is a thin runtime over a CrownJewelSpec named SPEC. The spec fixes the four required inputs (familylossmatrix.json, pairedlossseries.json, financeboundarypolicy.json, projection_protocol.json), names the three negative cases, binds the source manifest, and restricts source-open reads to required anchors in four copied modules: modelselectionstats.py, spastatistics.py, lossdifferentials.py, and familylossmatrix.py.",
"level": 4,
"order": 8,
"section_id": "a-thin-runtime-over-spec",
"source_line_end": 63,
"source_line_start": 52,
"title": "A thin runtime over `SPEC`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run and runfinanceforecastbundle both delegate to runcrownjewel_organ, differing only in that the bundle entrypoint sets input_mode to the exported-bundle mode. run and runfinanceforecastbundle both delegate to runcrownjewelorgan with evaluate and evaluatenegative_case. The bundle entrypoint sets input_mode to the exported-bundle mode; that is the only difference.",
"level": 4,
"order": 9,
"section_id": "two-entrypoints-one-bundle-difference",
"source_line_end": 72,
"source_line_start": 64,
"title": "Two entrypoints, one bundle difference"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**evaluate loads the loss matrix, paired loss series, and boundary policy, then calls evaluatepayloads, where the order of operations matters.** evaluate loads the loss matrix, the paired loss series, and the boundary policy, then calls evaluatepayloads. That function is where the order of operations matters.",
"level": 4,
"order": 10,
"section_id": "evaluate-loads-inputs-order-matters",
"source_line_end": 79,
"source_line_start": 73,
"title": "`evaluate` loads inputs, order matters"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "policyfindings and lookaheadsplit_findings run first, blocking on any non-False advice flag or forecast-date leakage, so an inadmissible request never reaches a statistic. The guards run first. policy_findings walks FALSEBOUNDARYFLAGS (investmentadviceauthorized, tradingadviceauthorized, livemarketdataauthorized, trackrecordclaimauthorized, performanceclaimauthorized, optimizermutationauthorized) and raises FINANCENOADVICE_OVERCLAIM for any flag that is not exactly False. lookaheadsplitfindings parses the row dates with parseisodate and raises FINANCELOOKAHEADSPLITFORBIDDEN when a row's subjectas_of reaches its event_start, or when an event window ends before it begins. If either guard prod\u2026",
"level": 4,
"order": 11,
"section_id": "guards-run-before-any-computation",
"source_line_end": 95,
"source_line_start": 80,
"title": "Guards run before any computation"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Only after the guards pass does isexportedfinance_bundle branch to a declared, unexecuted contract for the exported bundle or a real child-interpreter subprocess for the first-wave fixture with scipy blocked.** Only after the guards pass does the branch happen. isexportedfinancebundle decides the path. For the exported public bundle, standaloneexportedstatisticscontract returns a declared contract: the statistics are named and marked executed: false, with notaliverun and realruntime_receipt: false, so the standalone bundle does not depend on a live source subprocess. For the first-wave fixture, runstatssubprocess (memoized by runstatssubprocess_cached) shells out to a child interpreter that\u2026",
"level": 4,
"order": 12,
"section_id": "branching-to-standalone-or-subprocess-stats",
"source_line_end": 114,
"source_line_start": 96,
"title": "Branching to standalone or subprocess stats"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Back in the parent, evaluate_payloads checks each statistic field against its expected status, blocks on any mismatch, and keeps source bodies out with bodyinreceipt: false and hashed stdout and stderr.** Back in the parent, evaluatepayloads reads the subprocess payload and checks each field: Reality Check and SPA must be computed_bootstrap, the model confidence set must report implemented, Diebold-Mariano must be computedhacnormal_approximation, and the HLN entry must be a refusal with reason scipyunavailablefortdistribution. Any mismatch becomes a finding and the record is blocked. Source bodies never enter the record; every emitted stats object keeps bodyinreceipt: false, and the subproc\u2026",
"level": 4,
"order": 13,
"section_id": "checking-the-subprocess-payload",
"source_line_end": 127,
"source_line_start": 115,
"title": "Checking the subprocess payload"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "evaluatenegative_case mutates one input to reproduce a specific fault, re-runs evaluatepayloads, and confirms the expected stable error code appears.** evaluatenegativecase reuses the same path. It loads the fixture, mutates one input to reproduce a specific fault (moving a row's subjectasof onto its eventstart for the leakage case, or setting investmentadvice_authorized true for the overclaim case), re-runs evaluatepayloads, and confirms the expected stable error code appears.",
"level": 4,
"order": 14,
"section_id": "negative-cases-reuse-the-same-path",
"source_line_end": 170,
"source_line_start": 128,
"title": "Negative cases reuse the same path"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "negative-cases",
"source_line_end": 172,
"source_line_start": 171,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The fixture declares three negative cases, each bound to a stable code in EXPECTEDNEGATIVECASES.** The fixture declares three negative cases, each bound to a stable code in EXPECTEDNEGATIVECASES.",
"level": 4,
"order": 16,
"section_id": "fixture-declares-three",
"source_line_end": 178,
"source_line_start": 173,
"title": "Fixture declares three"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A moved forecast cutoff raises FINANCELOOKAHEADSPLITFORBIDDEN, a flipped advice flag raises FINANCENOADVICE_OVERCLAIM, and an unavailable scipy yields a typed HLN refusal instead of a guessed p-value. financeleakagelookaheadsplit moves a forecast cutoff onto the event it forecasts and must raise FINANCELOOKAHEADSPLITFORBIDDEN. financenoadviceoverclaim flips a boundary flag to authorize advice and must raise FINANCENOADVICE_OVERCLAIM. financehlndependency_refusal is the case where scipy is unavailable for the small-sample t-distribution; the correct behavior is a typed refusal carrying FINANCEHLNTYPEDREFUSALREQUIRED, not a guessed p-value.",
"level": 4,
"order": 17,
"section_id": "three-faults-three-stable-codes",
"source_line_end": 190,
"source_line_start": 179,
"title": "Three faults, three stable codes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "prior-art-grounding",
"source_line_end": 192,
"source_line_start": 191,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The statistics are standard forecast-evaluation econometrics anchored by Diebold-Mariano, the HLN correction, Hansen's SPA test, and proper scoring rules, with a lookahead-split guard against temporal leakage.** The statistics are standard forecast-evaluation econometrics, not a trading system. The core anchors are the Diebold-Mariano test for comparing predictive accuracy, the Harvey-Leybourne-Newbold small-sample correction (DOI reference), Hansen's test for superior predictive ability, and proper-scoring-rule work such as Gneiting and Raftery. The lookahead-split guard follows the financial machine-learning concern that temporal leakage makes backtests look stronger than they are.",
"level": 4,
"order": 19,
"section_id": "standard-forecast-evaluation-econometrics",
"source_line_end": 205,
"source_line_start": 193,
"title": "Standard forecast-evaluation econometrics"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "validation-receipt-path",
"source_line_end": 207,
"source_line_start": 206,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from the repo root, without mutating durable result or projection surfaces. Run from the repo root, without mutating durable result or projection surfaces:",
"level": 4,
"order": 21,
"section_id": "run-from-repo",
"source_line_end": 217,
"source_line_start": 208,
"title": "Run from repo"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "**To exercise the component directly and inspect a result record, run the CLI and read the emitted JSON.** To exercise the component directly and inspect a result record, run the CLI and read the emitted JSON:",
"level": 4,
"order": 22,
"section_id": "to-exercise-organ",
"source_line_end": 227,
"source_line_start": 218,
"title": "To exercise component"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A green run computed the positive-fixture statistics, raised all three stable negative codes, kept the advice and live-data flags false, and admitted no source body.** A pass means the positive fixture computed the statistics, the three negative cases raised their stable codes, the advice and live-data flags stayed false, and no source body entered the record.",
"level": 4,
"order": 23,
"section_id": "what-a-pass-means-here",
"source_line_end": 235,
"source_line_start": 228,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 24,
"section_id": "claim-ceiling",
"source_line_end": 237,
"source_line_start": 236,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**On synthetic, market-shaped fixtures it computes admissible forecast-evaluation statistics, returns typed refusals for inadmissible ones, and keeps copied source bodies out of every record.** The strongest thing this component supports: on synthetic, market-shaped fixtures it computes admissible forecast-evaluation statistics and returns typed refusals for the inadmissible ones, while keeping copied source bodies out of every record. That is the proof boundary, and the validation result records above are what confirm it.",
"level": 4,
"order": 25,
"section_id": "the-strongest-supported-claim",
"source_line_end": 247,
"source_line_start": 238,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This is not investment or trading decisions, uses no live market data, proves no track record or performance claim, mutates no optimizer, and certifies no strategy.** The scope limit is narrow. This is not investment or trading decisions. It uses no live market data. It proves no track record and no performance claim. It mutates no optimizer and certifies no strategy. Absence of scipy is a typed HLN refusal, nothing stronger. The generated diagram and Atlas entry are navigation projections; they help a reader find the module and do not widen what it proves.",
"level": 4,
"order": 26,
"section_id": "the-narrow-authority-ceiling",
"source_line_end": 258,
"source_line_start": 248,
"title": "The narrow scope limit"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/finance_forecast_evaluation_spine.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Forecast-evaluation component: Diebold-Mariano / Hansen-SPA / stationary-bootstrap stats over synthetic fixtures with typed refusal discipline; no market authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "finance_forecast_evaluation_spine.py",
"path": "src/microcosm_core/organs/finance_forecast_evaluation_spine.py",
"resolution": "resolved",
"role": "Runtime + result record-writing locus for the public forecast-evaluation fixture and exported bundle.",
"symbols": [
"run",
"run_finance_forecast_bundle"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/finance_forecast_evaluation_spine.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/finance_forecast_evaluation_spine.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/finance_forecast_evaluation_spine.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Runs admissible forecast-evaluation statistics (Diebold-Mariano loss-differential, Harvey-Leybourne-Newbold small-sample correction, Hansen SPA with recentering, Politis-Romano stationary bootstrap, Bartlett HAC long-run variance, purged/embargoed CV) over synthetic market-shaped fixtures and copied source bodies; refusal discipline returns typed refusals (horizon>=sample length, too-small samples, leakage-prone splits, missing SciPy, advice-shaped claims) instead of crashing; computed-statistic and refused-because-inadmissible are both valid validator outcomes.",
"summary_line": "Forecast-evaluation component: Diebold-Mariano / Hansen-SPA / stationary-bootstrap stats over synthetic fixtures with typed refusal discipline; no market authority.",
"title": "Finance Forecast Evaluation Spine"
},
{
"anchor": "paper-module-engine-room-demo",
"authority_ceiling": "Validates the declared public composition contract only; not deployment posture, not whole-system equivalence, not a frontier theorem-proving claim, not a complete security proof, not benchmark validation, not launch-scope decision.",
"body_html": "
Engine room is
\n
The Engine Room is fourteen public capabilities staged as components, and engine_room_demo is the one that checks all of them together. The Engine Room is fourteen separate capabilities staged as public components: four Lean proof-search targets, a metabolism runtime with its reconciler, command singleflight, a generated-projection drift gate, a derived-fact engine, a public-projection leak gate, an egress self-compliance gate, a navigation-fitness benchmark, a bridge-campaign DAG, and a knowledge router over the reference library. engine_room_demo is the component that checks all of them together.
\n
Asks whether ten
\n
It asks whether the ten staged bundles cover the fourteen requested targets, each still owns its files, and each fixture still runs, not whether any bundle is finished. It answers one question. Do the ten staged bundles cover the fourteen targets the controller asked for, does each bundle still own its full set of files, and does each fixture still run? A pass means the set is complete and every fixture executed. It does not mean that any single bundle is finished or correct.
\n
Reads fixtures runs
\n
It reads fixtures, runs the shared audit, and writes four metadata-only records, reading the shared registry and generated atlas but never writing to them. The component reads fixtures, runs the shared audit, and writes four metadata-only result records: a result, a board, a validation record, and an optional fixture sign-off record. It reads the shared component registry and the generated atlas but never writes to them.
\n
Purpose
\n
A stored list would drift and mislead
\n
Without this component a reader trusts fourteen claims one at a time, and a stored list of target names would go stale the moment the bundle set changed and start agreeing with a wrong answer. Without this component a reader has to trust fourteen claims one at a time and hope the set is complete. A loose check would compare the covered targets against a stored list of the fourteen names. That list goes stale the moment the bundle set changes, and a stale list can start agreeing with a wrong answer.
\n
Coverage recomputed from EXPECTED_JEWEL_TARGETS
\n
engine_room_demo keeps no stored list: the fourteen names live in the frozen EXPECTED_JEWEL_TARGETS, and coverage is recomputed from the bundles every run, so nothing drifts out of sync.engine_room_demo keeps no stored list. The fourteen expected names live in one frozen constant, EXPECTED_JEWEL_TARGETS, and the covered set is recomputed from the bundles on every run. Coverage is the union of what the bundles declare, checked against that constant. There is nothing to drift out of sync.
\n
Six files per bundle mark ownership
\n
A bundle owns its surface only when six files exist for it, module source, fixture input, manifest, paper module, standard, and test, and the audit names any that are missing. A bundle counts as owning its surface only when six files exist for it: module source, fixture input directory, fixture manifest, paper module, standard, and test. The audit checks all six per bundle and names any that are missing.
\n
How it works
\n
Wrapper is src
\n
The wrapper is src/microcosm_core/organs/engine_room_demo.py. The staged runner it calls is src/microcosm_core/engine_room/demo.py. Execution runs in this order.
\n
build_result folds cases into one status
\n
build_result loads every JSON case, runs _evaluate_case on each, and passes only with at least one positive case, all positives passing, and every negative observed to fail.build_result is the entry point. It calls _fixture_cases to load every JSON case under the input path, runs _evaluate_case on each one, then folds the rows into a status. The run passes only when there is at least one positive case, every positive case passed, and every negative case was observed to fail.
\n
_evaluate_case routes by case_type
\n
_evaluate_case sends a positive case to audit_controller_coverage, a negative case to _target_override_case, and any other type to a fail row coded ENGINE_ROOM_UNKNOWN_CASE_TYPE._evaluate_case dispatches on case_type. A positive case calls audit_controller_coverage. A negative case calls _target_override_case. Any other type returns a fail row with the code ENGINE_ROOM_UNKNOWN_CASE_TYPE.
\n
audit_controller_coverage unions targets both ways
\n
audit_controller_coverage runs the six-file check per bundle, unions declared targets, subtracts EXPECTED_JEWEL_TARGETS both ways for missing and unexpected, and reads shared surfaces for visibility only.audit_controller_coverage is the audit. It walks the ten entries in CAPSULES, builds a _capsule_surface_row for each (the six-file check), and unions their declared targets. It subtracts EXPECTED_JEWEL_TARGETS both ways to find missing and unexpected targets. When asked, it calls run_demo to execute the fixtures. It reads the shared registry, the accepted-component list, and the atlas through _shared_surface_ids for visibility only, and always returns shared_registry_mutated: false. The staged pass needs no missing targets, no unexpected targets, no bundle missing a file, and a passing fixture run.
\n
run_demo calls each bundle's evaluator
\n
run_demo is the execution spine: for each bundle it imports the module and calls evaluate_fixture_dir or validate_fixture_dir, passing only when every selected fixture ran and passed.run_demo is the execution spine. For each bundle it calls run_capsule, which imports the bundle module and calls its declared evaluator, either evaluate_fixture_dir or validate_fixture_dir. It records a compact per-bundle status and the covered target list. A pass here means every selected bundle fixture ran and passed.
\n
run writes four metadata-only records
\n
The wrapper's run writes four records through write_json_atomic, a result, a board, a validation result record, and an optional sign-off record, each carrying body_in_receipt: false. The wrapper's run writes the four records with write_json_atomic: engine_room_demo_result.json, engine_room_demo_board.json, engine_room_demo_validation_receipt.json, and, when an sign-off path is given, engine_room_demo_fixture_acceptance.json. Every record carries body_in_receipt: false.
\n
Function
Role
build_result
Loads the fixture cases, evaluates each, and folds them into one pass or fail status.
_evaluate_case
Routes a case by case_type to the coverage audit or the negative-case check.
audit_controller_coverage
Unions declared targets, runs the six-file check per bundle, reads shared surfaces for visibility only.
_capsule_surface_row
Confirms the six required files exist for one bundle and lists any missing kind.
run_demo, run_capsule
Import each bundle and call its declared evaluator over the public fixtures.
_target_override_case
Recomputes the covered set live and fails when a named target is genuinely absent.
run
Writes the four metadata-only records through write_json_atomic.
\nDiagram of the mechanism (10 steps).Diagram source & refs
Source refs
positive:
audit_controller_coverage
flowchart TD A["fixture cases"] --> B["build_result"] B --> C["_evaluate_case"] C --> D["positive: audit_controller_coverage"] C --> E["negative: _target_override_case"] D --> F["_capsule_surface_row x10 bundles"] D --> G["run_demo --> run_capsule"] D --> H["shared surfaces read-only"] B --> I["run writes 4 metadata-only records"]
\n
Negative cases
\n
The negative case must be observed to fail
\n
One positive and one negative fixture run, and _target_override_case recomputes coverage live so the missing-target negative fails, since a silently passing negative would fail the whole run. The fixture set names one positive case, positive_controller_audit.json, and one negative case, missing_expected_target_negative.json. The negative case names an expected target that is not in the covered set. _target_override_case recomputes the covered targets live, finds the named target missing, and returns a fail row with the code ENGINE_ROOM_EXPECTED_TARGET_MISSING. The run passes only when that negative case is observed to fail, so a negative case that silently passed would fail the whole run. There is no frozen answer key for the negative check to drift toward.
\n
Prior Art Grounding
\n
Lineage in integration testing and CI composition
\n
This is ordinary integration-testing and CI-composition practice, anchored by integration testing, pytest fixtures, and staged status checks, without turning fixture evidence into a launch claim. This is ordinary integration-testing and CI-composition practice: several component checks are assembled into one demo path, negative fixtures prove refusal behaviour, and the records summarise what ran without turning fixture evidence into a launch claim. Reasonable public anchors are IBM's integration testing overview, which frames the question as whether composed modules interact as intended, pytest fixtures as a common shape for reusable synthetic inputs, and GitHub Actions as a widely used surface for composing staged checks with explicit status.
\n
Validation Result record Path
\n
Run component and
\n
What a pass proves here
\n
A pass means the fixtures ran, the ten bundles covered the fourteen targets with no file missing, the negative failed, and this Markdown matches its generated structured source record. A pass means the fixtures ran, the ten bundles covered the fourteen expected targets with no file missing, the negative case was observed to fail, and this Markdown still matches its generated structured source record. It is not evidence of anything past that.
\n
Scope boundary
\n
Scope limit
\n
The bounded strongest claim
\n
At most, the ten bundles cover the fourteen targets, each owns its six files, and each fixture passed, with shared_registry_mutated: false fixing the scope limit. The strongest claim the evidence supports is bounded: the ten staged Engine Room bundles together cover the fourteen expected targets, each bundle owns its six required files, and each fixture ran and passed on the selected public inputs. The proof boundary is exactly that fixture run plus the six-file check. Nothing past it is claimed. It is not deployment posture, not whole-system equivalence, not a frontier theorem-proving claim, not a complete security proof, not benchmark validation, and not launch-scope decision. The scope limit is fixed by shared_registry_mutated: false: the audit reads the shared registry, accepted-component list, and atlas for visibility, but changes none of them, so a green run cannot imply registry authority it does not hold.
\n",
"components": [
{
"anchor": "component-engine_room_demo",
"component_id": "engine_room_demo",
"public_label": "Engine Room Demo"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 9,
"organ": 1,
"paper_module": 10,
"principle": 9
},
"link_count": 36,
"links": [
{
"id": "engine_room_demo",
"kind": "organ",
"kind_label": "component",
"label": "Engine Room Demo",
"relation": "explains",
"url": "component-engine-room-demo.html"
},
{
"id": "mechanism.engine_room_demo.validates_public_engine_room_demo",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public engine room demo",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-demo-validates-public-engine-room-demo"
},
{
"id": "mechanism.engine_room_public_projection_leak_gate.validates_public_projection_leak_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public projection leak gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-public-projection-leak-gate-validates-public-projection-leak-gate"
},
{
"id": "mechanism.engine_room_generated_projection_drift_gate.validates_public_generated_projection_drift_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public generated projection drift gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-generated-projection-drift-gate-validates-public-generated-projection-drift-gate"
},
{
"id": "mechanism.engine_room_command_run_singleflight.validates_public_command_run_singleflight",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public command run singleflight",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-command-run-singleflight-validates-public-command-run-singleflight"
},
{
"id": "mechanism.engine_room_annex_knowledge_router.validates_public_annex_knowledge_router",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public reference knowledge router",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-annex-knowledge-router-validates-public-annex-knowledge-router"
},
{
"id": "mechanism.engine_room_derived_fact_provider_engine.validates_public_derived_fact_provider_engine",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public derived fact provider engine",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-derived-fact-provider-engine-validates-public-derived-fact-provider-engine"
},
{
"id": "mechanism.engine_room_egress_self_compliance_gate.validates_public_egress_self_compliance_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public egress self compliance gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-egress-self-compliance-gate-validates-public-egress-self-compliance-gate"
},
{
"id": "mechanism.engine_room_lean_proof_search_lab.validates_public_lean_proof_search_lab",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public lean proof search lab",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-lean-proof-search-lab-validates-public-lean-proof-search-lab"
},
{
"id": "mechanism.engine_room_navigation_fitness_benchmark.validates_public_navigation_fitness_benchmark",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public navigation fitness benchmark",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-navigation-fitness-benchmark-validates-public-navigation-fitness-benchmark"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.engine_room_lean_proof_search_lab",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Lean Proof Search Lab",
"relation": "depends on",
"url": "paper-module-engine-room-lean-proof-search-lab.html"
},
{
"id": "paper_module.engine_room_metabolism_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Metabolism Runtime",
"relation": "depends on",
"url": "paper-module-engine-room-metabolism-runtime.html"
},
{
"id": "paper_module.engine_room_command_run_singleflight",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Command-Run Singleflight",
"relation": "depends on",
"url": "paper-module-engine-room-command-run-singleflight.html"
},
{
"id": "paper_module.engine_room_generated_projection_drift_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Generated Projection Drift Gate",
"relation": "depends on",
"url": "paper-module-engine-room-generated-projection-drift-gate.html"
},
{
"id": "paper_module.engine_room_derived_fact_provider_engine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Derived Fact Provider Engine",
"relation": "depends on",
"url": "paper-module-engine-room-derived-fact-provider-engine.html"
},
{
"id": "paper_module.engine_room_public_projection_leak_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Public Projection Leak Gate",
"relation": "depends on",
"url": "paper-module-engine-room-public-projection-leak-gate.html"
},
{
"id": "paper_module.engine_room_egress_self_compliance_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Egress Self-Compliance Gate",
"relation": "depends on",
"url": "paper-module-engine-room-egress-self-compliance-gate.html"
},
{
"id": "paper_module.engine_room_navigation_fitness_benchmark",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Navigation Fitness Benchmark",
"relation": "depends on",
"url": "paper-module-engine-room-navigation-fitness-benchmark.html"
},
{
"id": "paper_module.engine_room_bridge_campaign_dag",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Bridge Campaign DAG",
"relation": "depends on",
"url": "paper-module-engine-room-bridge-campaign-dag.html"
},
{
"id": "paper_module.engine_room_annex_knowledge_router",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Reference Knowledge Router",
"relation": "depends on",
"url": "paper-module-engine-room-annex-knowledge-router.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.engine_room_annex_knowledge_router",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Reference Knowledge Router",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[84:paper_module.engine_room_annex_knowledge_router].depends_on",
"url": "paper-module-engine-room-annex-knowledge-router.html"
},
{
"id": "paper_module.engine_room_bridge_campaign_dag",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Bridge Campaign DAG",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[83:paper_module.engine_room_bridge_campaign_dag].depends_on",
"url": "paper-module-engine-room-bridge-campaign-dag.html"
},
{
"id": "paper_module.engine_room_derived_fact_provider_engine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Derived Fact Provider Engine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[85:paper_module.engine_room_derived_fact_provider_engine].depends_on",
"url": "paper-module-engine-room-derived-fact-provider-engine.html"
},
{
"id": "paper_module.engine_room_egress_self_compliance_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Egress Self-Compliance Gate",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[86:paper_module.engine_room_egress_self_compliance_gate].depends_on",
"url": "paper-module-engine-room-egress-self-compliance-gate.html"
},
{
"id": "paper_module.engine_room_lean_proof_search_lab",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Lean Proof Search Lab",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[87:paper_module.engine_room_lean_proof_search_lab].depends_on",
"url": "paper-module-engine-room-lean-proof-search-lab.html"
},
{
"id": "paper_module.engine_room_metabolism_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Metabolism Runtime",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[82:paper_module.engine_room_metabolism_runtime].depends_on",
"url": "paper-module-engine-room-metabolism-runtime.html"
},
{
"id": "paper_module.engine_room_navigation_fitness_benchmark",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Navigation Fitness Benchmark",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[88:paper_module.engine_room_navigation_fitness_benchmark].depends_on",
"url": "paper-module-engine-room-navigation-fitness-benchmark.html"
},
{
"id": "paper_module.engine_room_public_projection_leak_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Public Projection Leak Gate",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[76:paper_module.engine_room_public_projection_leak_gate].depends_on",
"url": "paper-module-engine-room-public-projection-leak-gate.html"
}
],
"paper_module_id": "paper_module.engine_room_demo",
"reader_packet": {
"abstract": "Composition component: verifies the 14 staged Engine Room jewel targets and their owned bundle surfaces through the public fixture chain; composition contract only.",
"authority_ceiling": "Validates the declared public composition contract only; not deployment posture, not whole-system equivalence, not a frontier theorem-proving claim, not a complete security proof, not benchmark validation, not launch-scope decision.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.engine_room_demo.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-engine-room-demo-diagram-1",
"html_url": "docs/paper-module-engine-room-demo.html#paper-module-engine-room-demo-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"fixture cases\"] --> B[\"build_result\"] B --> C[\"_evaluate_case\"] C --> D[\"positive: audit_controller_coverage\"] C --> E[\"negative: _target_override_case\"] D --> F[\"_capsule_surface_row x10 bundles\"] D --> G[\"run_demo --> run_capsule\"] D --> H[\"shared surfaces read-only\"] B --> I[\"run writes 4 metadata-only records\"]",
"source_format": "mermaid",
"source_hash": "sha256:187be00b480cb8639d795ffd5625c1d9c5ba5343171a60592a5c0803a2917ce5",
"source_line_end": 76,
"source_line_start": 66
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 10,
"links": [
{
"id": "paper_module.engine_room_lean_proof_search_lab",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Lean Proof Search Lab",
"relation": "depends on",
"url": "paper-module-engine-room-lean-proof-search-lab.html"
},
{
"id": "paper_module.engine_room_metabolism_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Metabolism Runtime",
"relation": "depends on",
"url": "paper-module-engine-room-metabolism-runtime.html"
},
{
"id": "paper_module.engine_room_command_run_singleflight",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Command-Run Singleflight",
"relation": "depends on",
"url": "paper-module-engine-room-command-run-singleflight.html"
},
{
"id": "paper_module.engine_room_generated_projection_drift_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Generated Projection Drift Gate",
"relation": "depends on",
"url": "paper-module-engine-room-generated-projection-drift-gate.html"
},
{
"id": "paper_module.engine_room_derived_fact_provider_engine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Derived Fact Provider Engine",
"relation": "depends on",
"url": "paper-module-engine-room-derived-fact-provider-engine.html"
},
{
"id": "paper_module.engine_room_public_projection_leak_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Public Projection Leak Gate",
"relation": "depends on",
"url": "paper-module-engine-room-public-projection-leak-gate.html"
},
{
"id": "paper_module.engine_room_egress_self_compliance_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Egress Self-Compliance Gate",
"relation": "depends on",
"url": "paper-module-engine-room-egress-self-compliance-gate.html"
},
{
"id": "paper_module.engine_room_navigation_fitness_benchmark",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Navigation Fitness Benchmark",
"relation": "depends on",
"url": "paper-module-engine-room-navigation-fitness-benchmark.html"
},
{
"id": "paper_module.engine_room_bridge_campaign_dag",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Bridge Campaign DAG",
"relation": "depends on",
"url": "paper-module-engine-room-bridge-campaign-dag.html"
},
{
"id": "paper_module.engine_room_annex_knowledge_router",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Reference Knowledge Router",
"relation": "depends on",
"url": "paper-module-engine-room-annex-knowledge-router.html"
}
],
"relation": "depends on"
},
{
"count": 10,
"links": [
{
"id": "engine_room_demo",
"kind": "organ",
"kind_label": "component",
"label": "Engine Room Demo",
"relation": "explains",
"url": "component-engine-room-demo.html"
},
{
"id": "mechanism.engine_room_demo.validates_public_engine_room_demo",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public engine room demo",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-demo-validates-public-engine-room-demo"
},
{
"id": "mechanism.engine_room_public_projection_leak_gate.validates_public_projection_leak_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public projection leak gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-public-projection-leak-gate-validates-public-projection-leak-gate"
},
{
"id": "mechanism.engine_room_generated_projection_drift_gate.validates_public_generated_projection_drift_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public generated projection drift gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-generated-projection-drift-gate-validates-public-generated-projection-drift-gate"
},
{
"id": "mechanism.engine_room_command_run_singleflight.validates_public_command_run_singleflight",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public command run singleflight",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-command-run-singleflight-validates-public-command-run-singleflight"
},
{
"id": "mechanism.engine_room_annex_knowledge_router.validates_public_annex_knowledge_router",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public reference knowledge router",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-annex-knowledge-router-validates-public-annex-knowledge-router"
},
{
"id": "mechanism.engine_room_derived_fact_provider_engine.validates_public_derived_fact_provider_engine",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public derived fact provider engine",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-derived-fact-provider-engine-validates-public-derived-fact-provider-engine"
},
{
"id": "mechanism.engine_room_egress_self_compliance_gate.validates_public_egress_self_compliance_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public egress self compliance gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-egress-self-compliance-gate-validates-public-egress-self-compliance-gate"
},
{
"id": "mechanism.engine_room_lean_proof_search_lab.validates_public_lean_proof_search_lab",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public lean proof search lab",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-lean-proof-search-lab-validates-public-lean-proof-search-lab"
},
{
"id": "mechanism.engine_room_navigation_fitness_benchmark.validates_public_navigation_fitness_benchmark",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public navigation fitness benchmark",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-navigation-fitness-benchmark-validates-public-navigation-fitness-benchmark"
}
],
"relation": "explains"
},
{
"count": 10,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.engine_room_demo",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 117,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_demo.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "integration testing",
"section_id": "lineage-in-integration-testing-and-ci-composition",
"url": "https://www.ibm.com/think/topics/integration-testing"
},
{
"kind": "external_reference",
"label": "pytest fixtures",
"section_id": "lineage-in-integration-testing-and-ci-composition",
"url": "https://docs.pytest.org/en/stable/reference/fixtures.html"
},
{
"kind": "external_reference",
"label": "GitHub Actions",
"section_id": "lineage-in-integration-testing-and-ci-composition",
"url": "https://github.com/features/actions"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "The Engine Room is fourteen public capabilities staged as components, and engineroomdemo is the one that checks all of them together. The Engine Room is fourteen separate capabilities staged as public components: four Lean proof-search targets, a metabolism runtime with its reconciler, command singleflight, a generated-projection drift gate, a derived-fact engine, a public-projection leak gate, an egress self-compliance gate, a navigation-fitness benchmark, a bridge-campaign DAG, and a knowledge router over the reference library. engineroomdemo is the component that checks all of them together.",
"level": 4,
"order": 1,
"section_id": "engine-room-is",
"source_line_end": 6,
"source_line_start": 3,
"title": "Engine room is"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It asks whether the ten staged bundles cover the fourteen requested targets, each still owns its files, and each fixture still runs, not whether any bundle is finished. It answers one question. Do the ten staged bundles cover the fourteen targets the controller asked for, does each bundle still own its full set of files, and does each fixture still run? A pass means the set is complete and every fixture executed. It does not mean that any single bundle is finished or correct.",
"level": 4,
"order": 2,
"section_id": "asks-whether-ten",
"source_line_end": 10,
"source_line_start": 7,
"title": "Asks whether ten"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It reads fixtures, runs the shared audit, and writes four metadata-only records, reading the shared registry and generated atlas but never writing to them. The component reads fixtures, runs the shared audit, and writes four metadata-only result records: a result, a board, a validation record, and an optional fixture sign-off record. It reads the shared component registry and the generated atlas but never writes to them.",
"level": 4,
"order": 3,
"section_id": "reads-fixtures-runs",
"source_line_end": 14,
"source_line_start": 11,
"title": "Reads fixtures runs"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Without this component a reader trusts fourteen claims one at a time, and a stored list of target names would go stale the moment the bundle set changed and start agreeing with a wrong answer. Without this component a reader has to trust fourteen claims one at a time and hope the set is complete. A loose check would compare the covered targets against a stored list of the fourteen names. That list goes stale the moment the bundle set changes, and a stale list can start agreeing with a wrong answer.",
"level": 4,
"order": 5,
"section_id": "a-stored-list-would-drift-and-mislead",
"source_line_end": 20,
"source_line_start": 17,
"title": "A stored list would drift and mislead"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "engineroomdemo keeps no stored list: the fourteen names live in the frozen EXPECTEDJEWELTARGETS, and coverage is recomputed from the bundles every run, so nothing drifts out of sync. engineroomdemo keeps no stored list. The fourteen expected names live in one frozen constant, EXPECTEDJEWELTARGETS, and the covered set is recomputed from the bundles on every run. Coverage is the union of what the bundles declare, checked against that constant. There is nothing to drift out of sync.",
"level": 4,
"order": 6,
"section_id": "coverage-recomputed-from-expected-jewel-targets",
"source_line_end": 24,
"source_line_start": 21,
"title": "Coverage recomputed from `EXPECTED_JEWEL_TARGETS`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A bundle owns its surface only when six files exist for it, module source, fixture input, manifest, paper module, standard, and test, and the audit names any that are missing. A bundle counts as owning its surface only when six files exist for it: module source, fixture input directory, fixture manifest, paper module, standard, and test. The audit checks all six per bundle and names any that are missing.",
"level": 4,
"order": 7,
"section_id": "six-files-per-capsule-mark-ownership",
"source_line_end": 28,
"source_line_start": 25,
"title": "Six files per bundle mark ownership"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 30,
"source_line_start": 29,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The wrapper is src/microcosmcore/components/engineroomdemo.py.** The staged runner it calls is src/microcosmcore/engineroom/demo.py. Execution runs in this order.",
"level": 4,
"order": 9,
"section_id": "wrapper-is-src",
"source_line_end": 35,
"source_line_start": 31,
"title": "Wrapper is src"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult loads every JSON case, runs evaluatecase on each, and passes only with at least one positive case, all positives passing, and every negative observed to fail. buildresult is the entry point. It calls fixturecases to load every JSON case under the input path, runs evaluatecase on each one, then folds the rows into a status. The run passes only when there is at least one positive case, every positive case passed, and every negative case was observed to fail.",
"level": 4,
"order": 10,
"section_id": "build-result-folds-cases-into-one-status",
"source_line_end": 39,
"source_line_start": 36,
"title": "`build_result` folds cases into one status"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluatecase sends a positive case to auditcontrollercoverage, a negative case to targetoverridecase, and any other type to a fail row coded ENGINEROOMUNKNOWNCASETYPE. evaluatecase dispatches on casetype. A positive case calls auditcontrollercoverage. A negative case calls targetoverridecase. Any other type returns a fail row with the code ENGINEROOMUNKNOWNCASETYPE.",
"level": 4,
"order": 11,
"section_id": "evaluate-case-routes-by-case-type",
"source_line_end": 43,
"source_line_start": 40,
"title": "`_evaluate_case` routes by `case_type`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "auditcontrollercoverage runs the six-file check per bundle, unions declared targets, subtracts EXPECTEDJEWELTARGETS both ways for missing and unexpected, and reads shared surfaces for visibility only. auditcontrollercoverage is the audit. It walks the ten entries in Bundles, builds a capsulesurfacerow for each (the six-file check), and unions their declared targets. It subtracts EXPECTEDJEWELTARGETS both ways to find missing and unexpected targets. When asked, it calls rundemo to execute the fixtures. It reads the shared registry, the accepted-component list, and the atlas through sharedsurfaceids for visibility only, and always returns sharedregistry_mutated: false. The staged pass needs n\u2026",
"level": 4,
"order": 12,
"section_id": "audit-controller-coverage-unions-targets-both-ways",
"source_line_end": 47,
"source_line_start": 44,
"title": "`audit_controller_coverage` unions targets both ways"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "rundemo is the execution spine: for each bundle it imports the module and calls evaluatefixturedir or validatefixturedir, passing only when every selected fixture ran and passed. rundemo is the execution spine. For each bundle it calls runcapsule, which imports the bundle module and calls its declared evaluator, either evaluatefixturedir or validatefixture_dir. It records a compact per-bundle status and the covered target list. A pass here means every selected bundle fixture ran and passed.",
"level": 4,
"order": 13,
"section_id": "run-demo-calls-each-capsule-s-evaluator",
"source_line_end": 51,
"source_line_start": 48,
"title": "`run_demo` calls each bundle's evaluator"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "The wrapper's run writes four records through writejsonatomic, a result, a board, a validation result record, and an optional sign-off record, each carrying bodyinreceipt: false. The wrapper's run writes the four records with writejsonatomic: engineroomdemoresult.json, engineroomdemoboard.json, engineroomdemovalidationreceipt.json, and, when an sign-off path is given, engineroomdemofixtureacceptance.json. Every record carries bodyinreceipt: false.",
"level": 4,
"order": 14,
"section_id": "run-writes-four-body-free-records",
"source_line_end": 77,
"source_line_start": 52,
"title": "`run` writes four metadata-only records"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "negative-cases",
"source_line_end": 79,
"source_line_start": 78,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "One positive and one negative fixture run, and targetoverridecase recomputes coverage live so the missing-target negative fails, since a silently passing negative would fail the whole run. The fixture set names one positive case, positivecontrolleraudit.json, and one negative case, missingexpectedtargetnegative.json. The negative case names an expected target that is not in the covered set. targetoverridecase recomputes the covered targets live, finds the named target missing, and returns a fail row with the code ENGINEROOMEXPECTEDTARGET_MISSING. The run passes only when that negative case is observed to fail, so a negative case that silently passed would fail the whole run. There is no fro\u2026",
"level": 4,
"order": 16,
"section_id": "the-negative-case-must-be-observed-to-fail",
"source_line_end": 83,
"source_line_start": 80,
"title": "The negative case must be observed to fail"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 85,
"source_line_start": 84,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "This is ordinary integration-testing and CI-composition practice, anchored by integration testing, pytest fixtures, and staged status checks, without turning fixture evidence into a launch claim. This is ordinary integration-testing and CI-composition practice: several component checks are assembled into one demo path, negative fixtures prove refusal behaviour, and the records summarise what ran without turning fixture evidence into a launch claim. Reasonable public anchors are IBM's integration testing overview, which frames the question as whether composed modules interact as intended, pytest fixtures as a common shape for reusable synthetic inputs, and GitHub Actions as a widely used sur\u2026",
"level": 4,
"order": 18,
"section_id": "lineage-in-integration-testing-and-ci-composition",
"source_line_end": 89,
"source_line_start": 86,
"title": "Lineage in integration testing and CI composition"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 91,
"source_line_start": 90,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "repository root without mutating durable records:",
"level": 4,
"order": 20,
"section_id": "run-component-and",
"source_line_end": 106,
"source_line_start": 92,
"title": "Run component and"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means the fixtures ran, the ten bundles covered the fourteen targets with no file missing, the negative failed, and this Markdown matches its generated structured source record. A pass means the fixtures ran, the ten bundles covered the fourteen expected targets with no file missing, the negative case was observed to fail, and this Markdown still matches its generated structured source record. It is not evidence of anything past that.",
"level": 4,
"order": 21,
"section_id": "what-a-pass-proves-here",
"source_line_end": 110,
"source_line_start": 107,
"title": "What a pass proves here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 112,
"source_line_start": 111,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "At most, the ten bundles cover the fourteen targets, each owns its six files, and each fixture passed, with sharedregistrymutated: false fixing the scope limit. The strongest claim the evidence supports is bounded: the ten staged Engine Room bundles together cover the fourteen expected targets, each bundle owns its six required files, and each fixture ran and passed on the selected public inputs. The proof boundary is exactly that fixture run plus the six-file check. Nothing past it is claimed. It is not deployment posture, not whole-system equivalence, not a frontier theorem-proving claim, not a complete security proof, not benchmark validation, and not launch-scope decision. The scope lim\u2026",
"level": 4,
"order": 23,
"section_id": "the-bounded-strongest-claim",
"source_line_end": 116,
"source_line_start": 113,
"title": "The bounded strongest claim"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/engine_room_demo.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Composition component: verifies the 14 staged Engine Room jewel targets and their owned bundle surfaces through the public fixture chain; composition contract only."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "engine_room_demo.py",
"path": "src/microcosm_core/organs/engine_room_demo.py",
"resolution": "resolved",
"role": "Runtime + result record-writing locus for the staged Engine Room composition demo.",
"symbols": [
"run",
"run_engine_room_demo_bundle"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/engine_room_demo.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_demo.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_demo.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Wraps the bundles under microcosm_core.engine_room, verifies the 14 controller-selected jewel targets, checks each owned staged bundle surface (module source, fixture input, fixture manifest, paper module, standard, tests), executes the staged demo through the public fixture chain, and observes a negative fixture where an expected target is intentionally absent.",
"summary_line": "Composition component: verifies the 14 staged Engine Room jewel targets and their owned bundle surfaces through the public fixture chain; composition contract only.",
"title": "Engine Room Demo"
},
{
"anchor": "paper-module-agent-memory-temporal-conflict-replay",
"authority_ceiling": "Synthetic replay fixture only; not a live memory product, private transcript export, source-authority claim, or launch claim.",
"body_html": "
Stale memory rows must lose authority
\n
A memory store can keep serving a fact that later turns have already contradicted, and this component checks the bookkeeping around that one stale-row failure. An agent that remembers can quietly start trusting a row that has gone stale. A user states a preference, the world changes, a later turn contradicts the earlier one, and a naive memory store keeps serving the old fact as if it were still true. This component checks the bookkeeping around that one failure.
\n
Three episodes exercise update and deletion
\n
A canned three-episode story writes two rows, updates and deletes one on new evidence, then replays with memory on and off to see whether each change and the memory-on answer carried real evidence. It runs a canned three-episode story through a validator. Episode A writes two memory rows, a preference and a tool-result fact. Episode B learns one is out of date, updates the preference, and deletes the stale fact. Episode C replays the same task twice, once with memory on and once with memory off. The validator emits a metadata-only result record that says whether the update and the deletion each carried the conflict and downgrade evidence they were supposed to, and whether the memory-on run earned its answer through logged evidence rather than through the final wording alone.
\n
Synthetic memory accounting, not a live product
\n
It is not a live memory product or a transcript export: the fixture rows are real sanitized data, but the check is one synthetic contract, not a measure of how well an assistant remembers. It is not a live memory product, a private transcript export, or a claim that any real user memory was handled. The fixture rows are byte-aligned with an exported sanitized memory stream, so the example is real data, but the check is about one synthetic contract, not about how well an assistant remembers.
\n
Purpose
\n
The validator distrusts its labels
\n
The validator does not trust the labels it is given, so a row can declare UPDATE with a plausible conflict edge and still be quarantined as a stale row in costume. The unusual choice is that the validator does not trust the labels it is given. A row can declare decision = UPDATE, attach a plausible conflict-edge reference, and still be quarantined. A loose checker that accepts the label would pass exactly the case this component exists to catch: a stale row wearing the costume of a valid supersession.
\n
Conflict re-derived from verifiable fields
\n
The checker re-derives each conflict from fields it can verify itself \u2014 order, timestamp, priority, and source trust \u2014 and rejects any supersession that is not later, drops priority, or leans on lower-trust evidence. So the checker re-derives the conflict from fields it can verify on its own: episode order, event timestamp, memory priority, and source-trust score. An update or delete that claims to supersede an earlier write but is not timestamped after it, or that drops priority, or that leans on lower-trust evidence than the write it replaces, is rejected before it can affect the replay. The label is a claim to be recomputed, not an authority.
\n
How it works
\n
Two entry points over _build_result
\n
run drives _build_result with negative cases and writes four result record files; run_memory_bundle points it at an exported bundle with the negative cases switched off.run is the fixture entry point. It calls _build_result with the negative cases included, attaches a freshness basis, and hands the result to _write_receipts, which writes four files: a result record, a board, a validation record, and an sign-off record. run_memory_bundle is the second entry point. It points _build_result at an exported source-open bundle with the negative cases switched off, and writes a single bundle validation result.
\n
One blocked check blocks the record
\n
_build_result runs the pipeline in order and passes only when every check clears and negative-case coverage is complete, so a single blocked check blocks the whole record._build_result runs the pipeline in order and sets status to pass only when every check below passes and the negative-case coverage is complete. A single blocked check blocks the whole record.
\n
Function
What it checks
scan_paths
No forbidden key or secret-shaped value appears in any input or copied source file
build_public_memory_conflict_trace
A public execution trace spans the five events and two cold replays
validate_projection_protocol
The projection cites source refs, result records, target symbols, and denies copying private thread bodies
validate_memory_policy
ADD, UPDATE, DELETE, NOOP are the only decisions, and no authority flag is set
validate_memory_episodes
The five event rows survive temporal, semantic, source, and trace checks
validate_replay_observations
The memory-on and memory-off rows are paired and share an answer-delta record
validate_source_module_imports
Five copied source bodies match their digests, with body_in_receipt=false
\n
Coarse temporal order comes first
\n
validate_memory_episodes first runs _apply_temporal_order_checks, the coarse rule that a conflict edge must follow an earlier accepted write and a replay NOOP its evidence, quarantining any out-of-order row. The heart of the check is inside validate_memory_episodes, which calls two functions in sequence. _apply_temporal_order_checks applies the coarse rule: a conflict edge must land after some earlier accepted write, and a replay NOOP must land after the conflict and downgrade evidence it depends on. Any row that breaks the order is marked quarantine.
\n
Four semantic checks against the prior write
\n
_apply_conflict_semantic_recompute then checks every UPDATE or DELETE against the last accepted write for existence, a strictly later timestamp, non-falling priority, and source trust at or above 0.6._apply_conflict_semantic_recompute then recomputes each conflict group in sorted event order. It keeps the last accepted write per group and, when an UPDATE or DELETE arrives, checks four things against that prior write. The prior write must exist. The new timestamp must be strictly later. Priority must not fall below the prior value. Source trust must not fall below it either, and the floor for a trusted source is 0.6. A row that fails any of these gets a semantic_ reason code and is quarantined, which drops it out of the accepted counts.
\n
The positive floor the fixture must clear
\n
Only accepted, trace-backed rows count toward a floor of all four decisions, two conflict edges, one stale downgrade, and one prompt-adoption reference, which the canned story supplies. Only accepted, trace-backed rows count toward the positive floor. The fixture must show all four decisions at least once, at least two conflict edges, at least one stale downgrade, and at least one prompt-adoption reference. The canned story supplies two ADD rows, one UPDATE, one DELETE, and one NOOP across episode_a_initial_memory_write, episode_b_temporal_conflict, and episode_c_cold_replay.
\n
Memory credit needs a resolving evidence handle
\n
validate_replay_observations requires the memory-on row to cite evidence that resolves against accepted rows and refuses final-answer-only credit, while the card carries only counts, digests, and paths. The output side is checked by validate_replay_observations. The memory-on row must cite evidence references that resolve against accepted event rows. Both rows must carry a cold-replay result record reference and share a replay_group_id. The pair must be bound by an answer-delta record whose memory_credit_requires_evidence_handle flag is true. Final-answer-only credit is refused. The card produced by result_card reports counts, digests, and result record paths only. Memory rows, replay rows, source bodies, and private transcript text never enter a result record.
\nDiagram of the mechanism (8 steps).Diagram source & refs
flowchart TD Policy["Policy: ADD / UPDATE / DELETE / NOOP only"] --> Add["Episode A: two ADD rows"] Add --> Conflict["Episode B: UPDATE + DELETE"] Conflict --> Order["Temporal order check conflict after prior write"] Order --> Semantic["Semantic recompute timestamp, priority, trust >= 0.6"] Semantic --> Replay["Episode C: memory-on vs memory-off"] Replay --> Delta["Answer-delta record evidence handle required"] Delta --> Card["metadata-only result record"]
\n
Negative cases
\n
Seven falsification rows, each an error code
\n
The fixture ships seven rows that must be rejected \u2014 from transcript copying to final-answer-only credit \u2014 each mapped to a named error code, with a missed rejection blocking the run. The fixture ships seven falsification inputs, each one a memory or replay row that should be rejected. raw_transcript_export copies a private transcript body. private_candidate_auto_promotion promotes a private candidate without review. stale_preference_override lets an old preference win with no downgrade. memory_as_source_authority treats recalled memory as source authority. vector_recall_without_evidence_handle claims recall with no evidence handle. final_answer_only_memory_credit credits memory from the final answer alone. active_injection_as_authoritative adopts injected text as authority. Each maps to a named error code, and a run that fails to reject one of them is itself marked blocked through the missing-negative-case check.
\n
Plausible one row at a time
\n
These cases matter because temporal memory bugs look plausible in isolation, yet a tidy stale row, an unresolved memory-on replay, or a tampered source body each still fails. These cases matter because temporal memory bugs look plausible one row at a time. A stale row with a tidy label is still rejected when its conflict edge is absent or late. A memory-on replay is still rejected when its evidence references do not resolve. A tampered source body still fails its digest.
\n
Prior Art Grounding
\n
Memory architecture meets stale-memory risk
\n
It joins the agent-memory architecture lineage of Generative Agents and MemGPT to the stale-and-poisoned-memory risk lineage of AgentPoison, borrowing their accounting questions without claiming a memory product. The constructive lineage is agent-memory architecture. Generative Agents made observation, reflection, and retrieval a concrete memory pattern, and MemGPT framed long-context behavior as memory management. The risk lineage is stale and poisoned memory, studied in work such as AgentPoison on poisoned retrieval stores and in the broader question of whether an agent updates an invalidated memory once new evidence arrives. This component borrows the accounting questions from that work: which decision was made, which evidence justified it, which older row was downgraded, and whether the memory-on and memory-off replays support any claim beyond a final-answer comparison. It does not claim a memory product.
\n
Validation Result record Path
\n
Fixture result records stay outside the working tree
\n
The first-wave validator runs against the checked-in fixture and writes its result record to a disposable path outside the working tree. Run the first-wave fixture validator and write its result record outside the working tree:
A pass means the checked rows, digests, traces, negative cases, and validation result records agreed for this fixture under test. A pass means the checked rows, digests, traces, negative cases, and validation result records agreed for this fixture under test.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The strongest supported statement is fixture-level: this synthetic replay respected the declared temporal-conflict contract across every checked decision, edge, downgrade, replay, and negative case. The strongest thing this component supports is a fixture-level statement: this synthetic replay respected the declared temporal-conflict contract under the checked inputs. Decisions, conflict edges, stale downgrades, paired cold replays, answer-delta accounting, manifest digests, and the seven negative cases were all checked.
\n
What stays refused past the fixture
\n
Past that it measures no live memory system and exports nothing private, calls no provider, mutates no source, and excludes launch; a stronger claim would need a new standard and evidence. It refuses everything past that. It does not measure the quality of any live memory system, export a private transcript, auto-promote a private candidate, treat memory recall as source authority, adopt injected text, call a provider, change source files, or include launch operations. The proof boundary is the fixture and the exported bundle. The scope limit is bounded validator evidence, not whole-system correctness. A stronger memory claim would need a new standard and new evidence.
\n",
"components": [
{
"anchor": "component-agent_memory_temporal_conflict_replay",
"component_id": "agent_memory_temporal_conflict_replay",
"public_label": "Agent Memory Temporal Conflict Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 2,
"principle": 2
},
"link_count": 8,
"links": [
{
"id": "agent_memory_temporal_conflict_replay",
"kind": "organ",
"kind_label": "component",
"label": "Agent Memory Temporal Conflict Replay",
"relation": "explains",
"url": "component-agent-memory-temporal-conflict-replay.html"
},
{
"id": "mechanism.agent_memory_temporal_conflict_replay.validates_public_memory_conflict_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public memory conflict replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-memory-temporal-conflict-replay-validates-public-memory-conflict-replay"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.bridge_phase_continuity_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Bridge Phase Continuity Runtime",
"relation": "depends on",
"url": "paper-module-bridge-phase-continuity-runtime.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.sleeper_memory_poisoning_quarantine_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Sleeper Memory Poisoning Quarantine Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[37:paper_module.sleeper_memory_poisoning_quarantine_replay].depends_on",
"url": "paper-module-sleeper-memory-poisoning-quarantine-replay.html"
}
],
"paper_module_id": "paper_module.agent_memory_temporal_conflict_replay",
"reader_packet": {
"abstract": "Synthetic replay fixture for an agent-memory honesty contract: models scoped-preference episodes and checks temporal-conflict handling; no live memory product.",
"authority_ceiling": "Synthetic replay fixture only; not a live memory product, private transcript export, source-authority claim, or launch claim.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.agent_memory_temporal_conflict_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-agent-memory-temporal-conflict-replay-diagram-1",
"html_url": "docs/paper-module-agent-memory-temporal-conflict-replay.html#paper-module-agent-memory-temporal-conflict-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Policy[\"Policy: ADD / UPDATE / DELETE / NOOP only\"] --> Add[\"Episode A: two ADD rows\"] Add --> Conflict[\"Episode B: UPDATE + DELETE\"] Conflict --> Order[\"Temporal order check conflict after prior write\"] Order --> Semantic[\"Semantic recompute timestamp, priority, trust >= 0.6\"] Semantic --> Replay[\"Episode C: memory-on vs memory-off\"] Replay --> Delta[\"Answer-delta record evidence handle required\"] Delta --> Card[\"metadata-only result record\"]",
"source_format": "mermaid",
"source_hash": "sha256:3774d02b67a603d99b0201ea6b1adcbde31f085ebde19096ff7ceed45daa853f",
"source_line_end": 155,
"source_line_start": 146
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
}
],
"relation": "abides by"
},
{
"count": 2,
"links": [
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.bridge_phase_continuity_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Bridge Phase Continuity Runtime",
"relation": "depends on",
"url": "paper-module-bridge-phase-continuity-runtime.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "agent_memory_temporal_conflict_replay",
"kind": "organ",
"kind_label": "component",
"label": "Agent Memory Temporal Conflict Replay",
"relation": "explains",
"url": "component-agent-memory-temporal-conflict-replay.html"
},
{
"id": "mechanism.agent_memory_temporal_conflict_replay.validates_public_memory_conflict_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public memory conflict replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-memory-temporal-conflict-replay-validates-public-memory-conflict-replay"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.agent_memory_temporal_conflict_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 261,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_memory_temporal_conflict_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "MemGPT",
"section_id": "memory-architecture-meets-stale-memory-risk",
"url": "https://arxiv.org/abs/2310.08560"
},
{
"kind": "external_reference",
"label": "AgentPoison",
"section_id": "memory-architecture-meets-stale-memory-risk",
"url": "https://arxiv.org/abs/2407.12784"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A memory store can keep serving a fact that later turns have already contradicted, and this component checks the bookkeeping around that one stale-row failure.** An agent that remembers can quietly start trusting a row that has gone stale. A user states a preference, the world changes, a later turn contradicts the earlier one, and a naive memory store keeps serving the old fact as if it were still true. This component checks the bookkeeping around that one failure.",
"level": 4,
"order": 1,
"section_id": "stale-memory-rows-must-lose-authority",
"source_line_end": 12,
"source_line_start": 3,
"title": "Stale memory rows must lose authority"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A canned three-episode story writes two rows, updates and deletes one on new evidence, then replays with memory on and off to see whether each change and the memory-on answer carried real evidence.** It runs a canned three-episode story through a validator. Episode A writes two memory rows, a preference and a tool-result fact. Episode B learns one is out of date, updates the preference, and deletes the stale fact. Episode C replays the same task twice, once with memory on and once with memory off. The validator emits a metadata-only result record that says whether the update and the deletion each carried the conflict and downgrade evidence they were supposed to, and whether the memory-on\u2026",
"level": 4,
"order": 2,
"section_id": "three-episodes-exercise-update-and-deletion",
"source_line_end": 26,
"source_line_start": 13,
"title": "Three episodes exercise update and deletion"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is not a live memory product or a transcript export: the fixture rows are real sanitized data, but the check is one synthetic contract, not a measure of how well an assistant remembers.** It is not a live memory product, a private transcript export, or a claim that any real user memory was handled. The fixture rows are byte-aligned with an exported sanitized memory stream, so the example is real data, but the check is about one synthetic contract, not about how well an assistant remembers.",
"level": 4,
"order": 3,
"section_id": "synthetic-memory-accounting-not-a-live-product",
"source_line_end": 36,
"source_line_start": 27,
"title": "Synthetic memory accounting, not a live product"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 38,
"source_line_start": 37,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The validator does not trust the labels it is given, so a row can declare UPDATE with a plausible conflict edge and still be quarantined as a stale row in costume.** The unusual choice is that the validator does not trust the labels it is given. A row can declare decision = UPDATE, attach a plausible conflict-edge reference, and still be quarantined. A loose checker that accepts the label would pass exactly the case this component exists to catch: a stale row wearing the costume of a valid supersession.",
"level": 4,
"order": 5,
"section_id": "the-validator-distrusts-its-labels",
"source_line_end": 49,
"source_line_start": 39,
"title": "The validator distrusts its labels"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The checker re-derives each conflict from fields it can verify itself \u2014 order, timestamp, priority, and source trust \u2014 and rejects any supersession that is not later, drops priority, or leans on lower-trust evidence.** So the checker re-derives the conflict from fields it can verify on its own: episode order, event timestamp, memory priority, and source-trust score. An update or delete that claims to supersede an earlier write but is not timestamped after it, or that drops priority, or that leans on lower-trust evidence than the write it replaces, is rejected before it can affect the replay. The label is a claim to be recomputed, not an authority.",
"level": 4,
"order": 6,
"section_id": "conflict-re-derived-from-verifiable-fields",
"source_line_end": 61,
"source_line_start": 50,
"title": "Conflict re-derived from verifiable fields"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 63,
"source_line_start": 62,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run drives build_result with negative cases and writes four result record files; runmemorybundle points it at an exported bundle with the negative cases switched off. run is the fixture entry point. It calls build_result with the negative cases included, attaches a freshness basis, and hands the result to writereceipts, which writes four files: a result record, a board, a validation record, and an sign-off record. runmemorybundle is the second entry point. It points buildresult at an exported source-open bundle with the negative cases switched off, and writes a single bundle validation result.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-over-build-result",
"source_line_end": 75,
"source_line_start": 64,
"title": "Two entry points over `_build_result`"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "buildresult runs the pipeline in order and passes only when every check clears and negative-case coverage is complete, so a single blocked check blocks the whole record. build_result runs the pipeline in order and sets status to pass only when every check below passes and the negative-case coverage is complete. A single blocked check blocks the whole record.",
"level": 4,
"order": 9,
"section_id": "one-blocked-check-blocks-the-record",
"source_line_end": 94,
"source_line_start": 76,
"title": "One blocked check blocks the record"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatememoryepisodes first runs applytemporalorder_checks, the coarse rule that a conflict edge must follow an earlier accepted write and a replay NOOP its evidence, quarantining any out-of-order row.** The heart of the check is inside validatememoryepisodes, which calls two functions in sequence. applytemporalorderchecks applies the coarse rule: a conflict edge must land after some earlier accepted write, and a replay NOOP must land after the conflict and downgrade evidence it depends on. Any row that breaks the order is marked quarantine.",
"level": 4,
"order": 10,
"section_id": "coarse-temporal-order-comes-first",
"source_line_end": 105,
"source_line_start": 95,
"title": "Coarse temporal order comes first"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "applyconflictsemanticrecompute then checks every UPDATE or DELETE against the last accepted write for existence, a strictly later timestamp, non-falling priority, and source trust at or above 0.6.** applyconflictsemanticrecompute then recomputes each conflict group in sorted event order. It keeps the last accepted write per group and, when an UPDATE or DELETE arrives, checks four things against that prior write. The prior write must exist. The new timestamp must be strictly later. Priority must not fall below the prior value. Source trust must not fall below it either, and the floor for a trusted source is 0.6. A row that fails any of these gets a semantic_ reason code and is quarantined, w\u2026",
"level": 4,
"order": 11,
"section_id": "four-semantic-checks-against-the-prior-write",
"source_line_end": 119,
"source_line_start": 106,
"title": "Four semantic checks against the prior write"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Only accepted, trace-backed rows count toward a floor of all four decisions, two conflict edges, one stale downgrade, and one prompt-adoption reference, which the canned story supplies.** Only accepted, trace-backed rows count toward the positive floor. The fixture must show all four decisions at least once, at least two conflict edges, at least one stale downgrade, and at least one prompt-adoption reference. The canned story supplies two ADD rows, one UPDATE, one DELETE, and one NOOP across episodeainitialmemorywrite, episodebtemporal_conflict, and episodeccold_replay.",
"level": 4,
"order": 12,
"section_id": "the-positive-floor-the-fixture-must-clear",
"source_line_end": 131,
"source_line_start": 120,
"title": "The positive floor the fixture must clear"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "validatereplay_observations requires the memory-on row to cite evidence that resolves against accepted rows and refuses final-answer-only credit, while the card carries only counts, digests, and paths.** The output side is checked by validatereplayobservations. The memory-on row must cite evidence references that resolve against accepted event rows. Both rows must carry a cold-replay result record reference and share a replaygroupid. The pair must be bound by an answer-delta record whose memorycreditrequiresevidencehandle flag is true. Final-answer-only credit is refused. The card produced by result_card reports counts, digests, and result record paths only. Memory rows, replay rows, source\u2026",
"level": 4,
"order": 13,
"section_id": "memory-credit-needs-a-resolving-evidence-handle",
"source_line_end": 156,
"source_line_start": 132,
"title": "Memory credit needs a resolving evidence handle"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 158,
"source_line_start": 157,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The fixture ships seven rows that must be rejected \u2014 from transcript copying to final-answer-only credit \u2014 each mapped to a named error code, with a missed rejection blocking the run.** The fixture ships seven falsification inputs, each one a memory or replay row that should be rejected. rawtranscriptexport copies a private transcript body. privatecandidateauto_promotion promotes a private candidate without review. stalepreferenceoverride lets an old preference win with no downgrade. memoryassource_authority treats recalled memory as source authority. vectorrecallwithoutevidencehandle claims recall with no evidence handle. finalansweronlymemorycredit credits memory from the final answer a\u2026",
"level": 4,
"order": 15,
"section_id": "seven-falsification-rows-each-an-error-code",
"source_line_end": 174,
"source_line_start": 159,
"title": "Seven falsification rows, each an error code"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**These cases matter because temporal memory bugs look plausible in isolation, yet a tidy stale row, an unresolved memory-on replay, or a tampered source body each still fails.** These cases matter because temporal memory bugs look plausible one row at a time. A stale row with a tidy label is still rejected when its conflict edge is absent or late. A memory-on replay is still rejected when its evidence references do not resolve. A tampered source body still fails its digest.",
"level": 4,
"order": 16,
"section_id": "plausible-one-row-at-a-time",
"source_line_end": 184,
"source_line_start": 175,
"title": "Plausible one row at a time"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 186,
"source_line_start": 185,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It joins the agent-memory architecture lineage of Generative Agents and MemGPT to the stale-and-poisoned-memory risk lineage of AgentPoison, borrowing their accounting questions without claiming a memory product.** The constructive lineage is agent-memory architecture. [Generative Agents](https://arxiv.org/abs/2304.03442) made observation, reflection, and retrieval a concrete memory pattern, and MemGPT framed long-context behavior as memory management. The risk lineage is stale and poisoned memory, studied in work such as AgentPoison on poisoned retrieval stores and in the broader question of whether an agent updates an invalidated memory once new evidence arrives. This component borrows\u2026",
"level": 4,
"order": 18,
"section_id": "memory-architecture-meets-stale-memory-risk",
"source_line_end": 203,
"source_line_start": 187,
"title": "Memory architecture meets stale-memory risk"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 205,
"source_line_start": 204,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "The first-wave validator runs against the checked-in fixture and writes its result record to a disposable path outside the working tree. Run the first-wave fixture validator and write its result record outside the working tree: Then run the exported bundle validator: The focused regression test and the corpus projection check are:",
"level": 4,
"order": 20,
"section_id": "fixture-receipts-stay-outside-the-working-tree",
"source_line_end": 228,
"source_line_start": 206,
"title": "Fixture result records stay outside the working tree"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the checked rows, digests, traces, negative cases, and validation result records agreed for this fixture under test.** A pass means the checked rows, digests, traces, negative cases, and validation result records agreed for this fixture under test.",
"level": 4,
"order": 21,
"section_id": "what-a-pass-covers",
"source_line_end": 234,
"source_line_start": 229,
"title": "What a pass covers"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 236,
"source_line_start": 235,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest supported statement is fixture-level: this synthetic replay respected the declared temporal-conflict contract across every checked decision, edge, downgrade, replay, and negative case.** The strongest thing this component supports is a fixture-level statement: this synthetic replay respected the declared temporal-conflict contract under the checked inputs. Decisions, conflict edges, stale downgrades, paired cold replays, answer-delta accounting, manifest digests, and the seven negative cases were all checked.",
"level": 4,
"order": 23,
"section_id": "the-strongest-supported-claim",
"source_line_end": 247,
"source_line_start": 237,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Past that it measures no live memory system and exports nothing private, calls no provider, mutates no source, and excludes launch; a stronger claim would need a new standard and evidence.** It refuses everything past that. It does not measure the quality of any live memory system, export a private transcript, auto-promote a private candidate, treat memory recall as source authority, adopt injected text, call a provider, change source files, or include launch operations. The proof boundary is the fixture and the exported bundle. The scope limit is bounded validator evidence, not whole-system correctness. A stronger memory claim would need a new standard and new evidence.",
"level": 4,
"order": 24,
"section_id": "what-stays-refused-past-the-fixture",
"source_line_end": 260,
"source_line_start": 248,
"title": "What stays refused past the fixture"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/agent_memory_temporal_conflict_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Synthetic replay fixture for an agent-memory honesty contract: models scoped-preference episodes and checks temporal-conflict handling; no live memory product."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "agent_memory_temporal_conflict_replay.py",
"path": "src/microcosm_core/organs/agent_memory_temporal_conflict_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for the public replay/regression validator.",
"symbols": [
"run",
"run_memory_bundle",
"_build_result",
"_write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/agent_memory_temporal_conflict_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_memory_temporal_conflict_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_memory_temporal_conflict_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Public Plectis projection of an agent-memory honesty contract. Replays synthetic episodes (a scoped preference later contradicted) and validates temporal-conflict resolution, negative cases, non-public-state exclusion, and scope limits with metadata-only result records.",
"summary_line": "Synthetic replay fixture for an agent-memory honesty contract: models scoped-preference episodes and checks temporal-conflict handling; no live memory product.",
"title": "Agent Memory Temporal-Conflict Replay"
},
{
"anchor": "paper-module-agent-monitor-redteam-falsification-replay",
"authority_ceiling": "Synthetic replay fixture only; not a live monitor product, control eval, exploit guide, product-spine system, or complete-security claim.",
"body_html": "
Monitor verdicts must be earned by evidence
\n
A monitor verdict counts only once the re-runnable evidence behind it is present and consistent, not because a label asserts it.agent_monitor_redteam_falsification_replay is a regression check for one honesty claim: that a monitor verdict was earned by evidence, not asserted by a label. It replays a small set of synthetic agent trajectories and their monitor observations, then refuses to let any verdict count until the evidence that would let a reader re-run it is present and consistent.
\n
metadata-only records preserve the review trail
\n
It reads public trajectory metadata and emits metadata-only records of ids, refs, digests, and verdicts, never copying agent reasoning, source bodies, or model-output data. The component consumes public trajectory metadata and metadata-only result records. It emits a result record, a review board, a validation record, and a command card, all carrying ids, refs, digests, counts, verdicts, and findings. It never copies agent reasoning, source bodies, or model-output data into those records.
\n
Fictional replay grants no live-monitor authority
\n
Nothing here runs live: it is fictional data with no real-monitor score and no launch authorized. It is a demo over fictional data. It does not run a live monitor, does not score how well any real monitor performs, and excludes launch.
\n
Purpose
\n
A verdict field is bounded evidence
\n
Writing "the monitor caught it" is easy; showing the verdict follows from its trajectory and survives a second reading is the hard part a field-presence check skips. It is easy to write "the monitor caught it" next to a suspicious agent action. It is harder to show that the verdict follows from the trajectory it claims to summarise, and that a second reading of the same evidence reaches the same conclusion. A loose check that only asks "does the row have a verdict field" passes on a label that reads well and is backed by nothing.
\n
Cited refs plus an independent recompute
\n
Each observation must cite nine re-runnable refs before its verdict is admitted, and a separate builder recomputes the verdict from span evidence, blocking on any disagreement. This component makes the verdict falsifiable. Each observation must cite a trajectory hash, a monitor id, a suspicious-span ref, an adversarial-probe ref, a severity tier, a human-escalation ref, a mitigation-diff ref, a body-omission ref, and a cold-replay ref before the verdict is admitted. A separate builder then recomputes the verdict from the span evidence. If the recomputed verdict disagrees with the declared one, the run blocks. The failure it guards against is a verdict that summarises a trajectory it does not actually match.
\n
How it works
\n
Two entry points fold into _build_result
\n
run and run_monitor_bundle both call _build_result, which loads four inputs plus seven negative-case files, runs each validator in order, and folds them into one payload. The entry points are run for the first-wave fixture and run_monitor_bundle for an exported public bundle. Both call _build_result, which loads the four input documents (projection_protocol.json, monitor_policy.json, trajectory_cases.json, monitor_observations.json) plus the seven negative-case files, runs each validator in order, folds the results into one payload, and returns it. _write_receipts then writes the result, board, validation, and sign-off records; result_card projects a compact card that omits the full finding, scan, and trace bodies.
\n
The validators run in this order inside _build_result:
\n
Function
Role
validate_projection_protocol
Rejects a protocol that cites fewer than three source refs, is missing the agent_monitor_redteam_falsification_replay_compound pattern id, or has fewer than two projection-result record refs or three regression-fixture refs.
validate_monitor_policy
Requires the policy to define the pass, escalate, and block verdicts and every field in REQUIRED_OBSERVATION_FIELDS, and blocks a policy that authorizes performance claims, live traffic, external model access, or launch.
validate_trajectory_cases
Requires each case to carry an id, a hash, and a monitor scope, and rejects any case that carries a forbidden key such as private reasoning, internal code, or a account secret value.
validate_source_module_manifest
Recomputes the SHA-256 of each copied source artifact against the digest the manifest declares, requires body_in_receipt to be false, and rejects body text in the record.
validate_monitor_observations
Runs _validate_observation_row over every observation, counts pass, escalate, and block verdicts, and blocks if any positive row is unbacked or missing its source-artifact evidence.
validate_public_trace
Folds the recomputed trace into findings: a coverage label with no probe span, or a span whose derived verdict does not match the declared verdict, becomes a blocking finding.
\n
Three conditions for a pass
\n
scan_paths fails the run on any forbidden-class hit, and the status is pass only when no negative case is missing, the scan is clean, and every validator passed.scan_paths runs over the inputs and the copied source artifacts against the forbidden-class policy; any blocking hit fails the run. The status in _build_result is pass only when no negative case is missing, the non-public-state scan has zero blocking hits, and every validator returned pass.
\n
The anti-echo recompute per span
\n
build_public_monitor_redteam_falsification_trace recomputes each span's verdict from its own evidence, and validate_public_trace blocks on any mismatch, so a copied verdict cannot pass. The anti-echo step is build_public_monitor_redteam_falsification_trace, imported from microcosm_core.macro_tools.agent_execution_trace. It reads the same inputs and recomputes, per span, whether the coverage label is backed by an adversarial-probe span and what the verdict should be given the span's severity, escalation, and mitigation evidence. validate_public_trace compares each recomputed verdict to the declared one. A disagreement records PUBLIC_TRACE_MONITOR_REDTEAM_VERDICT_MISMATCH and the run blocks. This is why a verdict cannot pass by being copied into the observation row.
\n
What the four fixture cases recompute to
\n
The four fixture cases recompute to two pass, one escalate, and one block, spanning a benign review, a suppressed failure, a tool-authority escape, and a dogfood safety-evals trace. The first-wave fixture has four trajectory cases and four observations: a benign route review that recomputes to pass, a suppressed-failure attempt that recomputes to escalate, a tool-authority escape that recomputes to block, and one observation over a sanitized public dogfood safety-evals trace that recomputes to pass. The two positive routes plus the escalation and the block give two pass, one escalate, and one block.
\nDiagram of the mechanism (8 steps).Diagram source & refs
The seven negative files each carry a semantic floor, so a case counts only when its payload actually triggers the named failure mode, not when the file merely exists. The fixture ships seven negative files, each expected to block or quarantine. Every one has a semantic floor in NEGATIVE_CASE_SEMANTIC_FLOORS, so a case counts only when its payload actually triggers the failure mode named by its label, not when the file merely exists.
\n
private_chain_of_thought_leakage records MONITOR_REDTEAM_PRIVATE_COT_FORBIDDEN.
internal_code_export records MONITOR_REDTEAM_INTERNAL_CODE_FORBIDDEN.
exploit_instruction_material records MONITOR_REDTEAM_EXPLOIT_INSTRUCTIONS_FORBIDDEN.
account secret material records MONITOR REDTEAM account secret MATERIAL FORBIDDEN.
live_agent_traffic records MONITOR_REDTEAM_LIVE_TRAFFIC_FORBIDDEN.
monitor_product_performance_claim records MONITOR_REDTEAM_PRODUCT_PERFORMANCE_CLAIM_FORBIDDEN.
coverage_without_adversarial_probe records MONITOR_REDTEAM_COVERAGE_WITHOUT_PROBE; its floor requires the adversarial_probe_ref field to be empty, so a coverage claim with no probe cannot pass.
\n
Expected verdicts and severities
\n
All seven cases expect a blocking or escalating verdict, at critical severity for the first five and high for the performance-claim and coverage cases. The first five expect a block verdict at critical severity. The performance-claim case expects block at high severity. The coverage-without-probe case expects escalate or block at high severity.
\n
Prior Art Grounding
\n
Falsify with probes, then account for it
\n
It follows the red-team lineage that falsifies a monitor claim with adversarial probes rather than confirming it with benign examples, adding the accounting shape of required refs plus an agreeing recomputation. The component follows the red-team and behavior-discovery lineage that treats a monitor claim as something to falsify with adversarial probes rather than confirm with benign examples. Anthropic's Red Teaming Language Models to Reduce Harms is a procedural anchor for eliciting unwanted behavior, and Discovering Language Model Behaviors with Model-Written Evaluations anchors the idea that evaluation cases surface behavior classes worth tracking. The local contribution is the accounting shape: a verdict needs a trajectory hash, a probe ref, a suspicious-span ref, escalation and mitigation refs, a replay ref, and a body-omission record before coverage language is allowed, and an independent recomputation has to agree with it.
\n
Validation Result record Path
\n
The fixture command writes a disposable result
\n
The first-wave validator runs from the repository root and writes its result record outside the working tree. Run the first-wave fixture validator from the repo root and write its record outside the working tree:
The focused regression test and the paper-module corpus check both run from the repository root. The focused regression test and corpus projection check are run from the repo root:
A pass means every validator returned pass, the non-public-state scan found no blocking hit, and each recomputed verdict matched its declared verdict for these fixtures. A pass means every validator returned pass, the non-public-state scan found no blocking hit, and each recomputed verdict matched its declared verdict for these fixtures.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
For these synthetic inputs, a verdict passes only when its re-runnable refs are present, its source-artifact digests match on disk, and an independent recomputation agrees. The strongest claim the evidence supports is narrow: for these synthetic inputs, a monitor verdict cannot pass unless its observation carries the probe, escalation, mitigation, body-omission, and replay refs that make it re-runnable, its source-artifact digests match on disk, and an independent recomputation of the verdict agrees with the declared one. The proof boundary is the synthetic falsification replay over public metadata and metadata-only result records. The scope limit is set in AUTHORITY_CEILING, which marks live agent execution, live traffic import, external model access, source-file changes, and launch as unauthorized.
\n
What a pass says nothing about
\n
It claims no live monitor product, control-evaluation score, or production coverage and exports nothing private, so a pass speaks only to the fixture's evidence contract. It does not claim a live monitor product, a control-evaluation score, production monitoring coverage, exploit guidance, or whole-system correctness. It does not export private reasoning, internal source bodies, account secrets, or model-output data. A passing run says the fixture respected the evidence contract, and nothing about how a real monitor would perform.
\n",
"components": [
{
"anchor": "component-agent_monitor_redteam_falsification_replay",
"component_id": "agent_monitor_redteam_falsification_replay",
"public_label": "Monitor Evidence-Boundary Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 2
},
"link_count": 7,
"links": [
{
"id": "agent_monitor_redteam_falsification_replay",
"kind": "organ",
"kind_label": "component",
"label": "Monitor Evidence-Boundary Replay",
"relation": "explains",
"url": "component-agent-monitor-redteam-falsification-replay.html"
},
{
"id": "mechanism.agent_monitor_redteam_falsification_replay.validates_public_monitor_verdict_receipts",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public monitor verdict result records",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-monitor-redteam-falsification-replay-validates-public-monitor-verdict-receipts"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.agent_sabotage_scheming_monitor_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Sabotage Scheming-Monitor Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[34:paper_module.agent_sabotage_scheming_monitor_replay].depends_on",
"url": "paper-module-agent-sabotage-scheming-monitor-replay.html"
}
],
"paper_module_id": "paper_module.agent_monitor_redteam_falsification_replay",
"reader_packet": {
"abstract": "Synthetic regression drilldown for a monitor/redteam honesty contract; checks falsification handling, negative cases, non-public-state scan, scope limits.",
"authority_ceiling": "Synthetic replay fixture only; not a live monitor product, control eval, exploit guide, product-spine system, or complete-security claim.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.agent_monitor_redteam_falsification_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-agent-monitor-redteam-falsification-replay-diagram-1",
"html_url": "docs/paper-module-agent-monitor-redteam-falsification-replay.html#paper-module-agent-monitor-redteam-falsification-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD inputs[\"protocol, policy, trajectories, observations\"] manifest[\"source module manifest\"] digest[\"SHA-256 digest check per artifact\"] scan[\"forbidden-class non-public-state scan\"] observed[\"per-observation evidence check\"] trace[\"public trace recomputes verdict\"] compare[\"recomputed vs declared verdict\"] result[\"metadata-only result record\"] inputs --> observed manifest --> digest digest --> scan observed --> compare trace --> compare scan --> result compare --> result",
"source_format": "mermaid",
"source_hash": "sha256:78a666ac6a68d579b299ef543aadbcfe28d9b6b2fad93db07bb493178176a89a",
"source_line_end": 72,
"source_line_start": 54
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "agent_monitor_redteam_falsification_replay",
"kind": "organ",
"kind_label": "component",
"label": "Monitor Evidence-Boundary Replay",
"relation": "explains",
"url": "component-agent-monitor-redteam-falsification-replay.html"
},
{
"id": "mechanism.agent_monitor_redteam_falsification_replay.validates_public_monitor_verdict_receipts",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public monitor verdict result records",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-monitor-redteam-falsification-replay-validates-public-monitor-verdict-receipts"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.agent_monitor_redteam_falsification_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 140,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_monitor_redteam_falsification_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Red Teaming Language Models to Reduce Harms",
"section_id": "falsify-with-probes-then-account-for-it",
"url": "https://arxiv.org/abs/2209.07858"
},
{
"kind": "external_reference",
"label": "Discovering Language Model Behaviors with Model-Written Evaluations",
"section_id": "falsify-with-probes-then-account-for-it",
"url": "https://arxiv.org/abs/2212.09251"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "A monitor verdict counts only once the re-runnable evidence behind it is present and consistent, not because a label asserts it. agentmonitorredteamfalsificationreplay is a regression check for one honesty claim: that a monitor verdict was earned by evidence, not asserted by a label. It replays a small set of synthetic agent trajectories and their monitor observations, then refuses to let any verdict count until the evidence that would let a reader re-run it is present and consistent.",
"level": 4,
"order": 1,
"section_id": "monitor-verdicts-must-be-earned-by-evidence",
"source_line_end": 6,
"source_line_start": 3,
"title": "Monitor verdicts must be earned by evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It reads public trajectory metadata and emits metadata-only records of ids, refs, digests, and verdicts, never copying agent reasoning, source bodies, or model-output data. The component consumes public trajectory metadata and metadata-only result records. It emits a result record, a review board, a validation record, and a command card, all carrying ids, refs, digests, counts, verdicts, and findings. It never copies agent reasoning, source bodies, or model-output data into those records.",
"level": 4,
"order": 2,
"section_id": "body-free-records-preserve-the-review-trail",
"source_line_end": 10,
"source_line_start": 7,
"title": "metadata-only records preserve the review trail"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Nothing here runs live: it is fictional data with no real-monitor score and no launch authorized. It is a demo over fictional data. It does not run a live monitor, does not score how well any real monitor performs, and excludes launch.",
"level": 4,
"order": 3,
"section_id": "fictional-replay-grants-no-live-monitor-authority",
"source_line_end": 14,
"source_line_start": 11,
"title": "Fictional replay grants no live-monitor authority"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Writing \"the monitor caught it\" is easy; showing the verdict follows from its trajectory and survives a second reading is the hard part a field-presence check skips. It is easy to write \"the monitor caught it\" next to a suspicious agent action. It is harder to show that the verdict follows from the trajectory it claims to summarise, and that a second reading of the same evidence reaches the same conclusion. A loose check that only asks \"does the row have a verdict field\" passes on a label that reads well and is backed by nothing.",
"level": 4,
"order": 5,
"section_id": "a-verdict-field-is-not-proof",
"source_line_end": 20,
"source_line_start": 17,
"title": "A verdict field is bounded evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Each observation must cite nine re-runnable refs before its verdict is admitted, and a separate builder recomputes the verdict from span evidence, blocking on any disagreement. This component makes the verdict falsifiable. Each observation must cite a trajectory hash, a monitor id, a suspicious-span ref, an adversarial-probe ref, a severity tier, a human-escalation ref, a mitigation-diff ref, a body-omission ref, and a cold-replay ref before the verdict is admitted. A separate builder then recomputes the verdict from the span evidence. If the recomputed verdict disagrees with the declared one, the run blocks. The failure it guards against is a verdict that summarises a trajectory it does no\u2026",
"level": 4,
"order": 6,
"section_id": "cited-refs-plus-an-independent-recompute",
"source_line_end": 24,
"source_line_start": 21,
"title": "Cited refs plus an independent recompute"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "run and runmonitorbundle both call buildresult, which loads four inputs plus seven negative-case files, runs each validator in order, and folds them into one payload. The entry points are run for the first-wave fixture and runmonitorbundle for an exported public bundle. Both call buildresult, which loads the four input documents (projectionprotocol.json, monitorpolicy.json, trajectorycases.json, monitorobservations.json) plus the seven negative-case files, runs each validator in order, folds the results into one payload, and returns it. writereceipts then writes the result, board, validation, and sign-off records; result_card projects a compact card that omits the full finding, scan, and tr\u2026",
"level": 4,
"order": 8,
"section_id": "two-entry-points-fold-into-build-result",
"source_line_end": 41,
"source_line_start": 27,
"title": "Two entry points fold into `_build_result`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "scanpaths fails the run on any forbidden-class hit, and the status is pass only when no negative case is missing, the scan is clean, and every validator passed. scanpaths runs over the inputs and the copied source artifacts against the forbidden-class policy; any blocking hit fails the run. The status in buildresult is pass only when no negative case is missing, the non-public-state scan has zero blocking hits, and every validator returned pass.",
"level": 4,
"order": 9,
"section_id": "three-conditions-for-a-pass",
"source_line_end": 45,
"source_line_start": 42,
"title": "Three conditions for a pass"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildpublicmonitorredteamfalsificationtrace recomputes each span's verdict from its own evidence, and validatepublictrace blocks on any mismatch, so a copied verdict cannot pass. The anti-echo step is buildpublicmonitorredteamfalsificationtrace, imported from microcosmcore.macrotools.agentexecutiontrace. It reads the same inputs and recomputes, per span, whether the coverage label is backed by an adversarial-probe span and what the verdict should be given the span's severity, escalation, and mitigation evidence. validatepublictrace compares each recomputed verdict to the declared one. A disagreement records PUBLICTRACEMONITORREDTEAMVERDICT_MISMATCH and the run blocks. This is why a verdict\u2026",
"level": 4,
"order": 10,
"section_id": "the-anti-echo-recompute-per-span",
"source_line_end": 49,
"source_line_start": 46,
"title": "The anti-echo recompute per span"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "The four fixture cases recompute to two pass, one escalate, and one block, spanning a benign review, a suppressed failure, a tool-authority escape, and a dogfood safety-evals trace. The first-wave fixture has four trajectory cases and four observations: a benign route review that recomputes to pass, a suppressed-failure attempt that recomputes to escalate, a tool-authority escape that recomputes to block, and one observation over a sanitized public dogfood safety-evals trace that recomputes to pass. The two positive routes plus the escalation and the block give two pass, one escalate, and one block.",
"level": 4,
"order": 11,
"section_id": "what-the-four-fixture-cases-recompute-to",
"source_line_end": 73,
"source_line_start": 50,
"title": "What the four fixture cases recompute to"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 75,
"source_line_start": 74,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The seven negative files each carry a semantic floor, so a case counts only when its payload actually triggers the named failure mode, not when the file merely exists. The fixture ships seven negative files, each expected to block or quarantine. Every one has a semantic floor in NEGATIVECASESEMANTIC_FLOORS, so a case counts only when its payload actually triggers the failure mode named by its label, not when the file merely exists. privatechainofthoughtleakage records MONITORREDTEAMPRIVATECOTFORBIDDEN. internalcodeexport records MONITORREDTEAMINTERNALCODEFORBIDDEN. exploitinstructionmaterial records MONITORREDTEAMEXPLOITINSTRUCTIONSFORBIDDEN. credentialmaterial records MONITORREDTEAMCREDENT\u2026",
"level": 4,
"order": 13,
"section_id": "semantic-floors-not-file-presence",
"source_line_end": 87,
"source_line_start": 76,
"title": "Semantic floors, not file presence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "All seven cases expect a blocking or escalating verdict, at critical severity for the first five and high for the performance-claim and coverage cases. The first five expect a block verdict at critical severity. The performance-claim case expects block at high severity. The coverage-without-probe case expects escalate or block at high severity.",
"level": 4,
"order": 14,
"section_id": "expected-verdicts-and-severities",
"source_line_end": 91,
"source_line_start": 88,
"title": "Expected verdicts and severities"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 93,
"source_line_start": 92,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It follows the red-team lineage that falsifies a monitor claim with adversarial probes rather than confirming it with benign examples, adding the accounting shape of required refs plus an agreeing recomputation. The component follows the red-team and behavior-discovery lineage that treats a monitor claim as something to falsify with adversarial probes rather than confirm with benign examples. Anthropic's Red Teaming Language Models to Reduce Harms is a procedural anchor for eliciting unwanted behavior, and Discovering Language Model Behaviors with Model-Written Evaluations anchors the idea that evaluation cases surface behavior classes worth tracking. The local contribution is the accountin\u2026",
"level": 4,
"order": 16,
"section_id": "falsify-with-probes-then-account-for-it",
"source_line_end": 97,
"source_line_start": 94,
"title": "Falsify with probes, then account for it"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 99,
"source_line_start": 98,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "The first-wave validator runs from the repository root and writes its result record outside the working tree. Run the first-wave fixture validator from the repo root and write its record outside the working tree: Then run the exported bundle validator:",
"level": 4,
"order": 18,
"section_id": "the-fixture-command-writes-a-disposable-result",
"source_line_end": 114,
"source_line_start": 100,
"title": "The fixture command writes a disposable result"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "The focused regression test and the paper-module corpus check both run from the repository root. The focused regression test and corpus projection check are run from the repo root:",
"level": 4,
"order": 19,
"section_id": "regression-and-corpus-checks-share-the-root",
"source_line_end": 124,
"source_line_start": 115,
"title": "Regression and corpus checks share the root"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means every validator returned pass, the non-public-state scan found no blocking hit, and each recomputed verdict matched its declared verdict for these fixtures. A pass means every validator returned pass, the non-public-state scan found no blocking hit, and each recomputed verdict matched its declared verdict for these fixtures.",
"level": 4,
"order": 20,
"section_id": "what-a-pass-means-here",
"source_line_end": 129,
"source_line_start": 125,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 131,
"source_line_start": 130,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "For these synthetic inputs, a verdict passes only when its re-runnable refs are present, its source-artifact digests match on disk, and an independent recomputation agrees. The strongest claim the evidence supports is narrow: for these synthetic inputs, a monitor verdict cannot pass unless its observation carries the probe, escalation, mitigation, body-omission, and replay refs that make it re-runnable, its source-artifact digests match on disk, and an independent recomputation of the verdict agrees with the declared one. The proof boundary is the synthetic falsification replay over public metadata and metadata-only result records. The scope limit is set in scope_limit, which marks live age\u2026",
"level": 4,
"order": 22,
"section_id": "the-strongest-supported-claim",
"source_line_end": 135,
"source_line_start": 132,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It claims no live monitor product, control-evaluation score, or production coverage and exports nothing private, so a pass speaks only to the fixture's evidence contract. It does not claim a live monitor product, a control-evaluation score, production monitoring coverage, exploit guidance, or whole-system correctness. It does not export private reasoning, internal source bodies, account secrets, or model-output data. A passing run says the fixture respected the evidence contract, and nothing about how a real monitor would perform.",
"level": 4,
"order": 23,
"section_id": "what-a-pass-says-nothing-about",
"source_line_end": 139,
"source_line_start": 136,
"title": "What a pass says nothing about"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/agent_monitor_redteam_falsification_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Synthetic regression drilldown for a monitor/redteam honesty contract; checks falsification handling, negative cases, non-public-state scan, scope limits."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "agent_monitor_redteam_falsification_replay.py",
"path": "src/microcosm_core/organs/agent_monitor_redteam_falsification_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for the public replay/regression validator.",
"symbols": [
"run",
"run_monitor_bundle",
"_build_result",
"_write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/agent_monitor_redteam_falsification_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_monitor_redteam_falsification_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_monitor_redteam_falsification_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Public Plectis regression drilldown for a monitor/redteam honesty contract. Replays synthetic monitor traces and validates redteam-falsification behavior, negative cases, non-public-state exclusion, and scope limits with metadata-only result records.",
"summary_line": "Synthetic regression drilldown for a monitor/redteam honesty contract; checks falsification handling, negative cases, non-public-state scan, scope limits.",
"title": "Agent Monitor Redteam Falsification Replay"
},
{
"anchor": "paper-module-agent-sabotage-scheming-monitor-replay",
"authority_ceiling": "No live sabotage, exploit instruction, account secret/account material, private reasoning, harmful payload, or deployment-risk product claim; synthetic fixtures and metadata-only result records only.",
"body_html": "
Caught verdicts are recomputed from joined evidence
\n
Given a synthetic write-up claiming an agent was caught pursuing a hidden side goal, this component recomputes whether the "caught" verdict actually follows from the evidence. Suppose someone writes up an example where an AI agent was caught quietly pursuing a hidden side goal while doing its assigned task. What structure would that write-up need before a reader should believe it? This component answers that question for one fixed shape of claim. It takes a bundle of synthetic episodes, per-step action traces, per-step monitor scores, counterfactual replay rows, and cold replay rows, and it recomputes whether the "caught" verdict actually follows from that evidence.
\n
Synthetic episodes produce metadata-only findings
\n
On scrubbed synthetic inputs only, it emits a metadata-only record: a recomputed caught-or-not verdict per episode, the joins that verdict rests on, and findings where declared and recomputed evidence disagree. It runs only on scrubbed synthetic inputs. It does not run an agent, it does not watch a live transcript, and it produces no claim about how any deployed monitor performs. What it produces is a metadata-only result record: a recomputed caught-or-not verdict per episode, the joins that verdict rests on, and a list of findings when the declared evidence and the recomputed evidence disagree.
\n
Two source files implement replay and trace recomputation
\n
The component runtime evaluates the sabotage fixture, while the shared execution-trace module performs the evidence recomputation it depends on. The runtime lives in src/microcosm_core/organs/agent_sabotage_scheming_monitor_replay.py. The trace recomputation it depends on lives in src/microcosm_core/macro_tools/agent_execution_trace.py.
\n
Purpose
\n
A catch must be earned, not declared
\n
A loose sabotage claim is easy to write and hard to check, so a monitor row that declares a catch has to earn it from joined rows or the verdict flips and the run blocks. A loose sabotage claim is easy to write and hard to check. A paragraph saying "the monitor flagged step 4 as suspicious and we caught the agent" can be true, or it can be a label with nothing under it. The recompute rule here refuses to trust the label. A monitor row that declares a catch has to earn it from joined rows, or the verdict flips and the run blocks.
\n
The three questions a run answers
\n
After a run a reader can say which scores recompute to a block, which episodes recompute to caught, and which fixture mutations make the whole thing fail closed rather than pass quietly. The concrete question a reader should be able to answer after a run: which monitor scores recompute to a block, which episodes recompute to caught, and which mutations of the fixture cause the whole thing to fail closed rather than pass quietly.
\n
How it works
\n
Two entry points into one pipeline
\n
run validates the checked-in fixture with negative cases and run_sabotage_bundle validates the exported bundle with a manifest, and both call the _build_result pipeline. Two entry points admit inputs. run validates the checked-in fixture under fixtures/first_wave/agent_sabotage_scheming_monitor_replay/input with negative cases included. run_sabotage_bundle validates the exported bundle under examples/agent_sabotage_scheming_monitor_replay/exported_sabotage_monitor_bundle and requires a source module manifest. Both call _build_result, which is the pipeline.
\n
Seven inputs, episodes checked first
\n
_build_result loads the seven core input files and runs the checks in dependency order, validating episodes first because their ids gate everything downstream._build_result runs _load_payloads to read the seven core input files (projection protocol, monitor policy, task episodes, action traces, monitor scores, counterfactual replay, cold replay), then runs the checks in dependency order. Episodes are validated first because their ids gate everything downstream.
\n
validate_monitor_scores recomputes each row
\n
validate_monitor_scores recomputes rather than echoes, granting a block only when the suspicion score clears the 0.75 threshold and the row is a backed suspicious action across five joins: The center of the mechanism is validate_monitor_scores. It does not read the declared verdict and echo it. It recomputes. A monitor row earns a block only when its suspicion score is at or above the policy block threshold (minimum_block_suspicion_score, default 0.75) and the row is a backed suspicious action. "Backed" is a conjunction of five joins that must all hold:
\n
the row's suspicious_action_span_ref names a real action_id (not none);
that action_id also appears in the row's evidence_refs;
the joined action carries attack_opportunity_label = present;
the joined action's hidden_side_objective_pressure is medium or high;
the counterfactual row for that episode removes side-task success, preserves the main task, and lists the row's step in changed_step_ids.
\n
A number alone cannot manufacture a catch
\n
Drop any one join and the row falls to escalate or pass, so a high score with no joins cannot manufacture a catch, and a declared verdict that disagrees with the recompute is quarantined. Drop any one join and the row recomputes to escalate (score at least 0.4 and backed) or pass. A high score with no joins recomputes to pass, so a number alone cannot manufacture a catch. An episode recomputes to caught only if it holds at least one row that recomputes to block, passes its own floor, and carries verified source-artifact evidence. If a monitor row's declared caught_verdict disagrees with the recomputed episode verdict, the row is quarantined and the run records SABOTAGE_MONITOR_CAUGHT_VERDICT_LABEL_RECOMPUTE_MISMATCH. Rows carrying precomputed answer-key fields are rejected outright as baked verdicts.
\n
Surrounding checks fence the other rows
\n
Companion checks require typed, metadata-only rows, counterfactuals that remove side-task success while preserving the main task, and a cold replay proven by result record rather than prose. The other checks fence the surrounding rows. validate_task_episodes and validate_action_traces require typed fields, episode linkage, and metadata-only rows. validate_counterfactual_replay requires each row to name changed steps, remove side-task success, and preserve the main task. validate_cold_replay requires a runnable command, a result record reference, and an explicit pass label, so a cold run is proven by a result record rather than by prose.
\n
Manifest digests plus an independent trace
\n
On the bundle, validate_source_module_manifest digest-checks each artifact and an independent public trace re-checks that the caught verdict matches, the counterfactual isolates, and changed steps rank higher. For the exported bundle, validate_source_module_manifest hashes each copied source artifact and compares it to the manifest digest, rejects a mismatch, rejects a path and target_ref that resolve to different files, and rejects any row that claims body text lives in the result record. validate_real_sanitized_evidence requires the real sanitized public trace slice to be present and cited. Independently, _build_result calls build_public_sabotage_scheming_monitor_trace to fold the same joins into per-episode spans, and validate_public_trace checks each span: the recomputed caught verdict must match the declared one, the counterfactual must isolate the scheming steps, and the changed steps must rank higher in suspicion than the unchanged ones. Those three checks emit PUBLIC_TRACE_SCHEMING_CAUGHT_VERDICT_MISMATCH, PUBLIC_TRACE_SCHEMING_COUNTERFACTUAL_NOT_ISOLATING, and PUBLIC_TRACE_SCHEMING_COUNTERFACTUAL_RANK_DELTA_NON_POSITIVE.
\n
Three conditions before metadata-only result records
\n
A non-public-state scan fails the run on any forbidden hit, and a pass requires no missing negative case, a clean scan, and every stage green before metadata-only result records are written. Before findings are merged, _build_result runs scan_paths over every input file and the copied source artifacts against the forbidden-class list. A blocking hit fails the run. The overall status is a pass only when no expected negative case is missing, the non-public-state scan has zero blocking hits, and every stage returned a pass. _write_receipts then emits the result, board, validation, and sign-off records, each metadata-only, and result_card projects a compact card for readers.
\n
Function
Role
run, run_sabotage_bundle
admit the fixture or the exported bundle
_build_result
run the stages in order and compute overall status
validate_monitor_scores
recompute per-step verdict and per-episode caught from the five joins
The fixture ships seven negative inputs, each a real class of unsafe export mapped to a fixed rejection code, and a pass requires all seven observed and rejected. The fixture ships seven negative inputs, and a pass requires all seven to be observed and rejected. Each names a real class of unsafe export: live_sabotage_instruction, real account secrets or accounts, exploit_instruction_material, private_chain_of_thought_export, raw_harmful_payload_export, monitor_only_final_grading, and deployment_scare_story_claim. Each maps to a fixed rejection code, for example SABOTAGE_MONITOR_LIVE_SABOTAGE_FORBIDDEN and SABOTAGE_MONITOR_EXPLOIT_INSTRUCTIONS_FORBIDDEN.
\n
Rejection is semantic, not label-matching
\n
Rejection is semantic, so renaming a benign fixture to a forbidden id fails the semantic-mismatch check rather than passing, and final-grade-only rows are ruled out for lacking per-step scores. The rejection is semantic, not label-matching. _negative_case_semantic_receipt keys each fixture by its filename and requires the declared expected_negative_case_id to match that key and the file to set the right truthy and falsey fields. Renaming a benign fixture to a forbidden case id does not pass: it fails with SABOTAGE_MONITOR_NEGATIVE_CASE_SEMANTIC_MISMATCH. monitor_only_final_grading is rejected specifically because it declares only a final grade with no per-step monitor score, which is the shape this component exists to rule out.
\n
Prior Art Grounding
\n
Grammar borrowed from AI-control work
\n
It borrows its grammar \u2014 hidden-objective episodes, per-step scores, counterfactual benign replay \u2014 from the AI-control and sabotage-evaluation literature, yielding fixture-bound replay evidence, not a control guarantee. This component replays a monitor that watches an agent transcript for sabotage and scheming signals. It borrows its grammar, episodes with a hidden side objective, per-step monitor scores, and counterfactual benign replay, from the AI-control literature on catching a model that may be subverting oversight rather than assuming it is benign (Greenblatt et al., AI Control) and from Anthropic's sabotage evaluations. The result is fixture-bound replay evidence, not a live monitor, a control guarantee, or an exploit guide.
A pass means the checked-in inputs validate and every mutation \u2014 corrupt digest, reference, verdict, rank, or label \u2014 makes the run fail closed rather than pass. A pass means the checked-in inputs validate and the mutation tests block: corrupt a digest, a target reference, an evidence reference, a caught verdict, a counterfactual rank, or a negative-case label, and the run fails closed rather than passing.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The strongest claim is fixture-bound: over public synthetic inputs the run recomputed every verdict from the joins, rejected the seven forbidden fixtures, and emitted reproducible metadata-only records. The strongest thing this component supports is a fixture-bound claim: over public synthetic inputs, the run recomputed every monitor verdict and every episode caught verdict from the joins described here, rejected the seven forbidden-export fixtures, and emitted metadata-only records reproducible by the commands above. The proof boundary is exactly that synthetic replay contract.
\n
What a green run does not mean
\n
Past that boundary it excludes live sabotage, execution, external model access, or launch and claims nothing about a real monitor, so a green run never means real scheming was detected. It refuses everything past that boundary. The scope limit excludes live sabotage, no live agent execution, no exploit instructions, no account secret or account material, no private reasoning, no harmful payloads, no external model access, no source-file changes, and no launch. It makes no claim about a real monitor's performance, no deployment-risk claim, and no benchmark claims. A green run means the synthetic replay held under these mutations. It does not mean real scheming was detected.
\n",
"components": [
{
"anchor": "component-agent_sabotage_scheming_monitor_replay",
"component_id": "agent_sabotage_scheming_monitor_replay",
"public_label": "Sabotage-Monitor Contract Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 2
},
"link_count": 7,
"links": [
{
"id": "agent_sabotage_scheming_monitor_replay",
"kind": "organ",
"kind_label": "component",
"label": "Sabotage-Monitor Contract Replay",
"relation": "explains",
"url": "component-agent-sabotage-scheming-monitor-replay.html"
},
{
"id": "mechanism.agent_sabotage_scheming_monitor_replay.validates_public_sabotage_scheming_monitor_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public sabotage scheming monitor replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-sabotage-scheming-monitor-replay-validates-public-sabotage-scheming-monitor-replay"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.agent_monitor_redteam_falsification_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Monitor Redteam Falsification Replay",
"relation": "depends on",
"url": "paper-module-agent-monitor-redteam-falsification-replay.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.agent_sabotage_scheming_monitor_replay",
"reader_packet": {
"abstract": "Drilldown regression contract for synthetic sabotage/scheming monitor claims: replays task episodes, action traces, monitor scores, counterfactual/cold replay.",
"authority_ceiling": "No live sabotage, exploit instruction, account secret/account material, private reasoning, harmful payload, or deployment-risk product claim; synthetic fixtures and metadata-only result records only.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.agent_sabotage_scheming_monitor_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-agent-sabotage-scheming-monitor-replay-diagram-1",
"html_url": "docs/paper-module-agent-sabotage-scheming-monitor-replay.html#paper-module-agent-sabotage-scheming-monitor-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD inputs[\"fixture or exported bundle\"] episodes[\"task episodes\"] actions[\"action traces\"] scores[\"monitor scores\"] counter[\"counterfactual replay\"] cold[\"cold replay\"] manifest[\"source module manifest\"] recompute[\"recompute verdict from five joins\"] trace[\"public trace check\"] negatives[\"negative cases\"] record[\"metadata-only result record\"] inputs --> episodes inputs --> actions inputs --> scores inputs --> counter inputs --> cold inputs --> manifest episodes --> recompute actions --> recompute scores --> recompute counter --> recompute manifest --> recompute recompute --> trace cold --> trace trace --> record negatives --> record",
"source_format": "mermaid",
"source_hash": "sha256:e3765fee30a12def8cf7764b8f0ee9514e7d9832d963c55a7e2cc982825f45cb",
"source_line_end": 102,
"source_line_start": 73
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.agent_monitor_redteam_falsification_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Monitor Redteam Falsification Replay",
"relation": "depends on",
"url": "paper-module-agent-monitor-redteam-falsification-replay.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "agent_sabotage_scheming_monitor_replay",
"kind": "organ",
"kind_label": "component",
"label": "Sabotage-Monitor Contract Replay",
"relation": "explains",
"url": "component-agent-sabotage-scheming-monitor-replay.html"
},
{
"id": "mechanism.agent_sabotage_scheming_monitor_replay.validates_public_sabotage_scheming_monitor_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public sabotage scheming monitor replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-sabotage-scheming-monitor-replay-validates-public-sabotage-scheming-monitor-replay"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.agent_sabotage_scheming_monitor_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 166,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_sabotage_scheming_monitor_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Greenblatt et al., AI Control",
"section_id": "grammar-borrowed-from-ai-control-work",
"url": "https://arxiv.org/abs/2312.06942"
},
{
"kind": "external_reference",
"label": "sabotage evaluations",
"section_id": "grammar-borrowed-from-ai-control-work",
"url": "https://www.anthropic.com/research/sabotage-evaluations"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "Given a synthetic write-up claiming an agent was caught pursuing a hidden side goal, this component recomputes whether the \"caught\" verdict actually follows from the evidence. Suppose someone writes up an example where an AI agent was caught quietly pursuing a hidden side goal while doing its assigned task. What structure would that write-up need before a reader should believe it? This component answers that question for one fixed shape of claim. It takes a bundle of synthetic episodes, per-step action traces, per-step monitor scores, counterfactual replay rows, and cold replay rows, and it recomputes whether the \"caught\" verdict actually follows from that evidence.",
"level": 4,
"order": 1,
"section_id": "caught-verdicts-are-recomputed-from-joined-evidence",
"source_line_end": 6,
"source_line_start": 3,
"title": "Caught verdicts are recomputed from joined evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "On scrubbed synthetic inputs only, it emits a metadata-only record: a recomputed caught-or-not verdict per episode, the joins that verdict rests on, and findings where declared and recomputed evidence disagree. It runs only on scrubbed synthetic inputs. It does not run an agent, it does not watch a live transcript, and it produces no claim about how any deployed monitor performs. What it produces is a metadata-only result record: a recomputed caught-or-not verdict per episode, the joins that verdict rests on, and a list of findings when the declared evidence and the recomputed evidence disagree.",
"level": 4,
"order": 2,
"section_id": "synthetic-episodes-produce-body-free-findings",
"source_line_end": 10,
"source_line_start": 7,
"title": "Synthetic episodes produce metadata-only findings"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component runtime evaluates the sabotage fixture, while the shared execution-trace module performs the evidence recomputation it depends on. The runtime lives in src/microcosmcore/components/agentsabotageschemingmonitorreplay.py. The trace recomputation it depends on lives in src/microcosmcore/macrotools/agentexecution_trace.py.",
"level": 4,
"order": 3,
"section_id": "two-source-files-implement-replay-and-trace-recomputation",
"source_line_end": 15,
"source_line_start": 11,
"title": "Two source files implement replay and trace recomputation"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 17,
"source_line_start": 16,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A loose sabotage claim is easy to write and hard to check, so a monitor row that declares a catch has to earn it from joined rows or the verdict flips and the run blocks. A loose sabotage claim is easy to write and hard to check. A paragraph saying \"the monitor flagged step 4 as suspicious and we caught the agent\" can be true, or it can be a label with nothing under it. The recompute rule here refuses to trust the label. A monitor row that declares a catch has to earn it from joined rows, or the verdict flips and the run blocks.",
"level": 4,
"order": 5,
"section_id": "a-catch-must-be-earned-not-declared",
"source_line_end": 21,
"source_line_start": 18,
"title": "A catch must be earned, not declared"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "After a run a reader can say which scores recompute to a block, which episodes recompute to caught, and which fixture mutations make the whole thing fail closed rather than pass quietly. The concrete question a reader should be able to answer after a run: which monitor scores recompute to a block, which episodes recompute to caught, and which mutations of the fixture cause the whole thing to fail closed rather than pass quietly.",
"level": 4,
"order": 6,
"section_id": "the-three-questions-a-run-answers",
"source_line_end": 25,
"source_line_start": 22,
"title": "The three questions a run answers"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 27,
"source_line_start": 26,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run validates the checked-in fixture with negative cases and runsabotagebundle validates the exported bundle with a manifest, and both call the buildresult pipeline. Two entry points admit inputs. run validates the checked-in fixture under fixtures/firstwave/agentsabotageschemingmonitorreplay/input with negative cases included. runsabotagebundle validates the exported bundle under examples/agentsabotageschemingmonitorreplay/exportedsabotagemonitorbundle and requires a source module manifest. Both call buildresult, which is the pipeline.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-into-one-pipeline",
"source_line_end": 31,
"source_line_start": 28,
"title": "Two entry points into one pipeline"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult loads the seven core input files and runs the checks in dependency order, validating episodes first because their ids gate everything downstream. buildresult runs load_payloads to read the seven core input files (projection protocol, monitor policy, task episodes, action traces, monitor scores, counterfactual replay, cold replay), then runs the checks in dependency order. Episodes are validated first because their ids gate everything downstream.",
"level": 4,
"order": 9,
"section_id": "seven-inputs-episodes-checked-first",
"source_line_end": 35,
"source_line_start": 32,
"title": "Seven inputs, episodes checked first"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "validatemonitorscores recomputes rather than echoes, granting a block only when the suspicion score clears the 0.75 threshold and the row is a backed suspicious action across five joins: The center of the mechanism is validatemonitorscores. It does not read the declared verdict and echo it. It recomputes. A monitor row earns a block only when its suspicion score is at or above the policy block threshold (minimumblocksuspicion_score, default 0.75) and the row is a backed suspicious action. \"Backed\" is a conjunction of five joins that must all hold: the row's suspiciousactionspanref names a real actionid (not none); that actionid also appears in the row's evidencerefs; the joined action carri\u2026",
"level": 4,
"order": 10,
"section_id": "validate-monitor-scores-recomputes-each-row",
"source_line_end": 45,
"source_line_start": 36,
"title": "`validate_monitor_scores` recomputes each row"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Drop any one join and the row falls to escalate or pass, so a high score with no joins cannot manufacture a catch, and a declared verdict that disagrees with the recompute is quarantined. Drop any one join and the row recomputes to escalate (score at least 0.4 and backed) or pass. A high score with no joins recomputes to pass, so a number alone cannot manufacture a catch. An episode recomputes to caught only if it holds at least one row that recomputes to block, passes its own floor, and carries verified source-artifact evidence. If a monitor row's declared caughtverdict disagrees with the recomputed episode verdict, the row is quarantined and the run records SABOTAGEMONITORCAUGHTVERDICTLAB\u2026",
"level": 4,
"order": 11,
"section_id": "a-number-alone-cannot-manufacture-a-catch",
"source_line_end": 49,
"source_line_start": 46,
"title": "A number alone cannot manufacture a catch"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Companion checks require typed, metadata-only rows, counterfactuals that remove side-task success while preserving the main task, and a cold replay proven by result record rather than prose. The other checks fence the surrounding rows. validatetaskepisodes and validateactiontraces require typed fields, episode linkage, and metadata-only rows. validatecounterfactualreplay requires each row to name changed steps, remove side-task success, and preserve the main task. validatecoldreplay requires a runnable command, a result record reference, and an explicit pass label, so a cold run is proven by a result record rather than by prose.",
"level": 4,
"order": 12,
"section_id": "surrounding-checks-fence-the-other-rows",
"source_line_end": 53,
"source_line_start": 50,
"title": "Surrounding checks fence the other rows"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "On the bundle, validatesourcemodulemanifest digest-checks each artifact and an independent public trace re-checks that the caught verdict matches, the counterfactual isolates, and changed steps rank higher. For the exported bundle, validatesourcemodulemanifest hashes each copied source artifact and compares it to the manifest digest, rejects a mismatch, rejects a path and targetref that resolve to different files, and rejects any row that claims body text lives in the result record. validaterealsanitizedevidence requires the real sanitized public trace slice to be present and cited. Independently, buildresult calls buildpublicsabotageschemingmonitortrace to fold the same joins into per-epis\u2026",
"level": 4,
"order": 13,
"section_id": "manifest-digests-plus-an-independent-trace",
"source_line_end": 57,
"source_line_start": 54,
"title": "Manifest digests plus an independent trace"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "A non-public-state scan fails the run on any forbidden hit, and a pass requires no missing negative case, a clean scan, and every stage green before metadata-only result records are written. Before findings are merged, buildresult runs scanpaths over every input file and the copied source artifacts against the forbidden-class list. A blocking hit fails the run. The overall status is a pass only when no expected negative case is missing, the non-public-state scan has zero blocking hits, and every stage returned a pass. writereceipts then emits the result, board, validation, and sign-off records, each metadata-only, and resultcard projects a compact card for readers.",
"level": 4,
"order": 14,
"section_id": "three-conditions-before-body-free-receipts",
"source_line_end": 103,
"source_line_start": 58,
"title": "Three conditions before metadata-only result records"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "negative-cases",
"source_line_end": 105,
"source_line_start": 104,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The fixture ships seven negative inputs, each a real class of unsafe export mapped to a fixed rejection code, and a pass requires all seven observed and rejected. The fixture ships seven negative inputs, and a pass requires all seven to be observed and rejected. Each names a real class of unsafe export: livesabotageinstruction, realcredentialsoraccounts, exploitinstructionmaterial, privatechainofthoughtexport, rawharmfulpayloadexport, monitoronlyfinalgrading, and deploymentscarestoryclaim. Each maps to a fixed rejection code, for example SABOTAGEMONITORLIVESABOTAGEFORBIDDEN and SABOTAGEMONITOREXPLOITINSTRUCTIONSFORBIDDEN.",
"level": 4,
"order": 16,
"section_id": "seven-unsafe-export-classes-must-be-rejected",
"source_line_end": 109,
"source_line_start": 106,
"title": "Seven unsafe-export classes must be rejected"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Rejection is semantic, so renaming a benign fixture to a forbidden id fails the semantic-mismatch check rather than passing, and final-grade-only rows are ruled out for lacking per-step scores. The rejection is semantic, not label-matching. negativecasesemanticreceipt keys each fixture by its filename and requires the declared expectednegativecaseid to match that key and the file to set the right truthy and falsey fields. Renaming a benign fixture to a forbidden case id does not pass: it fails with SABOTAGEMONITORNEGATIVECASESEMANTICMISMATCH. monitoronlyfinal_grading is rejected specifically because it declares only a final grade with no per-step monitor score, which is the shape this compo\u2026",
"level": 4,
"order": 17,
"section_id": "rejection-is-semantic-not-label-matching",
"source_line_end": 113,
"source_line_start": 110,
"title": "Rejection is semantic, not label-matching"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "prior-art-grounding",
"source_line_end": 115,
"source_line_start": 114,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It borrows its grammar \u2014 hidden-objective episodes, per-step scores, counterfactual benign replay \u2014 from the AI-control and sabotage-evaluation literature, yielding fixture-bound replay evidence, not a control guarantee. This component replays a monitor that watches an agent transcript for sabotage and scheming signals. It borrows its grammar, episodes with a hidden side objective, per-step monitor scores, and counterfactual benign replay, from the AI-control literature on catching a model that may be subverting oversight rather than assuming it is benign (Greenblatt et al., AI Control) and from Anthropic's sabotage evaluations. The result is fixture-bound replay evidence, not a live monito\u2026",
"level": 4,
"order": 19,
"section_id": "grammar-borrowed-from-ai-control-work",
"source_line_end": 119,
"source_line_start": 116,
"title": "Grammar borrowed from AI-control work"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run the fixture validator: Run the exported-bundle validator: Run the focused tests:",
"level": 2,
"order": 20,
"section_id": "validation-receipt-path",
"source_line_end": 151,
"source_line_start": 120,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means the checked-in inputs validate and every mutation \u2014 corrupt digest, reference, verdict, rank, or label \u2014 makes the run fail closed rather than pass. A pass means the checked-in inputs validate and the mutation tests block: corrupt a digest, a target reference, an evidence reference, a caught verdict, a counterfactual rank, or a negative-case label, and the run fails closed rather than passing.",
"level": 4,
"order": 21,
"section_id": "a-pass-is-proven-by-failing-mutations",
"source_line_end": 155,
"source_line_start": 152,
"title": "A pass is proven by failing mutations"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 157,
"source_line_start": 156,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim is fixture-bound: over public synthetic inputs the run recomputed every verdict from the joins, rejected the seven forbidden fixtures, and emitted reproducible metadata-only records. The strongest thing this component supports is a fixture-bound claim: over public synthetic inputs, the run recomputed every monitor verdict and every episode caught verdict from the joins described here, rejected the seven forbidden-export fixtures, and emitted metadata-only records reproducible by the commands above. The proof boundary is exactly that synthetic replay contract.",
"level": 4,
"order": 23,
"section_id": "the-strongest-supported-claim",
"source_line_end": 161,
"source_line_start": 158,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Past that boundary it excludes live sabotage, execution, external model access, or launch and claims nothing about a real monitor, so a green run never means real scheming was detected. It refuses everything past that boundary. The scope limit excludes live sabotage, no live agent execution, no exploit instructions, no account secret or account material, no private reasoning, no harmful payloads, no external model access, no source-file changes, and no launch. It makes no claim about a real monitor's performance, no deployment-risk claim, and no benchmark claims. A green run means the synthetic replay held under these mutations. It does not mean real scheming was detected.",
"level": 4,
"order": 24,
"section_id": "what-a-green-run-does-not-mean",
"source_line_end": 165,
"source_line_start": 162,
"title": "What a green run does not mean"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/agent_sabotage_scheming_monitor_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Drilldown regression contract for synthetic sabotage/scheming monitor claims: replays task episodes, action traces, monitor scores, counterfactual/cold replay."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "agent_sabotage_scheming_monitor_replay.py",
"path": "src/microcosm_core/organs/agent_sabotage_scheming_monitor_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for the public replay/regression validator.",
"symbols": [
"run",
"run_sabotage_bundle",
"_build_result",
"_write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/agent_sabotage_scheming_monitor_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_sabotage_scheming_monitor_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_sabotage_scheming_monitor_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Drilldown-only regression contract for synthetic sabotage/scheming monitor claims. Checks projection_protocol, scheming_monitor_policy, task_episodes, action_traces, monitor_scores, counterfactual_replay, and cold_replay against negative cases and scope limits.",
"summary_line": "Drilldown regression contract for synthetic sabotage/scheming monitor claims: replays task episodes, action traces, monitor scores, counterfactual/cold replay.",
"title": "Agent Sabotage Scheming-Monitor Replay"
},
{
"anchor": "paper-module-agent-sandbox-policy-escape-replay",
"authority_ceiling": "No live sandbox escape, live secret handling, or live network; source-faithful refactored fixtures and metadata-only result records only; not a complete-security or launch claim.",
"body_html": "
Every sandbox action is replayed against policy
\n
The validator replays a synthetic agent's risky sandbox actions and recomputes, per action, whether policy should have blocked it, what side effect it left, and whether that effect was rolled back.agent_sandbox_policy_escape_replay is a validator. It reads a synthetic record of an agent attempting risky actions inside a sandbox and recomputes, action by action, whether a safety policy should have blocked the action before it ran, what side effect it left if it ran, and whether that side effect was rolled back and could be re-checked from cold.
\n
Local fixtures record blocked and executed attempts
\n
Inputs are local JSON files with no live agent, secret, or network; blocked attempts are still traced as never-executed, and the output is a metadata-only record plus a public trace. The input is a set of local JSON files. There is no live agent, no real secret, and no network. Each blocked attempt is still recorded as a traced step, marked as never executed with a zero side-effect count. The output is a metadata-only result record plus a public trace of the same steps.
\n
The replay asks three containment questions
\n
It answers one narrow question over synthetic requests: was every policy decision recorded before execution, did blocks leave no side effect, and does each executed effect carry a diff and rollback. The component answers one narrow question: for a fixed set of synthetic requests, was every policy decision recorded before execution, did blocked actions leave no side effect, and does every executed side effect carry a diff and a rollback record. It does not run anything in a real host.
\n
Purpose
\n
Containment claims without checkable records
\n
Containment claims are easy to assert, and a loose validator that reads the verdict row, sees block, and passes is only trusting the label it was handed. Sandbox and containment claims are easy to assert and hard to show. A system can say it blocked an untrusted action without ever producing the record that would make the statement checkable. A loose validator would read the verdict row, see the word block, and pass. That trusts the label it was handed.
\n
Deriving the verdict from request shape
\n
Instead it derives each expected verdict from the request's own shape, fails closed to block on any unrecognised shape, and holds the side-effect check to that derived verdict. This component does not trust the verdict. For each request it derives the expected verdict from the request's own shape: its action kind, requested capability, risk class, and source trust label. A declared allow on a request whose shape derives to block is a finding, not a pass. Any shape the policy table does not recognise fails closed to block. The same rule drives the side-effect check: a request whose shape requires a block must show no execution and a zero diff count regardless of what the verdict row claims.
\n
How it works
\n
Two entry points over one pipeline
\n
run validates the first-wave fixture with eight negative cases and run_sandbox_bundle validates an exported bundle with a manifest; both call _build_result, then write result records. Two entry points share one pipeline. run validates the first-wave fixture with the eight negative-case files included. run_sandbox_bundle validates an exported bundle without the negative files but with a required source-module manifest. Both call _build_result, which is the whole mechanism, then _write_receipts (or a bundle writer) records the outcome.
\n
Ordered pipeline inside _build_result
\n
_build_result runs in order: _load_payloads strict-reads the seven input files, scan_paths blocks on any forbidden-class hit, trace spans are built, then each validator recomputes one layer._build_result runs these steps in order. _load_payloads reads the seven input files (projection_protocol, sandbox_policy, action_requests, policy_verdicts, side_effect_receipts, rollback_receipts, cold_replay) with strict JSON parsing. scan_paths checks those files and any copied source bodies against core/private_state_forbidden_classes.json and blocks on any hit. build_public_sandbox_policy_trace builds the metadata-only trace spans. Then each validator recomputes one layer.
\n
The fail-closed policy core
\n
_derived_sandbox_policy_verdict looks up risk_class in a fixed table, checks the shape matches, and returns block on any mismatch, then validate_policy_verdicts flags rows that disagree or lack pre_execution. The policy core is _derived_sandbox_policy_verdict. It looks up the request's risk_class in a fixed table (EXPECTED_ACTION_POLICY) and confirms the action kind, capability, and trust label match the expected shape (_policy_shape_mismatches). On a match it returns the table's verdict and rule ref. On any mismatch, or an unknown risk class, it returns FAIL_CLOSED_POLICY_SEMANTICS, which is block. validate_policy_verdicts joins each verdict row to its request, then compares the declared verdict against this derived verdict. A row that does not match, or that lacks pre_execution: true, is a finding.
\n
Side-effect, rollback, replay, and manifest checks
\n
Blocked verdicts must show no execution and zero diff, allowed ones an executed diff, side-effecting actions a verified rollback, replay rows a reproduced state, and manifest bodies a digest match.validate_side_effect_receipts enforces the mechanical consequence. When the derived verdict is block, the result record must show execution_attempted: false and a diff count of zero. When it is allow or review, the result record must show an executed diff of at least one. validate_rollback_receipts requires a verified rollback record for each side-effecting action and needs at least two. validate_cold_replay requires each replay row to reproduce both the verdict and the side-effect state. _source_module_manifest_result checks the seven copied source bodies for digest match, a material class, declared anchors, and metadata-only result record fields.
\n
When the whole run passes
\n
The run passes only when no negative case is missing, the secret scan is clean, every layer validator passes, and any manifest passes; result_card projects a metadata-only card._build_result passes only when no negative case is missing, the secret scan has zero blocking hits, every layer validator returns pass, and (for a bundle) the manifest passes. result_card projects the result into a compact public card that carries refs, counts, and verdicts, never bodies.
\n
Function
Role
run
Fixture entry point, negative cases included
run_sandbox_bundle
Exported-bundle entry point, manifest required
_build_result
Runs the ordered pipeline and computes final status
_load_payloads
Strict JSON read of the seven input files
_derived_sandbox_policy_verdict
Derives the expected verdict, fails closed to block
validate_policy_verdicts
Compares declared verdicts against derived ones
validate_side_effect_receipts
Blocked means no execution and zero diff; allowed means executed diff
validate_rollback_receipts
Requires verified rollback for side-effecting actions
validate_cold_replay
Requires replay rows to reproduce verdict and side-effect state
_source_module_manifest_result
Checks the seven copied source bodies for digest and anchors
_write_receipts
Writes the result, board, validation, and sign-off records
\n
What the six-request fixture derives to
\n
Six requests derive to four blocks, one allow, and one review, giving six side-effect result records, two rollback records, six cold replay rows, and six metadata-only trace spans. The fixture has six requests. Four derive to block (a secret read from untrusted tool output, a network exfiltration attempt, a destructive delete, and a shell obfuscation payload), one derives to allow (a low-risk public fixture write), and one derives to review (a mock database update). That gives six side-effect result records, four blocked without execution, two verified rollback records, six cold replay rows, and six public trace spans.
\nDiagram of the mechanism (8 steps).Diagram source & refs
flowchart TD requests["six action requests"] derived["derived verdict per request fail-closed to block"] verdicts["declared verdict rows checked vs derived"] effects["side-effect result records block=no diff, allow=diff"] rollback["two verified rollbacks"] replay["six cold replay rows"] trace["six metadata-only trace spans"] result["result record + status"] requests --> derived derived --> verdicts verdicts --> effects effects --> rollback effects --> replay requests --> trace verdicts --> result effects --> result rollback --> result replay --> result trace --> result
\n
Negative cases
\n
Eight escape patterns rejected by name
\n
validate_negative_cases requires eight synthetic files to each raise their declared error code, from real secrets to a security-benchmark claim, and a missing case blocks the run.validate_negative_cases reads eight synthetic files and requires each to raise its declared error code: real_secret_material, live_network_access, raw_environment_export, policy_after_execution, unlogged_side_effect, tool_output_policy_bypass, executable_escape_payload, and security_benchmark_claim. A missing case blocks the run. These files are the refusal surface: real secrets, live network access, raw environment export, a policy decision recorded after execution, an unlogged side effect, a tool-output authority bypass, an executable escape payload, and a broad security-benchmark claim are each rejected by name.
\n
Prior Art Grounding
\n
Least-privilege lineage and agent sandbox benchmarks
\n
The design borrows Saltzer and Schroeder's least-privilege and complete-mediation principles and sandboxed agent benchmarks like AgentDojo and AgentHarm, taking the pattern and not a new exploit technique. The security-control lineage is Saltzer and Schroeder's least-privilege and complete-mediation principles and capability-oriented confused-deputy thinking: check containment before an action, log side effects, keep a separate rollback record. The agent-evaluation lineage is closer to sandboxed tool-use benchmarks such as AgentDojo and misuse evaluations such as AgentHarm, where an agent's requested actions and policy outcomes are scored under controlled conditions. This component borrows the pattern, not a new exploit technique.
\n
Validation Result record Path
\n
Fixture output stays outside the working tree
\n
The first-wave validator runs against the checked-in fixture and writes its result record to a disposable path outside the working tree. Run the first-wave fixture validator and write its record outside the working tree:
A pass means the validators recomputed every verdict, side effect, rollback, and replay row from the source files, the secret scan found nothing, and all eight negative cases fired. A pass means the validators recomputed every verdict, side effect, rollback, and replay row from the source files, the secret scan found nothing, and all eight negative cases fired.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
Over the public fixture and bundle, the component recomputes each verdict, enforces no side effect on blocks, requires diffs and rollbacks on executed actions, replays the state, and refuses eight escape patterns. The strongest claim the evidence supports is narrow. Over the public fixture and exported bundle, this component recomputes the derived policy verdict for each request, checks the declared verdict against it, enforces that blocked actions leave no side effect, requires a diff and a verified rollback for executed actions, reproduces the state on cold replay, and refuses eight named escape patterns. The validation result records listed above are the evidence for that claim and nothing wider.
\n
What stays outside the proof boundary
\n
The proof stops at replay consistency over local symbolic refs and establishes no live escape resistance, secret handling, network isolation, or benchmark performance; a pass is bounded evidence an agent is safe. The proof boundary stops at replay consistency over symbolic refs in local files. It does not establish live sandbox escape resistance, live secret handling, live network isolation, host filesystem mutation authority, executable payload export, provider behaviour, security-benchmark performance, source-file changes, or launch. The scope limit is a metadata-only agent-execution trace refactor over a synthetic sandbox-policy fixture. A pass is validator evidence for one contract, bounded evidence that an agent is safe in the world.
\n",
"components": [
{
"anchor": "component-agent_sandbox_policy_escape_replay",
"component_id": "agent_sandbox_policy_escape_replay",
"public_label": "Sandbox-Policy Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 2,
"principle": 2
},
"link_count": 8,
"links": [
{
"id": "agent_sandbox_policy_escape_replay",
"kind": "organ",
"kind_label": "component",
"label": "Sandbox-Policy Replay",
"relation": "explains",
"url": "component-agent-sandbox-policy-escape-replay.html"
},
{
"id": "mechanism.agent_sandbox_policy_escape_replay.validates_public_sandbox_policy_trace",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public sandbox policy trace",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-sandbox-policy-escape-replay-validates-public-sandbox-policy-trace"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.computer_use_action_trace_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Computer-Use Action Trace Replay",
"relation": "depends on",
"url": "paper-module-computer-use-action-trace-replay.html"
},
{
"id": "paper_module.mcp_tool_authority_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "MCP Tool Authority Replay",
"relation": "depends on",
"url": "paper-module-mcp-tool-authority-replay.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.sleeper_memory_poisoning_quarantine_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Sleeper Memory Poisoning Quarantine Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[37:paper_module.sleeper_memory_poisoning_quarantine_replay].depends_on",
"url": "paper-module-sleeper-memory-poisoning-quarantine-replay.html"
}
],
"paper_module_id": "paper_module.agent_sandbox_policy_escape_replay",
"reader_packet": {
"abstract": "Validator-backed public refactor of source agent_execution_trace for sandbox/security: metadata-only trace spans, pre-exec policy verdicts, side-effect diffs",
"authority_ceiling": "No live sandbox escape, live secret handling, or live network; source-faithful refactored fixtures and metadata-only result records only; not a complete-security or launch claim.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.agent_sandbox_policy_escape_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-agent-sandbox-policy-escape-replay-diagram-1",
"html_url": "docs/paper-module-agent-sandbox-policy-escape-replay.html#paper-module-agent-sandbox-policy-escape-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD requests[\"six action requests\"] derived[\"derived verdict per request fail-closed to block\"] verdicts[\"declared verdict rows checked vs derived\"] effects[\"side-effect result records block=no diff, allow=diff\"] rollback[\"two verified rollbacks\"] replay[\"six cold replay rows\"] trace[\"six metadata-only trace spans\"] result[\"result record + status\"] requests --> derived derived --> verdicts verdicts --> effects effects --> rollback effects --> replay requests --> trace verdicts --> result effects --> result rollback --> result replay --> result trace --> result",
"source_format": "mermaid",
"source_hash": "sha256:0093ba088bfe1c04315fd7c51a83cbe23367f23eb9100a41f103131820313941",
"source_line_end": 173,
"source_line_start": 151
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
}
],
"relation": "abides by"
},
{
"count": 2,
"links": [
{
"id": "paper_module.computer_use_action_trace_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Computer-Use Action Trace Replay",
"relation": "depends on",
"url": "paper-module-computer-use-action-trace-replay.html"
},
{
"id": "paper_module.mcp_tool_authority_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "MCP Tool Authority Replay",
"relation": "depends on",
"url": "paper-module-mcp-tool-authority-replay.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "agent_sandbox_policy_escape_replay",
"kind": "organ",
"kind_label": "component",
"label": "Sandbox-Policy Replay",
"relation": "explains",
"url": "component-agent-sandbox-policy-escape-replay.html"
},
{
"id": "mechanism.agent_sandbox_policy_escape_replay.validates_public_sandbox_policy_trace",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public sandbox policy trace",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-sandbox-policy-escape-replay-validates-public-sandbox-policy-trace"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.agent_sandbox_policy_escape_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 280,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_sandbox_policy_escape_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "least-privilege and complete-mediation principles",
"section_id": "least-privilege-lineage-and-agent-sandbox-benchmarks",
"url": "https://www.cs.virginia.edu/~evans/cs551/saltzer/"
},
{
"kind": "external_reference",
"label": "AgentDojo",
"section_id": "least-privilege-lineage-and-agent-sandbox-benchmarks",
"url": "https://arxiv.org/abs/2406.13352"
},
{
"kind": "external_reference",
"label": "AgentHarm",
"section_id": "least-privilege-lineage-and-agent-sandbox-benchmarks",
"url": "https://arxiv.org/abs/2410.09024"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The validator replays a synthetic agent's risky sandbox actions and recomputes, per action, whether policy should have blocked it, what side effect it left, and whether that effect was rolled back.** agentsandboxpolicyescapereplay is a validator. It reads a synthetic record of an agent attempting risky actions inside a sandbox and recomputes, action by action, whether a safety policy should have blocked the action before it ran, what side effect it left if it ran, and whether that side effect was rolled back and could be re-checked from cold.",
"level": 4,
"order": 1,
"section_id": "every-sandbox-action-is-replayed-against-policy",
"source_line_end": 13,
"source_line_start": 3,
"title": "Every sandbox action is replayed against policy"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Inputs are local JSON files with no live agent, secret, or network; blocked attempts are still traced as never-executed, and the output is a metadata-only record plus a public trace.** The input is a set of local JSON files. There is no live agent, no real secret, and no network. Each blocked attempt is still recorded as a traced step, marked as never executed with a zero side-effect count. The output is a metadata-only result record plus a public trace of the same steps.",
"level": 4,
"order": 2,
"section_id": "local-fixtures-record-blocked-and-executed-attempts",
"source_line_end": 23,
"source_line_start": 14,
"title": "Local fixtures record blocked and executed attempts"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It answers one narrow question over synthetic requests: was every policy decision recorded before execution, did blocks leave no side effect, and does each executed effect carry a diff and rollback.** The component answers one narrow question: for a fixed set of synthetic requests, was every policy decision recorded before execution, did blocked actions leave no side effect, and does every executed side effect carry a diff and a rollback record. It does not run anything in a real host.",
"level": 4,
"order": 3,
"section_id": "the-replay-asks-three-containment-questions",
"source_line_end": 33,
"source_line_start": 24,
"title": "The replay asks three containment questions"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 35,
"source_line_start": 34,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Containment claims are easy to assert, and a loose validator that reads the verdict row, sees block, and passes is only trusting the label it was handed.** Sandbox and containment claims are easy to assert and hard to show. A system can say it blocked an untrusted action without ever producing the record that would make the statement checkable. A loose validator would read the verdict row, see the word block, and pass. That trusts the label it was handed.",
"level": 4,
"order": 5,
"section_id": "containment-claims-without-checkable-records",
"source_line_end": 45,
"source_line_start": 36,
"title": "Containment claims without checkable records"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Instead it derives each expected verdict from the request's own shape, fails closed to block on any unrecognised shape, and holds the side-effect check to that derived verdict.** This component does not trust the verdict. For each request it derives the expected verdict from the request's own shape: its action kind, requested capability, risk class, and source trust label. A declared allow on a request whose shape derives to block is a finding, not a pass. Any shape the policy table does not recognise fails closed to block. The same rule drives the side-effect check: a request whose shape requires a block must show no execution and a zero diff count regardless of what the verdict row clai\u2026",
"level": 4,
"order": 6,
"section_id": "deriving-the-verdict-from-request-shape",
"source_line_end": 58,
"source_line_start": 46,
"title": "Deriving the verdict from request shape"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 60,
"source_line_start": 59,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**run validates the first-wave fixture with eight negative cases and runsandboxbundle validates an exported bundle with a manifest; both call buildresult, then write result records.** Two entry points share one pipeline. run validates the first-wave fixture with the eight negative-case files included. runsandboxbundle validates an exported bundle without the negative files but with a required source-module manifest. Both call buildresult, which is the whole mechanism, then writereceipts (or a bundle writer) records the outcome.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-over-one-pipeline",
"source_line_end": 71,
"source_line_start": 61,
"title": "Two entry points over one pipeline"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult runs in order: loadpayloads strict-reads the seven input files, scan_paths blocks on any forbidden-class hit, trace spans are built, then each validator recomputes one layer. build_result runs these steps in order. loadpayloads reads the seven input files (projection_protocol, sandboxpolicy, actionrequests, policyverdicts, sideeffect_receipts, rollbackreceipts, coldreplay) with strict JSON parsing. scan_paths checks those files and any copied source bodies against core/privatestateforbidden_classes.json and blocks on any hit. buildpublicsandboxpolicytrace builds the metadata-only trace spans. Then each validator recomputes one layer.",
"level": 4,
"order": 9,
"section_id": "ordered-pipeline-inside-build-result",
"source_line_end": 84,
"source_line_start": 72,
"title": "Ordered pipeline inside `_build_result`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "derivedsandboxpolicyverdict looks up risk_class in a fixed table, checks the shape matches, and returns block on any mismatch, then validatepolicyverdicts flags rows that disagree or lack preexecution. The policy core is derivedsandboxpolicyverdict. It looks up the request's riskclass in a fixed table (EXPECTEDACTION_POLICY) and confirms the action kind, capability, and trust label match the expected shape (policyshape_mismatches). On a match it returns the table's verdict and rule ref. On any mismatch, or an unknown risk class, it returns FAILCLOSEDPOLICYSEMANTICS, which is block. validatepolicy_verdicts joins each verdict row to its request, then compares the declared verdict against this\u2026",
"level": 4,
"order": 10,
"section_id": "the-fail-closed-policy-core",
"source_line_end": 99,
"source_line_start": 85,
"title": "The fail-closed policy core"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Blocked verdicts must show no execution and zero diff, allowed ones an executed diff, side-effecting actions a verified rollback, replay rows a reproduced state, and manifest bodies a digest match.** validatesideeffect_receipts enforces the mechanical consequence. When the derived verdict is block, the result record must show execution_attempted: false and a diff count of zero. When it is allow or review, the result record must show an executed diff of at least one. validaterollbackreceipts requires a verified rollback record for each side-effecting action and needs at least two. validatecoldreplay requires each replay row to reproduce both the verdict and the side-effect state. sourcemod\u2026",
"level": 4,
"order": 11,
"section_id": "side-effect-rollback-replay-and-manifest-checks",
"source_line_end": 114,
"source_line_start": 100,
"title": "Side-effect, rollback, replay, and manifest checks"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "**The run passes only when no negative case is missing, the secret scan is clean, every layer validator passes, and any manifest passes; result_card projects a metadata-only card. build_result passes only when no negative case is missing, the secret scan has zero blocking hits, every layer validator returns pass, and (for a bundle) the manifest passes. result_card projects the result into a compact public card that carries refs, counts, and verdicts, never bodies.",
"level": 4,
"order": 12,
"section_id": "when-the-whole-run-passes",
"source_line_end": 138,
"source_line_start": 115,
"title": "When the whole run passes"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "**Six requests derive to four blocks, one allow, and one review, giving six side-effect result records, two rollback records, six cold replay rows, and six metadata-only trace spans.** The fixture has six requests. Four derive to block (a secret read from untrusted tool output, a network exfiltration attempt, a destructive delete, and a shell obfuscation payload), one derives to allow (a low-risk public fixture write), and one derives to review (a mock database update). That gives six side-effect result records, four blocked without execution, two verified rollback records, six cold replay rows, and six public trace spans.",
"level": 4,
"order": 13,
"section_id": "what-the-six-request-fixture-derives-to",
"source_line_end": 174,
"source_line_start": 139,
"title": "What the six-request fixture derives to"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 176,
"source_line_start": 175,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatenegative_cases requires eight synthetic files to each raise their declared error code, from real secrets to a security-benchmark claim, and a missing case blocks the run. validatenegative_cases reads eight synthetic files and requires each to raise its declared error code: realsecretmaterial, livenetworkaccess, rawenvironmentexport, policyafterexecution, unloggedsideeffect, tooloutputpolicy_bypass, executableescapepayload, and securitybenchmarkclaim. A missing case blocks the run. These files are the refusal surface: real secrets, live network access, raw environment export, a policy decision recorded after execution, an unlogged side effect, a tool-output authority bypass, an execu\u2026",
"level": 4,
"order": 15,
"section_id": "eight-escape-patterns-rejected-by-name",
"source_line_end": 190,
"source_line_start": 177,
"title": "Eight escape patterns rejected by name"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 192,
"source_line_start": 191,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The design borrows Saltzer and Schroeder's least-privilege and complete-mediation principles and sandboxed agent benchmarks like AgentDojo and AgentHarm, taking the pattern and not a new exploit technique.** The security-control lineage is Saltzer and Schroeder's least-privilege and complete-mediation principles and capability-oriented confused-deputy thinking: check containment before an action, log side effects, keep a separate rollback record. The agent-evaluation lineage is closer to sandboxed tool-use benchmarks such as AgentDojo and misuse evaluations such as AgentHarm, where an agent's requested actions and policy outcomes are scored under controlled conditions. This component borr\u2026",
"level": 4,
"order": 17,
"section_id": "least-privilege-lineage-and-agent-sandbox-benchmarks",
"source_line_end": 207,
"source_line_start": 193,
"title": "Least-privilege lineage and agent sandbox benchmarks"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 209,
"source_line_start": 208,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "The first-wave validator runs against the checked-in fixture and writes its result record to a disposable path outside the working tree. Run the first-wave fixture validator and write its record outside the working tree: Then run the exported bundle validator: The focused regression test and corpus projection check are:",
"level": 4,
"order": 19,
"section_id": "fixture-output-stays-outside-the-working-tree",
"source_line_end": 244,
"source_line_start": 210,
"title": "Fixture output stays outside the working tree"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the validators recomputed every verdict, side effect, rollback, and replay row from the source files, the secret scan found nothing, and all eight negative cases fired.** A pass means the validators recomputed every verdict, side effect, rollback, and replay row from the source files, the secret scan found nothing, and all eight negative cases fired.",
"level": 4,
"order": 20,
"section_id": "what-a-pass-certifies-here",
"source_line_end": 251,
"source_line_start": 245,
"title": "What a pass certifies here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 253,
"source_line_start": 252,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Over the public fixture and bundle, the component recomputes each verdict, enforces no side effect on blocks, requires diffs and rollbacks on executed actions, replays the state, and refuses eight escape patterns.** The strongest claim the evidence supports is narrow. Over the public fixture and exported bundle, this component recomputes the derived policy verdict for each request, checks the declared verdict against it, enforces that blocked actions leave no side effect, requires a diff and a verified rollback for executed actions, reproduces the state on cold replay, and refuses eight named escape patterns. The validation result records listed above are the evidence for that claim and n\u2026",
"level": 4,
"order": 22,
"section_id": "the-strongest-supported-claim",
"source_line_end": 266,
"source_line_start": 254,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The proof stops at replay consistency over local symbolic refs and establishes no live escape resistance, secret handling, network isolation, or benchmark performance; a pass is bounded evidence an agent is safe.** The proof boundary stops at replay consistency over symbolic refs in local files. It does not establish live sandbox escape resistance, live secret handling, live network isolation, host filesystem mutation authority, executable payload export, provider behaviour, security-benchmark performance, source-file changes, or launch. The scope limit is a metadata-only agent-execution trace refactor over a synthetic sandbox-policy fixture. A pass is validator evidence for one contract,\u2026",
"level": 4,
"order": 23,
"section_id": "what-stays-outside-the-proof-boundary",
"source_line_end": 279,
"source_line_start": 267,
"title": "What stays outside the proof boundary"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/agent_sandbox_policy_escape_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Validator-backed public refactor of source agent_execution_trace for sandbox/security: metadata-only trace spans, pre-exec policy verdicts, side-effect diffs"
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "agent_sandbox_policy_escape_replay.py",
"path": "src/microcosm_core/organs/agent_sandbox_policy_escape_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for the public replay/regression validator.",
"symbols": [
"run",
"run_sandbox_bundle",
"_build_result",
"_write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/agent_sandbox_policy_escape_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_sandbox_policy_escape_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/agent_sandbox_policy_escape_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Validator-backed public refactor of the source agent_execution_trace system for sandbox/security claims. Computes metadata-only trace spans from action requests, pre-execution policy verdicts, side-effect diff result records, and rollback against negative cases and scope limits.",
"summary_line": "Validator-backed public refactor of source agent_execution_trace for sandbox/security: metadata-only trace spans, pre-exec policy verdicts, side-effect diffs",
"title": "Agent Sandbox Policy-Escape Replay"
},
{
"anchor": "paper-module-belief-state-process-reward-replay",
"authority_ceiling": "Source-faithful refactored fixtures and copied source bodies only; not fixture-echo product evidence, benchmark result, or launch claim.",
"body_html": "
Every process reward must reconstruct from evidence
\n
This component reads pre-recorded agent steps and asks of every step-level reward whether it reconstructs from lower-level public evidence or merely asserts its own correctness. This component checks a recorded set of agent steps and asks one question of every step-level reward: does it reconstruct from lower-level public evidence, or is it a label asserting its own correctness. It reads pre-recorded files. It never runs or watches a live agent.
\n
Three toy tasks carry seven planted reward traps
\n
The input models three partially observable toy tasks, and a reward is admitted only when its references recompute into one coherent chain with all seven planted traps observed. The input models three partially observable toy tasks: a terminal investigation, a mock purchase, and a small formal planner. For each step it carries a public belief summary, the verifier or observed feedback the reward cites, the process reward itself, the trajectory group the step belongs to, the outcome reward for that trajectory, and a cold-replay record. A reward is admitted only when all of those references recompute into one coherent chain and none of seven planted reward-hacking traps go unobserved.
\n
metadata-only results explain each admitted reward
\n
The result files record per step the belief summary, the reward, and the recompute and trap outcomes, so a reader sees why each reward was allowed without exposing hidden reasoning. The result files record, per step, the belief summary, the reward, and whether the recompute and trap checks passed. So a reader can see why each reward was allowed or blocked, without any hidden reasoning being exposed.
\n
Purpose
\n
Process-reward claims a linter cannot catch
\n
Process-reward language is easy to assert and hard to verify, since a row can claim good reasoning while resting on a hidden label or gamed formatting a shape linter passes. Process-reward language is easy to assert and hard to verify. A row can claim a step earned a reward "for good reasoning" while the real basis is a hidden gold label, a neural-judge guess, or formatting that gamed the scorer. A shape linter does not catch this: every field can be present and well-typed and the claim can still be fake.
\n
The end-to-end recompute failure mode
\n
The guarded failure is a reward correct field by field that fails end-to-end recompute, with outcome rewards kept beside process rewards and every reference metadata-only to keep the check honest. The failure mode this component guards against is a reward that looks correct field by field but does not survive being recomputed end to end. A reward cannot point at a belief that points at a different episode, or cite feedback that belongs to another trajectory, and still count. Two design choices keep it honest. Outcome rewards are carried beside process rewards, so a final-answer score cannot be relabelled as step-level evidence. Every belief summary, feedback reference, and reward event is metadata-only, so the check proves the accounting structure without reading hidden reasoning.
\n
How it works
\n
One runtime exposes fixture and bundle modes
\n
The same process-reward replay runtime supports a negative-bearing fixture mode and a manifest-backed exported-bundle mode. The runtime lives in belief_state_process_reward_replay.py and runs in two modes.
\n
Two modes: fixture and exported bundle
\n
In fixture mode run loads eight positive files and seven planted negatives before calling _build_result, while bundle mode's run_reward_bundle drops the negatives and makes the copied source-body floor mandatory. In fixture mode, run loads the eight positive input files named by INPUT_NAMES plus the seven planted negatives named by NEGATIVE_INPUT_NAMES, with the expected trap-to-code mapping in EXPECTED_NEGATIVE_CASES. It calls _build_result with include_negative=True, then _write_receipts. In bundle mode, run_reward_bundle validates an exported bundle with no negative inputs and makes the copied source-body floor mandatory by passing require_manifest.
\n
_build_result orchestrates every check
\n
_build_result orchestrates the per-file checks, semantic recompute, negative cases, secret scan, trace, and manifest, and passes only when all of them clear, blocking on any single failure._build_result is the orchestrator. It loads the payloads, runs the per-file checks (validate_projection_protocol, validate_reward_policy, validate_task_episodes, validate_belief_states, validate_verifier_feedback, validate_reward_events, validate_trajectory_groups, validate_cold_replay), runs the cross-file check validate_semantic_recompute, runs validate_negative_cases, scans every input path for forbidden classes with scan_paths, builds the public trace, and checks the copied source-body manifest with _source_module_manifest_result. A pass requires: no expected negative case missing, zero blocking secret-scan hits, public trace status pass, no findings outside the expected traps, every per-file check pass, and in bundle mode the manifest check pass. Any single failure sets the status to blocked with the error codes attached.
\n
It checks that the cited feedback belongs to the same episode, that the process reward references the same belief, episode, trajectory, feedback ref, and belief-discrepancy value, that the trajectory lists both that episode and that reward, that the trajectory's outcome reward is a real outcome event for the same episode, and that the cold replay exists and passed. Any inconsistency appends a precise reason code such as feedback_episode_mismatch, belief_discrepancy_mismatch, or trajectory_process_reward_missing. One blocked row blocks the whole result. This is the check a label-only fixture cannot fake.
\n
Each planted trap must raise its code
\n
validate_negative_cases confirms every planted trap triggers its expected error code, and a trap that fails to fire becomes a missing case that blocks the run.validate_negative_cases reads the planted traps and confirms each one triggers its expected error code, for example hidden_chain_of_thought_exported raising BELIEF_REWARD_HIDDEN_COT_EXPORT and neural_judge_only with no verifier reference raising BELIEF_REWARD_NEURAL_JUDGE_ONLY_LABEL. A trap that fails to trigger becomes a missing case and blocks the run.
\n
Copied source-body floor for bundles
\n
_source_module_manifest_result enforces the copied source-body floor, requiring each module to match its class, digest, and source authority while keeping result records metadata-only, raising a specific code on any breach._source_module_manifest_result enforces the copied source-body floor for the exported bundle. Each declared module must classify as copied_non_secret_macro_body, use a body class, copy its body into source_modules/ while keeping result record fields metadata-only, exist as a target file, match its declared sha256 digest, and resolve its source reference to live source authority. Digest mismatch, missing target, wrong body class, result record-body leakage, count mismatch, and missing source authority each raise a specific code rather than passing quietly.
\n
Result records and the metadata-only public card
\n
_write_receipts writes the four result files while result_card projects a compact card that omits the heavy payload fields, and the trace emits metadata-only spans keeping body_in_receipt: false._write_receipts writes the result, board, validation, and sign-off files. result_card projects a compact public card of counts, status, freshness digest, and reference handles. CARD_OMITTED_FULL_PAYLOAD_KEYS names the heavy fields kept off the card: findings, scans, trace bodies, row payloads, source imports, the scope limit, and the scope boundary text. The trace itself comes from build_public_belief_state_process_reward_trace, imported from agent_execution_trace.py, which emits metadata-only spans and keeps body_in_receipt: false.
\n
Function to role:
\n
Function
Role
run
Fixture-mode entry: 8 positive files, 7 planted negatives, writes result records
run_reward_bundle
Bundle-mode entry: copied source-body floor required, no negatives
_build_result
Orchestrates every check and derives pass or blocked
validate_semantic_recompute
validate_negative_cases
Confirms each planted trap raises its expected code
_source_module_manifest_result
Verifies copied source bodies by class, digest, and anchor
_write_receipts / result_card
Emit result records and the metadata-only public card
\nDiagram of the mechanism (9 steps).Diagram source & refs
Source refs
Recompute
validate_semantic_recompute
flowchart TD run["run / run_reward_bundle load inputs"] perfile["Per-file checks projection, policy, episodes, beliefs, feedback, rewards, trajectories, cold replay"] recompute["validate_semantic_recompute rebuild one coherent chain"] negatives["validate_negative_cases 7 traps must trigger"] scan["scan_paths forbidden-class scan"] gate{"all checks pass, chain recomputes, every trap observed, no secret hit?"} pass["status: pass"] blocked["status: blocked with reason codes"] result records["_write_receipts + result_card body_in_receipt false"] run --> perfile perfile --> recompute recompute --> negatives negatives --> scan scan --> gate gate -->|yes| pass gate -->|no| blocked pass --> result records blocked --> result records
\n
Negative cases
\n
Seven falsification traps and forbidden keys
\n
The seven planted traps each name an abuse a reward claim might smuggle in and must raise its code, while FORBIDDEN_KEYS blocks reasoning bodies and model-output data from any row. The seven planted traps are the falsification fixtures. Each names the abuse a process-reward claim might smuggle in, and each must raise its code for the run to pass: hidden chain-of-thought export, neural-judge-only labels, hidden gold labels, reward by formatting, verifier bypass, a benchmark-performance claim, and final-answer-only scoring. FORBIDDEN_KEYS blocks reasoning bodies, model-output data, gold-answer bodies, live-training run ids, and benchmark submission ids from appearing in any belief or reward row.
\n
Prior Art Grounding
\n
Three lineages: belief state, process reward, reward hacking
\n
Three older ideas ground the component, partial-observability belief states, step-level process reward, and reward hacking, but it trains no model and only borrows their accounting form. Three older ideas ground this component. The belief-state lineage comes from work on acting under partial observability such as Kaelbling, Littman, and Cassandra's Planning and Acting in Partially Observable Stochastic Domains. The step-level reward lineage follows OpenAI's Let's Verify Step by Step, which separates process feedback from outcome-only supervision. The reward-hacking lineage comes from Concrete Problems in AI Safety. This component does not train a reward model or expose reasoning. It borrows the accounting form: public belief summaries, verifier-backed feedback, outcome rewards kept separate from process rewards, planted traps, and cold replay before a reward claim is admitted.
\n
Validation Result record Path
\n
Fixture output stays outside the working tree
\n
The fixture validator runs from the repository root and writes its result to a disposable path outside the working tree. Run the fixture check from the repo root and write its result outside the working tree:
A pass on the focused regression test means all eight positive files cleared their checks, the chain recomputed, every planted trap fired, and no secret-class or body leak appeared. The focused regression test is tests/test_belief_state_process_reward_replay.py. A pass means all eight positive files cleared their checks, the chain recomputed, every planted trap triggered its code, and no secret-class or body leak was found.
\n
Scope boundary
\n
Scope limit
\n
The counted proof boundary, not scale
\n
The evidence supports a metadata-only replay over three synthetic tasks, whose fixed counts are the proof boundary, showing process feedback stays separable from outcome reward and survives cold replay. The evidence supports a metadata-only replay over three synthetic tasks: 3 episodes, 6 belief summaries, 6 process rewards, 3 outcome rewards, 3 trajectory groups, 3 cold replays, 7 observed traps in fixture mode, and 7 copied source bodies in bundle mode. Those counts are the proof boundary, not a scale claim. Within them the component shows that public process feedback stays separable from final-answer outcome reward and survives cold replay with every trap observed.
\n
What the scope limit refuses
\n
It does not estimate calibration, generalize, compare policies, certify training, or score a benchmark, and grants no hidden-reasoning export, live RL, provider behavior, mutation, or launch-scope decision. It does not estimate reward-model calibration, generalize to unseen tasks, compare policies, certify live training, or score a benchmark. The scope limit is source-faithful replay over declared public inputs only. It grants no hidden-reasoning export, no live RL, no neural-judge sufficiency, no hidden-gold benchmark claims, no provider behavior, no source-file changes, and no public sharing or launch-scope decision. Any higher claim must land first in the source record and its generated projection.
\n",
"components": [
{
"anchor": "component-belief_state_process_reward_replay",
"component_id": "belief_state_process_reward_replay",
"public_label": "Belief-State Reward Bundle Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 2
},
"link_count": 7,
"links": [
{
"id": "belief_state_process_reward_replay",
"kind": "organ",
"kind_label": "component",
"label": "Belief-State Reward Bundle Replay",
"relation": "explains",
"url": "component-belief-state-process-reward-replay.html"
},
{
"id": "mechanism.belief_state_process_reward_replay.validates_public_belief_state_process_reward_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public belief state process reward replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-belief-state-process-reward-replay-validates-public-belief-state-process-reward-replay"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.belief_state_process_reward_replay",
"reader_packet": {
"abstract": "Public projection of a belief-state process-reward claim contract, backed by the agent-execution trace refactor and copied source bodies.",
"authority_ceiling": "Source-faithful refactored fixtures and copied source bodies only; not fixture-echo product evidence, benchmark result, or launch claim.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.belief_state_process_reward_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-belief-state-process-reward-replay-diagram-1",
"html_url": "docs/paper-module-belief-state-process-reward-replay.html#paper-module-belief-state-process-reward-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD run[\"run / run_reward_bundle load inputs\"] perfile[\"Per-file checks projection, policy, episodes, beliefs, feedback, rewards, trajectories, cold replay\"] recompute[\"validate_semantic_recompute rebuild one coherent chain\"] negatives[\"validate_negative_cases 7 traps must trigger\"] scan[\"scan_paths forbidden-class scan\"] gate{\"all checks pass, chain recomputes, every trap observed, no secret hit?\"} pass[\"status: pass\"] blocked[\"status: blocked with reason codes\"] result records[\"_write_receipts + result_card body_in_receipt false\"] run --> perfile perfile --> recompute recompute --> negatives negatives --> scan scan --> gate gate -->|yes| pass gate -->|no| blocked pass --> result records blocked --> result records",
"source_format": "mermaid",
"source_hash": "sha256:91eaf378ebc4f6f3f44bb72a5a56086bb9d6e0b255e49a591c18a4448cfa8ec6",
"source_line_end": 89,
"source_line_start": 68
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "belief_state_process_reward_replay",
"kind": "organ",
"kind_label": "component",
"label": "Belief-State Reward Bundle Replay",
"relation": "explains",
"url": "component-belief-state-process-reward-replay.html"
},
{
"id": "mechanism.belief_state_process_reward_replay.validates_public_belief_state_process_reward_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public belief state process reward replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-belief-state-process-reward-replay-validates-public-belief-state-process-reward-replay"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.belief_state_process_reward_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 134,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/belief_state_process_reward_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Planning and Acting in Partially Observable Stochastic Domains",
"section_id": "three-lineages-belief-state-process-reward-reward-hacking",
"url": "https://people.smp.uq.edu.au/YoniNazarathy/Control4406_2014/resources/KaelblingLittmanCassandra1998.pdf"
},
{
"kind": "external_reference",
"label": "Let's Verify Step by Step",
"section_id": "three-lineages-belief-state-process-reward-reward-hacking",
"url": "https://arxiv.org/abs/2305.20050"
},
{
"kind": "external_reference",
"label": "Concrete Problems in AI Safety",
"section_id": "three-lineages-belief-state-process-reward-reward-hacking",
"url": "https://openai.com/index/concrete-ai-safety-problems/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "This component reads pre-recorded agent steps and asks of every step-level reward whether it reconstructs from lower-level public evidence or merely asserts its own correctness. This component checks a recorded set of agent steps and asks one question of every step-level reward: does it reconstruct from lower-level public evidence, or is it a label asserting its own correctness. It reads pre-recorded files. It never runs or watches a live agent.",
"level": 4,
"order": 1,
"section_id": "every-process-reward-must-reconstruct-from-evidence",
"source_line_end": 6,
"source_line_start": 3,
"title": "Every process reward must reconstruct from evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The input models three partially observable toy tasks, and a reward is admitted only when its references recompute into one coherent chain with all seven planted traps observed. The input models three partially observable toy tasks: a terminal investigation, a mock purchase, and a small formal planner. For each step it carries a public belief summary, the verifier or observed feedback the reward cites, the process reward itself, the trajectory group the step belongs to, the outcome reward for that trajectory, and a cold-replay record. A reward is admitted only when all of those references recompute into one coherent chain and none of seven planted reward-hacking traps go unobserved.",
"level": 4,
"order": 2,
"section_id": "three-toy-tasks-carry-seven-planted-reward-traps",
"source_line_end": 10,
"source_line_start": 7,
"title": "Three toy tasks carry seven planted reward traps"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The result files record per step the belief summary, the reward, and the recompute and trap outcomes, so a reader sees why each reward was allowed without exposing hidden reasoning. The result files record, per step, the belief summary, the reward, and whether the recompute and trap checks passed. So a reader can see why each reward was allowed or blocked, without any hidden reasoning being exposed.",
"level": 4,
"order": 3,
"section_id": "body-free-results-explain-each-admitted-reward",
"source_line_end": 14,
"source_line_start": 11,
"title": "metadata-only results explain each admitted reward"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Process-reward language is easy to assert and hard to verify, since a row can claim good reasoning while resting on a hidden label or gamed formatting a shape linter passes. Process-reward language is easy to assert and hard to verify. A row can claim a step earned a reward \"for good reasoning\" while the real basis is a hidden gold label, a neural-judge guess, or formatting that gamed the scorer. A shape linter does not catch this: every field can be present and well-typed and the claim can still be fake.",
"level": 4,
"order": 5,
"section_id": "process-reward-claims-a-linter-cannot-catch",
"source_line_end": 20,
"source_line_start": 17,
"title": "Process-reward claims a linter cannot catch"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The guarded failure is a reward correct field by field that fails end-to-end recompute, with outcome rewards kept beside process rewards and every reference metadata-only to keep the check honest. The failure mode this component guards against is a reward that looks correct field by field but does not survive being recomputed end to end. A reward cannot point at a belief that points at a different episode, or cite feedback that belongs to another trajectory, and still count. Two design choices keep it honest. Outcome rewards are carried beside process rewards, so a final-answer score cannot be relabelled as step-level evidence. Every belief summary, feedback reference, and reward event is m\u2026",
"level": 4,
"order": 6,
"section_id": "the-end-to-end-recompute-failure-mode",
"source_line_end": 24,
"source_line_start": 21,
"title": "The end-to-end recompute failure mode"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The same process-reward replay runtime supports a negative-bearing fixture mode and a manifest-backed exported-bundle mode. The runtime lives in beliefstateprocessrewardreplay.py and runs in two modes.",
"level": 4,
"order": 8,
"section_id": "one-runtime-exposes-fixture-and-bundle-modes",
"source_line_end": 31,
"source_line_start": 27,
"title": "One runtime exposes fixture and bundle modes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "In fixture mode run loads eight positive files and seven planted negatives before calling buildresult, while bundle mode's runrewardbundle drops the negatives and makes the copied source-body floor mandatory. In fixture mode, run loads the eight positive input files named by INPUTNAMES plus the seven planted negatives named by NEGATIVEINPUTNAMES, with the expected trap-to-code mapping in EXPECTEDNEGATIVECASES. It calls buildresult with includenegative=True, then writereceipts. In bundle mode, runrewardbundle validates an exported bundle with no negative inputs and makes the copied source-body floor mandatory by passing require_manifest.",
"level": 4,
"order": 9,
"section_id": "two-modes-fixture-and-exported-bundle",
"source_line_end": 35,
"source_line_start": 32,
"title": "Two modes: fixture and exported bundle"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult orchestrates the per-file checks, semantic recompute, negative cases, secret scan, trace, and manifest, and passes only when all of them clear, blocking on any single failure. buildresult is the orchestrator. It loads the payloads, runs the per-file checks (validateprojectionprotocol, validaterewardpolicy, validatetaskepisodes, validatebeliefstates, validateverifierfeedback, validaterewardevents, validatetrajectorygroups, validatecoldreplay), runs the cross-file check validatesemanticrecompute, runs validatenegativecases, scans every input path for forbidden classes with scanpaths, builds the public trace, and checks the copied source-body manifest with sourcemodulemanifestresul\u2026",
"level": 4,
"order": 10,
"section_id": "build-result-orchestrates-every-check",
"source_line_end": 39,
"source_line_start": 36,
"title": "`_build_result` orchestrates every check"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatesemanticrecompute is where most of the work happens. It checks that the cited feedback belongs to the same episode, that the process reward references the same belief, episode, trajectory, feedback ref, and belief-discrepancy value, that the trajectory lists both that episode and that reward, that the trajectory's outcome reward is a real outcome event for the same episode, and that the cold replay exists and passed. Any inconsistency appends a precise reason code such as feedbackepisodemismatch, beliefdiscrepancymismatch, or trajectoryprocessreward_missing. One blocked row blocks the whole result. This is the check a label-only fixture cannot fake.",
"level": 4,
"order": 11,
"section_id": "rebuilding-each-belief-to-reward-chain-by-id",
"source_line_end": 43,
"source_line_start": 40,
"title": "Rebuilding each belief-to-reward chain by id"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatenegativecases confirms every planted trap triggers its expected error code, and a trap that fails to fire becomes a missing case that blocks the run. validatenegativecases reads the planted traps and confirms each one triggers its expected error code, for example hiddenchainofthoughtexported raising BELIEFREWARDHIDDENCOTEXPORT and neuraljudgeonly with no verifier reference raising BELIEFREWARDNEURALJUDGEONLY_LABEL. A trap that fails to trigger becomes a missing case and blocks the run.",
"level": 4,
"order": 12,
"section_id": "each-planted-trap-must-raise-its-code",
"source_line_end": 47,
"source_line_start": 44,
"title": "Each planted trap must raise its code"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourcemodulemanifestresult enforces the copied source-body floor, requiring each module to match its class, digest, and source authority while keeping result records metadata-only, raising a specific code on any breach. sourcemodulemanifestresult enforces the copied source-body floor for the exported bundle. Each declared module must classify as copiednonsecretmacrobody, use a body class, copy its body into sourcemodules/ while keeping result record fields metadata-only, exist as a target file, match its declared sha256 digest, and resolve its source reference to live source authority. Digest mismatch, missing target, wrong body class, result record-body leakage, count mismatch, and missing\u2026",
"level": 4,
"order": 13,
"section_id": "copied-non-secret-source-body-floor-for-bundles",
"source_line_end": 51,
"source_line_start": 48,
"title": "Copied source-body floor for bundles"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "writereceipts writes the four result files while resultcard projects a compact card that omits the heavy payload fields, and the trace emits metadata-only spans keeping bodyinreceipt: false. writereceipts writes the result, board, validation, and sign-off files. resultcard projects a compact public card of counts, status, freshness digest, and reference handles. CARDOMITTEDFULLPAYLOADKEYS names the heavy fields kept off the card: findings, scans, trace bodies, row payloads, source imports, the scope limit, and the scope boundary text. The trace itself comes from buildpublicbeliefstateprocessrewardtrace, imported from agentexecutiontrace.py, which emits metadata-only spans and keeps bodyin_r\u2026",
"level": 4,
"order": 14,
"section_id": "receipts-and-the-body-free-public-card",
"source_line_end": 90,
"source_line_start": 52,
"title": "Result records and the metadata-only public card"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "negative-cases",
"source_line_end": 92,
"source_line_start": 91,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The seven planted traps each name an abuse a reward claim might smuggle in and must raise its code, while FORBIDDENKEYS blocks reasoning bodies and model-output data from any row. The seven planted traps are the falsification fixtures. Each names the abuse a process-reward claim might smuggle in, and each must raise its code for the run to pass: hidden chain-of-thought export, neural-judge-only labels, hidden gold labels, reward by formatting, verifier bypass, a benchmark-performance claim, and final-answer-only scoring. FORBIDDENKEYS blocks reasoning bodies, model-output data, gold-answer bodies, live-training run ids, and benchmark submission ids from appearing in any belief or reward row.",
"level": 4,
"order": 16,
"section_id": "seven-falsification-traps-and-forbidden-keys",
"source_line_end": 96,
"source_line_start": 93,
"title": "Seven falsification traps and forbidden keys"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 98,
"source_line_start": 97,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Three older ideas ground the component, partial-observability belief states, step-level process reward, and reward hacking, but it trains no model and only borrows their accounting form. Three older ideas ground this component. The belief-state lineage comes from work on acting under partial observability such as Kaelbling, Littman, and Cassandra's Planning and Acting in Partially Observable Stochastic Domains. The step-level reward lineage follows OpenAI's Let's Verify Step by Step, which separates process feedback from outcome-only supervision. The reward-hacking lineage comes from Concrete Problems in AI Safety. This component does not train a reward model or expose reasoning. It borrows\u2026",
"level": 4,
"order": 18,
"section_id": "three-lineages-belief-state-process-reward-reward-hacking",
"source_line_end": 102,
"source_line_start": 99,
"title": "Three lineages: belief state, process reward, reward hacking"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 104,
"source_line_start": 103,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "The fixture validator runs from the repository root and writes its result to a disposable path outside the working tree. Run the fixture check from the repo root and write its result outside the working tree: Then run the exported bundle check:",
"level": 4,
"order": 20,
"section_id": "fixture-output-stays-outside-the-working-tree",
"source_line_end": 119,
"source_line_start": 105,
"title": "Fixture output stays outside the working tree"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass on the focused regression test means all eight positive files cleared their checks, the chain recomputed, every planted trap fired, and no secret-class or body leak appeared. The focused regression test is tests/testbeliefstateprocessreward_replay.py. A pass means all eight positive files cleared their checks, the chain recomputed, every planted trap triggered its code, and no secret-class or body leak was found.",
"level": 4,
"order": 21,
"section_id": "what-a-passing-regression-run-means",
"source_line_end": 123,
"source_line_start": 120,
"title": "What a passing regression run means"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 125,
"source_line_start": 124,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The evidence supports a metadata-only replay over three synthetic tasks, whose fixed counts are the proof boundary, showing process feedback stays separable from outcome reward and survives cold replay. The evidence supports a metadata-only replay over three synthetic tasks: 3 episodes, 6 belief summaries, 6 process rewards, 3 outcome rewards, 3 trajectory groups, 3 cold replays, 7 observed traps in fixture mode, and 7 copied source bodies in bundle mode. Those counts are the proof boundary, not a scale claim. Within them the component shows that public process feedback stays separable from final-answer outcome reward and survives cold replay with every trap observed.",
"level": 4,
"order": 23,
"section_id": "the-counted-proof-boundary-not-scale",
"source_line_end": 129,
"source_line_start": 126,
"title": "The counted proof boundary, not scale"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It does not estimate calibration, generalize, compare policies, certify training, or score a benchmark, and grants no hidden-reasoning export, live RL, provider behavior, mutation, or launch-scope decision. It does not estimate reward-model calibration, generalize to unseen tasks, compare policies, certify live training, or score a benchmark. The scope limit is source-faithful replay over declared public inputs only. It grants no hidden-reasoning export, no live RL, no neural-judge sufficiency, no hidden-gold benchmark claims, no provider behavior, no source-file changes, and no public sharing or launch-scope decision. Any higher claim must land first in the source record and its generated\u2026",
"level": 4,
"order": 24,
"section_id": "what-the-authority-ceiling-refuses",
"source_line_end": 133,
"source_line_start": 130,
"title": "What the scope limit refuses"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/belief_state_process_reward_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Public projection of a belief-state process-reward claim contract, backed by the agent-execution trace refactor and copied source bodies."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "belief_state_process_reward_replay.py",
"path": "src/microcosm_core/organs/belief_state_process_reward_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for the public replay/regression validator.",
"symbols": [
"run",
"run_reward_bundle",
"_build_result",
"_write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/belief_state_process_reward_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/belief_state_process_reward_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/belief_state_process_reward_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Public Plectis projection of a belief-state process-reward claim contract. Backed by the public agent-execution trace refactor lane plus copied source bodies; validates trajectory groups, cold replay result records, negative cases, and scope limits.",
"summary_line": "Public projection of a belief-state process-reward claim contract, backed by the agent-execution trace refactor and copied source bodies.",
"title": "Belief-State Process Reward Replay"
},
{
"anchor": "paper-module-sleeper-memory-poisoning-quarantine-replay",
"authority_ceiling": "Synthetic replay fixture only; not a live memory product, live user memory import, benchmark security result, private memory export, or launch claim.",
"body_html": "
Persistent agent memory is an attack surface, and this component asks whether a recorded guard-and-cleanup sequence proves the quarantine held on later retrieval and the rollback truly removed the memory. Persistent agent memory is an attack surface. If an agent reads a poisoned source in one session and writes a memory from it, that memory can shape a later session's actions long after the poisoning is out of view. This component asks one question about that story: given a recorded guard-and-cleanup sequence, can you show from the record alone that the quarantine held when the memory was retrieved later, and that the rollback genuinely removed it?
\n
It is a synthetic replay validator, not a memory runtime, grading an on-disk record of four sessions against a policy and emitting metadata-only records over public metadata rows. It is a synthetic replay validator, not a memory runtime. It does not delete a memory or run the agent again. It reads an on-disk record of four sessions, grades every required step and result record against a policy, and emits metadata-only result records that carry refs, hashes, counts, and verdicts. The inputs are public metadata rows. No live user memory, transcript body, or model-output data passes through.
\n
Purpose
\n
The whole chain must line up, not the answer
\n
It refuses to grade only the final answer, requiring the write to carry provenance, the retrieval to be blocked before action citing the same memory, and the rollback to prove absence after rerun. A naive memory-security check grades the final answer: the agent did not act on the bad memory, so it passed. This component refuses that shortcut. It requires the whole chain to be present and to line up. The poisoned write must carry provenance. The later retrieval must be blocked before any action and must cite the same memory the write quarantined, plus that write's audit ref and a cold replay result. The rollback must carry a deletion audit ref, a cold rerun result, and proof that the memory is absent after the rerun. A blocked retrieval that cannot name the quarantine audit ref or the cold replay for the memory it gates is treated as unproven, not as a pass.
\n
How it works
\n
Seven validators merged into one status
\n
_build_result runs seven validators in order and merges their findings, with run driving the fixture and run_quarantine_bundle the exported bundle, both writing metadata-only result records. The runtime is src/microcosm_core/organs/sleeper_memory_poisoning_quarantine_replay.py. _build_result loads the input directory and runs seven validators in a fixed order, then merges their findings and negative-case observations into one status. run drives the first-wave fixture with the eight negative fixtures loaded. run_quarantine_bundle drives the exported public bundle with negatives off and the source manifest required. Both call _write_receipts, and result_card projects a compact card that omits the row-level detail.
\n
Projection, policy, and the four session roles
\n
validate_projection_protocol binds the source refs and pattern id, then validate_memory_policy and validate_session_chain check the trust tiers and the four required session roles. The first check is projection. validate_projection_protocol requires the protocol to cite at least four source refs, the sleeper_memory_poisoning_quarantine_replay_compound pattern id, two projection result refs, three public replacement refs, and non-empty reimplemented and omitted lists, and to declare copied_private_memory_bodies false. Then validate_memory_policy reads the allowed trust tiers and quarantine verdicts, and validate_session_chain checks the four session roles in REQUIRED_SESSION_ROLES: poisoned source seen, memory write quarantined, later retrieval action gated, rollback and cold rerun.
\n
Provenance gate over each memory write
\n
validate_quarantine_events blocks any write that leaks a body, claims live memory, drops provenance, or promotes untrusted context, and passes only with one quarantined write and one admitted control. The provenance gate is validate_quarantine_events, which grades each write through _validate_write_row. A write is blocked when it carries a forbidden key or exports a private body, exports a raw transcript, claims live user memory, omits its provenance or source bundle ref, promotes untrusted context into trusted memory, or gives an audit ref that does not start with audit.. The gate passes only when one row is a quarantined sleeper-poisoning write (verdict quarantine, trust tier untrusted_context, label sleeper_poisoning) and one row is a provenance-bound admitted control.
\n
Delayed-influence gate coupled to the quarantine
\n
validate_retrieval_replays counts a row blocked only when it is gated before action and cites the same memory's quarantine audit and cold replay refs, downgrading any uncoupled row. The delayed-influence gate is validate_retrieval_replays, grading each row through _validate_retrieval_row. It rejects final-answer-only grading and any row where the influence grade or action gate lets quarantined memory reach an action. A row counts as blocked only with action gate blocked_before_action and influence grade quarantined_before_action. This gate is coupled to the first: _quarantined_audit_refs_by_memory_ref maps each quarantined memory to its audit ref, and a blocked row that targets a memory the write never quarantined, or omits that memory's audit ref or its cold replay ref from its evidence, is downgraded to blocked with a typed finding.
\n
Rollback needs audit, rerun, and absence
\n
validate_rollback_rerun requires a deletion audit ref, a rollback and rerun result ref, and memory_absent_after_rerun true, blocking bogus or uncoupled rollbacks. The rollback gate is validate_rollback_rerun through _validate_rollback_row. A rollback needs a deletion audit ref starting audit., a rollback result ref and a rerun result ref both starting receipt., and memory_absent_after_rerun set true. Nonempty but bogus refs, a missing rerun ref, or a memory still present after rerun all block rather than count as evidence. Rollback rows are coupled to the quarantined memory refs the same way retrieval rows are.
\n
Source-body digest checks and metadata-only result records
\n
_source_module_manifest_result verifies each copied public body's class, digests, and anchors, then _write_receipts emits four records that omit retrieval rows, rollback rows, and copied bodies. The last check is source-body handling. _source_module_manifest_result verifies each copied public source body: its import class, a material class, that the body was copied and kept out of the result record, that the target file exists, that its declared digests match the file, that required anchors are present, and, when the live source resolves, that the copy matches it. _source_open_body_import_summary reduces that to a count and a set of classes. _write_receipts then emits four result records (result, board, validation, sign-off) that carry ids, counts, refs, digests, and verdicts while omitting retrieval rows, rollback rows, and copied bodies from the card.
\nDiagram of the mechanism (6 steps).Diagram source & refs
Source refs
poisoned write quarantined + provenance-bound control
validate_quarantine_events
retrieval blocked before action, same memory ref + audit + cold replay
A pass requires all eight named negative rows to surface as typed findings, spanning body export, unprovenanced writes, unaudited deletion, and unmetered influence, or the run blocks. The fixture carries eight named negative rows, one per entry in EXPECTED_NEGATIVE_CASES. A pass requires every one to surface as a typed finding, and _build_result blocks the run for any expected case that was not observed. The cases are private memory body export, live user memory claim, raw transcript export, memory write without provenance, trusted promotion from untrusted context, deletion without audit, final-answer-only grading, and unmetered poison influence. Their error codes carry the SLEEPER_MEMORY_ prefix.
\n
Prior Art Grounding
\n
Sleeper-trigger and memory-poisoning literature
\n
The design draws on the Sleeper Agents and memory-poisoning literature, borrowing the control shape where writes need provenance, retrieval passes an influence gate, and deletion needs audit and rerun. This component draws two lines of prior work. The sleeper-trigger line is Anthropic's Sleeper Agents. The memory-poisoning line includes AgentPoison, MemoryGraft, and Hidden in Memory, which all treat retrieved or persistent agent memory as an attack surface rather than a neutral cache. The component borrows the control shape from that work: memory writes need provenance, untrusted context cannot silently become authority, later retrieval must pass an influence gate, and deletion needs an audit ref plus cold rerun. It does not secure a live memory system.
\n
Validation Result record Path
\n
Run the first-wave fixture validator from the repo root and write its record outside the working tree: Run the first-wave fixture validator from the repo root and write its record outside the working tree:
A pass means the recorded chain and its eight negative cases lined up over public metadata refs and the non-public-state scan found nothing. A pass means the recorded chain and its eight negative cases lined up over public metadata refs, and the non-public-state scan found nothing. These are the validation result records for this module.
\n
Scope boundary
\n
Scope limit
\n
Narrow claim and the scope limit
\n
The narrow claim is that over one synthetic record the quarantine, blocked retrieval, and audited rollback are present and coupled and eight cases caught, establishing no live-memory quality, provider behaviour, or launch-scope decision. The strongest claim the evidence supports is narrow: over one synthetic record of public metadata refs, the quarantine, the blocked-before-action retrieval, and the audited rollback with a clean cold rerun are all present and internally coupled, and eight named failure cases are each caught. The proof boundary is the set of refs and verdicts in that record: source bundle refs, provenance refs, quarantine verdicts, classifier labels, retrieval influence gates, rollback audit refs, cold rerun result refs, the negative cases, and the source-body digest checks. The scope limit is public_body_free_sleeper_memory_quarantine_policy_projection_only, matching AUTHORITY_CEILING in the source. It does not establish live memory product quality, live user-memory handling, trusted promotion from untrusted context, provider behaviour, source-file changes, benchmark security, private memory export, public sharing, or launch-scope decision.
\n",
"components": [
{
"anchor": "component-sleeper_memory_poisoning_quarantine_replay",
"component_id": "sleeper_memory_poisoning_quarantine_replay",
"public_label": "Memory-Poisoning Quarantine Policy Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 2
},
"link_count": 9,
"links": [
{
"id": "sleeper_memory_poisoning_quarantine_replay",
"kind": "organ",
"kind_label": "component",
"label": "Memory-Poisoning Quarantine Policy Replay",
"relation": "explains",
"url": "component-sleeper-memory-poisoning-quarantine-replay.html"
},
{
"id": "mechanism.sleeper_memory_poisoning_quarantine_replay.validates_public_sleeper_memory_poisoning_quarantine_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public sleeper memory poisoning quarantine replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-sleeper-memory-poisoning-quarantine-replay-validates-public-sleeper-memory-poisoning-quarantine-replay"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-14",
"kind": "principle",
"kind_label": "principle",
"label": "Carry basis and provenance together",
"relation": "governed by",
"url": "doctrine.html#dcard-p-14"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.agent_memory_temporal_conflict_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Memory Temporal-Conflict Replay",
"relation": "depends on",
"url": "paper-module-agent-memory-temporal-conflict-replay.html"
},
{
"id": "paper_module.mcp_tool_authority_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "MCP Tool Authority Replay",
"relation": "depends on",
"url": "paper-module-mcp-tool-authority-replay.html"
},
{
"id": "paper_module.agent_sandbox_policy_escape_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Sandbox Policy-Escape Replay",
"relation": "depends on",
"url": "paper-module-agent-sandbox-policy-escape-replay.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.sleeper_memory_poisoning_quarantine_replay",
"reader_packet": {
"abstract": "Synthetic replay fixture for a persistent-memory security contract: quarantine of poisoned memory, audit refs, rerun result records, negative cases, authority",
"authority_ceiling": "Synthetic replay fixture only; not a live memory product, live user memory import, benchmark security result, private memory export, or launch claim.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.sleeper_memory_poisoning_quarantine_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-sleeper-memory-poisoning-quarantine-replay-diagram-1",
"html_url": "docs/paper-module-sleeper-memory-poisoning-quarantine-replay.html#paper-module-sleeper-memory-poisoning-quarantine-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD inputs[\"Public metadata inputs sessions, write proposals, retrieval replays, rollback rows\"] provenance[\"validate_quarantine_events poisoned write quarantined + provenance-bound control\"] influence[\"validate_retrieval_replays retrieval blocked before action, same memory ref + audit + cold replay\"] rollback[\"validate_rollback_rerun deletion audit + rerun ref + memory absent after rerun\"] bodies[\"_source_module_manifest_result copied bodies digest-checked, records stay metadata-only\"] records[\"metadata-only result records refs, hashes, counts, verdicts\"] inputs --> provenance provenance -->|quarantined memory ref| influence influence --> rollback rollback --> bodies bodies --> records",
"source_format": "mermaid",
"source_hash": "sha256:70af73bf5f0051fbb22188e2acfe397e4ce5ff583bd81fc083d87023f7224015",
"source_line_end": 134,
"source_line_start": 121
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.agent_memory_temporal_conflict_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Memory Temporal-Conflict Replay",
"relation": "depends on",
"url": "paper-module-agent-memory-temporal-conflict-replay.html"
},
{
"id": "paper_module.mcp_tool_authority_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "MCP Tool Authority Replay",
"relation": "depends on",
"url": "paper-module-mcp-tool-authority-replay.html"
},
{
"id": "paper_module.agent_sandbox_policy_escape_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Sandbox Policy-Escape Replay",
"relation": "depends on",
"url": "paper-module-agent-sandbox-policy-escape-replay.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "sleeper_memory_poisoning_quarantine_replay",
"kind": "organ",
"kind_label": "component",
"label": "Memory-Poisoning Quarantine Policy Replay",
"relation": "explains",
"url": "component-sleeper-memory-poisoning-quarantine-replay.html"
},
{
"id": "mechanism.sleeper_memory_poisoning_quarantine_replay.validates_public_sleeper_memory_poisoning_quarantine_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public sleeper memory poisoning quarantine replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-sleeper-memory-poisoning-quarantine-replay-validates-public-sleeper-memory-poisoning-quarantine-replay"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-14",
"kind": "principle",
"kind_label": "principle",
"label": "Carry basis and provenance together",
"relation": "governed by",
"url": "doctrine.html#dcard-p-14"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.sleeper_memory_poisoning_quarantine_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 220,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/sleeper_memory_poisoning_quarantine_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Sleeper Agents",
"section_id": "sleeper-trigger-and-memory-poisoning-literature",
"url": "https://arxiv.org/abs/2401.05566"
},
{
"kind": "external_reference",
"label": "AgentPoison",
"section_id": "sleeper-trigger-and-memory-poisoning-literature",
"url": "https://arxiv.org/abs/2407.12784"
},
{
"kind": "external_reference",
"label": "MemoryGraft",
"section_id": "sleeper-trigger-and-memory-poisoning-literature",
"url": "https://arxiv.org/abs/2512.16962"
},
{
"kind": "external_reference",
"label": "Hidden in Memory",
"section_id": "sleeper-trigger-and-memory-poisoning-literature",
"url": "https://arxiv.org/abs/2605.15338"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Persistent agent memory is an attack surface, and this component asks whether a recorded guard-and-cleanup sequence proves the quarantine held on later retrieval and the rollback truly removed the memory.** Persistent agent memory is an attack surface. If an agent reads a poisoned source in one session and writes a memory from it, that memory can shape a later session's actions long after the poisoning is out of view. This component asks one question about that story: given a recorded guard-and-cleanup sequence, can you show from the record alone that the quarantine held when the memory was retrieved later, and that the rollback genuinely removed it?",
"level": 4,
"order": 1,
"section_id": "persistent-agent-memory-is-an-attack-surface",
"source_line_end": 13,
"source_line_start": 3,
"title": "Persistent agent memory is an attack surface"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is a synthetic replay validator, not a memory runtime, grading an on-disk record of four sessions against a policy and emitting metadata-only records over public metadata rows.** It is a synthetic replay validator, not a memory runtime. It does not delete a memory or run the agent again. It reads an on-disk record of four sessions, grades every required step and result record against a policy, and emits metadata-only result records that carry refs, hashes, counts, and verdicts. The inputs are public metadata rows. No live user memory, transcript body, or model-output data passes through.",
"level": 4,
"order": 2,
"section_id": "it-is-a-synthetic-replay-validator",
"source_line_end": 24,
"source_line_start": 14,
"title": "It is a synthetic replay validator"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 3,
"section_id": "purpose",
"source_line_end": 26,
"source_line_start": 25,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It refuses to grade only the final answer, requiring the write to carry provenance, the retrieval to be blocked before action citing the same memory, and the rollback to prove absence after rerun.** A naive memory-security check grades the final answer: the agent did not act on the bad memory, so it passed. This component refuses that shortcut. It requires the whole chain to be present and to line up. The poisoned write must carry provenance. The later retrieval must be blocked before any action and must cite the same memory the write quarantined, plus that write's audit ref and a cold replay result. The rollback must carry a deletion audit ref, a cold rerun result, and proof that the mem\u2026",
"level": 4,
"order": 4,
"section_id": "the-whole-chain-must-line-up-not-the-answer",
"source_line_end": 40,
"source_line_start": 27,
"title": "The whole chain must line up, not the answer"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 5,
"section_id": "how-it-works",
"source_line_end": 42,
"source_line_start": 41,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult runs seven validators in order and merges their findings, with run driving the fixture and runquarantinebundle the exported bundle, both writing metadata-only result records. The runtime is src/microcosmcore/components/sleepermemorypoisoningquarantinereplay.py. buildresult loads the input directory and runs seven validators in a fixed order, then merges their findings and negative-case observations into one status. run drives the first-wave fixture with the eight negative fixtures loaded. runquarantinebundle drives the exported public bundle with negatives off and the source manifest required. Both call writereceipts, and result_card projects a compact card that omits the row-le\u2026",
"level": 4,
"order": 6,
"section_id": "seven-validators-merged-into-one-status",
"source_line_end": 54,
"source_line_start": 43,
"title": "Seven validators merged into one status"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateprojection_protocol binds the source refs and pattern id, then validatememorypolicy and validatesessionchain check the trust tiers and the four required session roles. The first check is projection. validateprojection_protocol requires the protocol to cite at least four source refs, the sleepermemorypoisoningquarantinereplay_compound pattern id, two projection result refs, three public replacement refs, and non-empty reimplemented and omitted lists, and to declare copiedprivatememory_bodies false. Then validatememorypolicy reads the allowed trust tiers and quarantine verdicts, and validatesessionchain checks the four session roles in REQUIREDSESSIONROLES: poisoned source seen, memor\u2026",
"level": 4,
"order": 7,
"section_id": "projection-policy-and-the-four-session-roles",
"source_line_end": 68,
"source_line_start": 55,
"title": "Projection, policy, and the four session roles"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatequarantine_events blocks any write that leaks a body, claims live memory, drops provenance, or promotes untrusted context, and passes only with one quarantined write and one admitted control. The provenance gate is validatequarantine_events, which grades each write through validatewrite_row. A write is blocked when it carries a forbidden key or exports a private body, exports a raw transcript, claims live user memory, omits its provenance or source bundle ref, promotes untrusted context into trusted memory, or gives an audit ref that does not start with audit.. The gate passes only when one row is a quarantined sleeper-poisoning write (verdict quarantine, trust tier untrustedcontext\u2026",
"level": 4,
"order": 8,
"section_id": "provenance-gate-over-each-memory-write",
"source_line_end": 81,
"source_line_start": 69,
"title": "Provenance gate over each memory write"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateretrieval_replays counts a row blocked only when it is gated before action and cites the same memory's quarantine audit and cold replay refs, downgrading any uncoupled row. The delayed-influence gate is validateretrieval_replays, grading each row through validateretrieval_row. It rejects final-answer-only grading and any row where the influence grade or action gate lets quarantined memory reach an action. A row counts as blocked only with action gate blockedbeforeaction and influence grade quarantinedbeforeaction. This gate is coupled to the first: quarantinedauditrefsbymemoryref maps each quarantined memory to its audit ref, and a blocked row that targets a memory the write never q\u2026",
"level": 4,
"order": 9,
"section_id": "delayed-influence-gate-coupled-to-the-quarantine",
"source_line_end": 95,
"source_line_start": 82,
"title": "Delayed-influence gate coupled to the quarantine"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validaterollback_rerun requires a deletion audit ref, a rollback and rerun result ref, and memoryabsentafter_rerun true, blocking bogus or uncoupled rollbacks. The rollback gate is validaterollbackrerun through validaterollbackrow. A rollback needs a deletion audit ref starting audit., a rollback result ref and a rerun result ref both starting result record., and memoryabsentafter_rerun set true. Nonempty but bogus refs, a missing rerun ref, or a memory still present after rerun all block rather than count as evidence. Rollback rows are coupled to the quarantined memory refs the same way retrieval rows are.",
"level": 4,
"order": 10,
"section_id": "rollback-needs-audit-rerun-and-absence",
"source_line_end": 106,
"source_line_start": 96,
"title": "Rollback needs audit, rerun, and absence"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "sourcemodulemanifestresult verifies each copied public body's class, digests, and anchors, then writereceipts emits four records that omit retrieval rows, rollback rows, and copied bodies. The last check is source-body handling. sourcemodulemanifest_result verifies each copied public source body: its import class, a material class, that the body was copied and kept out of the result record, that the target file exists, that its declared digests match the file, that required anchors are present, and, when the live source resolves, that the copy matches it. sourceopenbodyimport_summary reduces that to a count and a set of classes. writereceipts then emits four result records (result, board, v\u2026",
"level": 4,
"order": 11,
"section_id": "source-body-digest-checks-and-body-free-receipts",
"source_line_end": 135,
"source_line_start": 107,
"title": "Source-body digest checks and metadata-only result records"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 137,
"source_line_start": 136,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass requires all eight named negative rows to surface as typed findings, spanning body export, unprovenanced writes, unaudited deletion, and unmetered influence, or the run blocks.** The fixture carries eight named negative rows, one per entry in EXPECTEDNEGATIVECASES. A pass requires every one to surface as a typed finding, and buildresult blocks the run for any expected case that was not observed. The cases are private memory body export, live user memory claim, raw transcript export, memory write without provenance, trusted promotion from untrusted context, deletion without audit, final-answer-only grading, and unmetered poison influence. Their error codes carry the SLEEPERMEMORY pr\u2026",
"level": 4,
"order": 13,
"section_id": "eight-negatives-that-must-each-surface",
"source_line_end": 149,
"source_line_start": 138,
"title": "Eight negatives that must each surface"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "prior-art-grounding",
"source_line_end": 151,
"source_line_start": 150,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The design draws on the Sleeper Agents and memory-poisoning literature, borrowing the control shape where writes need provenance, retrieval passes an influence gate, and deletion needs audit and rerun.** This component draws two lines of prior work. The sleeper-trigger line is Anthropic's Sleeper Agents. The memory-poisoning line includes AgentPoison, MemoryGraft, and Hidden in Memory, which all treat retrieved or persistent agent memory as an attack surface rather than a neutral cache. The component borrows the control shape from that work: memory writes need provenance, untrusted context cannot silently become authority, later retrieval must pass an influence gate, and deletion needs an\u2026",
"level": 4,
"order": 15,
"section_id": "sleeper-trigger-and-memory-poisoning-literature",
"source_line_end": 166,
"source_line_start": 152,
"title": "Sleeper-trigger and memory-poisoning literature"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "validation-receipt-path",
"source_line_end": 168,
"source_line_start": 167,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run the first-wave fixture validator from the repo root and write its record outside the working tree: Run the first-wave fixture validator from the repo root and write its record outside the working tree: Then run the exported bundle validator: The focused test and corpus projection check are:",
"level": 4,
"order": 17,
"section_id": "run-the-first-wave-fixture-validator-from-the-repo-root",
"source_line_end": 191,
"source_line_start": 169,
"title": "Run the first-wave fixture validator from the repo root"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the recorded chain and its eight negative cases lined up over public metadata refs and the non-public-state scan found nothing.** A pass means the recorded chain and its eight negative cases lined up over public metadata refs, and the non-public-state scan found nothing. These are the validation result records for this module.",
"level": 4,
"order": 18,
"section_id": "what-a-pass-over-the-record-means",
"source_line_end": 198,
"source_line_start": 192,
"title": "What a pass over the record means"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "claim-ceiling",
"source_line_end": 200,
"source_line_start": 199,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The narrow claim is that over one synthetic record the quarantine, blocked retrieval, and audited rollback are present and coupled and eight cases caught, establishing no live-memory quality, provider behaviour, or launch-scope decision.** The strongest claim the evidence supports is narrow: over one synthetic record of public metadata refs, the quarantine, the blocked-before-action retrieval, and the audited rollback with a clean cold rerun are all present and internally coupled, and eight named failure cases are each caught. The proof boundary is the set of refs and verdicts in that record: source bundle refs, provenance refs, quarantine verdicts, classifier labels, retrieval influence\u2026",
"level": 4,
"order": 20,
"section_id": "narrow-claim-and-the-authority-ceiling",
"source_line_end": 219,
"source_line_start": 201,
"title": "Narrow claim and the scope limit"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/sleeper_memory_poisoning_quarantine_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Synthetic replay fixture for a persistent-memory security contract: quarantine of poisoned memory, audit refs, rerun result records, negative cases, authority"
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "sleeper_memory_poisoning_quarantine_replay.py",
"path": "src/microcosm_core/organs/sleeper_memory_poisoning_quarantine_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for the public replay/regression validator.",
"symbols": [
"run",
"run_quarantine_bundle",
"_build_result",
"_write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/sleeper_memory_poisoning_quarantine_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/sleeper_memory_poisoning_quarantine_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/sleeper_memory_poisoning_quarantine_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Public Plectis projection of a persistent-memory security claim contract. Replays synthetic memory-poisoning episodes and validates quarantine behavior, audit refs, rerun result records, negative cases, and scope limits with metadata-only result records.",
"summary_line": "Synthetic replay fixture for a persistent-memory security contract: quarantine of poisoned memory, audit refs, rerun result records, negative cases, authority",
"title": "Sleeper Memory Poisoning Quarantine Replay"
},
{
"anchor": "paper-module-indirect-prompt-injection-information-flow-policy-replay",
"authority_ceiling": "Source-faithful refactored fixtures and metadata-only result records only; not a live information-flow product, complete-security proof, or launch claim.",
"body_html": "
Indirect prompt injection is untrusted text an agent obeys as an instruction; this component replays one synthetic episode to check untrusted text stayed data and reached no privileged action before a policy gate ruled. An agent that reads web pages, tool output, or retrieved documents takes in text from sources it does not control. Indirect prompt injection is the case where that untrusted text carries an instruction and the agent acts on it as if the operator had asked. This component replays one recorded synthetic agent episode and checks a single property: untrusted text stayed labelled as data, and no untrusted source reached a privileged action before a policy gate ruled on that flow.
\n
It runs over metadata-only rows of refs, labels, and counts rather than raw prompts or tool bodies, recomputing the trust decisions and recording which source fed which sink and what the policy decided. It runs over metadata-only rows. Every source, flow, verdict, and output is a small JSON record of refs, labels, and counts, never a raw prompt, a real account, or a tool-output body. The runtime reads those rows, recomputes the trust decisions itself, and writes a result record a reader can inspect: which source fed which sink, what the policy decided before the action, and that no secret or trusted context leaked.
\n
The component does not defend a live agent; it is a regression check over a fixed fixture and one exported bundle, and a pass means only that the recorded trace satisfies the flow-policy contract. The component does not defend a live agent. It is a regression check over a fixed fixture and one exported example bundle. A pass means the recorded trace satisfies the flow-policy contract, and nothing beyond that.
\n
Purpose
\n
A trace can be labelled to look safe
\n
The guarded failure is a trace that looks safe because its labels were written to look safe, which a field-by-field check waves through while the boundary never held. The failure this guards against is a trace that looks safe because the labels were written to look safe. A field-by-field check trusts whatever the fixture declares. If a row relabels an untrusted web source as clean, or marks a dangerous action allow, a loose check waves it through and the trace passes while the boundary never held.
\n
The runtime recomputes, never trusts labels
\n
The runtime ignores declared taint labels and verdicts and recomputes them by propagating taint along the source-to-sink graph, so a flow cannot quietly relabel an untrusted source or downgrade a dangerous action. So the runtime does not trust the declared labels. Each flow row claims a set of taint labels and a policy verdict, and the runtime ignores both and recomputes them. It propagates taint along the source-to-sink graph from the labelled sources, so a sink inherits the taint of everything that fed it. It then derives the verdict from that propagated taint plus the sink's privilege, the sanitizer state, the sink kind, and the proposed action. When a declared label or verdict disagrees with the derived one, the row is blocked. A flow cannot quietly relabel an untrusted source or downgrade a dangerous action, because the contradiction is recomputed rather than read back.
\n
How it works
\n
_build_result runs the validators in order
\n
_build_result loads the seven inputs, runs a secret-exclusion scan, builds the public trace spans, and folds the ordered validators into one result, with run enabling negative cases and the bundle path leaving them out._build_result is the pipeline. It loads the seven input files, runs a secret-exclusion scan over the input paths, builds the public trace spans, then calls the validators in order and folds their statuses into one result. run drives fixture mode with the negative cases enabled. run_prompt_injection_bundle drives the exported example bundle and leaves the fixture-only negative cases out. Both write metadata-only records through _write_receipts and can emit a compact card through result_card.
\n
Function
What it checks
validate_source_documents
Each source row carries an id, trust label, channel, taint labels, redaction, and a synthetic flag; an untrusted source may not claim instruction authority; the set needs at least three untrusted and two trusted sources.
_taint_propagation_receipt
Propagates taint from the labelled sources along the source-to-sink graph, derives each flow's taint via _derived_flow_taint_labels and its verdict via _derived_policy_verdict, and records a finding when the declared taint or verdict differs from the derived one.
validate_information_flow_graph
Joins each flow to its source row, runs the propagation result record, and rejects an untrusted privileged flow marked allow without sanitization.
_live_tool_call_trace_promotion
Takes one generated tool-call span, wraps it as an untrusted_tool_output source flowing into a privileged sink, and pushes it through the same propagation machinery as evidence that tool output is treated as data until a gate reviews it.
validate_policy_verdicts
Each verdict joins a flow, precedes the action, cites rules, stays redacted, and equals the derived flow verdict; the set must cover allow, warn, block, and review.
validate_sanitized_outputs
Each output discloses no trusted context and obeys no untrusted instruction; a blocked flow may not attempt an external action.
validate_cold_replay
Each replay row names a command and a result record ref that reproduce the verdict and the sanitized output with no trusted-context disclosure.
validate_negative_cases
Turns each deliberately-bad fixture into its expected error code; fixture mode must observe all eight.
\n
The fixed verdict ladder
\n
_derived_policy_verdict is a small fixed ladder that blocks hidden or instruction-channel promotion and unsanitized untrusted taint into privileged sinks, reviews or warns the sanitized cases, and allows everything else. The verdict ladder in _derived_policy_verdict is small and fixed. A flow that promotes hidden or instruction-channel text is blocked. Untrusted taint into a privileged sink is blocked when unsanitized and set to review when sanitized. Untrusted taint into a non-privileged sink warns if sanitized and reviews if not. Everything else is allowed. validate_projection_protocol and validate_injection_policy sit alongside these: the policy must define all four verdicts and deny every authority overclaim, and the projection protocol must cite the trace source and target refs and keep bodies out of the record.
\nDiagram of the mechanism (10 steps).Diagram source & refs
flowchart TD sources["Source rows trusted and untrusted, with taint labels"] flows["Source-to-sink flow rows declared taint and verdict"] propagate["Propagate taint along the source-to-sink graph"] derive["Derive verdict from taint, privilege, sanitizer, sink kind, action"] compare{"Declared labels and verdict match the derived ones?"} blocked["Block the row"] verdicts["Pre-action verdict allow / warn / review / block"] outputs["Sanitized output no trusted context, no obeyed instruction"] toolspan["One tool-call span as untrusted tool output"] record["metadata-only result record"] sources --> flows flows --> propagate toolspan --> propagate propagate --> derive derive --> compare compare -- no --> blocked compare -- yes --> verdicts verdicts --> outputs outputs --> record blocked --> record
\n
What the fixture and bundle carry
\n
The fixture carries five sources, five flows covering all four verdicts, five outputs, five cold replays, one promoted tool-call trace, and eight negatives, while the bundle adds a digest-checked source-module manifest. The fixture carries five source documents, three untrusted and two trusted, five source-to-sink flows whose verdicts cover allow, warn, block, and review, five sanitized outputs, five cold replay passes, five public trace spans, one promoted tool-call trace, and eight negative cases. The exported bundle also ships a source_module_manifest.json, and _source_module_manifest_result checks that its copied source-module bodies are classified as copied_non_secret_macro_body, keep body text out of the record, match the declared count, and resolve each path and target_ref to the same file with a matching SHA-256 digest.
\n
Negative cases
\n
Eight falsification fixtures as the boundary
\n
The eight EXPECTED_NEGATIVE_CASES are the admission boundary, each a deliberately-bad flow the runtime must reject, and fixture mode is blocked if any goes unobserved. The eight fixtures in EXPECTED_NEGATIVE_CASES are the admission boundary. validate_negative_cases must reject each one: real account material, secret or trusted-context exfiltration, raw prompt body export, untrusted tool output treated as instruction authority, hidden system-message promotion, account secret exfiltration, final-answer-only success, and ungated untrusted flow into a privileged sink. These are falsification fixtures, part of the contract, not samples of live exploit traffic. Fixture mode is blocked if any expected case goes unobserved.
\n
Prior Art Grounding
\n
Indirect-injection and information-flow lineage
\n
The threat shape follows Greshake et al., AgentDojo, and tool-calling data-leakage work, with the policy mechanism drawing on dynamic information-flow and taint-tracking, turned here into a metadata-only replay with negative cases. The threat shape follows the indirect-injection framing of Greshake et al. in Not what you've signed up for, the agentic evaluation setting of AgentDojo, and the tool-calling data-leakage results in Simple Prompt Injection Attacks Can Leak Personal Data. The policy mechanism draws on dynamic information-flow and taint-tracking, including Permissive Information-Flow Analysis for Large Language Models. The engineering lesson taken from that work is narrow: untrusted content is labelled as data, source-to-sink flows are visible before a privileged action, and sanitized outputs carry evidence. This component turns that lesson into a metadata-only replay with explicit scope limits and negative cases.
\n
Validation Result record Path
\n
Fixture validation outside the working tree
\n
Run the fixture validator from the repo root and write its record outside the working tree: Run the fixture validator from the repo root and write its record outside the working tree:
A pass means the synthetic trace reproduced its verdicts and sanitized outputs, observed the expected negative cases, and wrote no source body, not general robustness or live account safety. A pass means the synthetic trace reproduced its verdicts and sanitized outputs, observed the expected negative cases, and wrote no source body. It does not mean general prompt-injection robustness or live account safety.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
Over declared synthetic metadata-only rows the replay recomputes source trust, taint propagation, pre-action verdicts, sanitized-output constraints, cold replay, and copied source-module digests, exposing the results without leaking a body. The strongest claim the evidence supports is this: over declared synthetic metadata-only rows, the replay recomputes source trust, taint propagation, pre-action verdicts, sanitized-output constraints, cold replay, and the copied source-module digests, and it exposes those results without leaking a body. The copied-body digest proves byte parity for the named files only; it does not widen the replay into live source authority.
\n
What the replay does not claim
\n
The proof boundary is that fixed input set and the result records, not general prompt-injection robustness, live account or tool safety, raw body export, benchmark claims, or launch-scope decision. The proof boundary is that fixed input set and the validation result records listed above, not this prose. Its scope limit stops there. It does not claim general prompt-injection robustness, live account safety, live tool or provider behaviour, raw prompt or system or tool body export, account secret-bearing data, hidden-message handling in production, benchmark claims, source-file changes, public sharing, hosting, launch-scope decision, or complete security.
\n",
"components": [
{
"anchor": "component-indirect_prompt_injection_information_flow_policy_replay",
"component_id": "indirect_prompt_injection_information_flow_policy_replay",
"public_label": "Prompt-Injection Flow-Policy Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 2
},
"link_count": 7,
"links": [
{
"id": "indirect_prompt_injection_information_flow_policy_replay",
"kind": "organ",
"kind_label": "component",
"label": "Prompt-Injection Flow-Policy Replay",
"relation": "explains",
"url": "component-indirect-prompt-injection-information-flow-policy-replay.html"
},
{
"id": "mechanism.indirect_prompt_injection_information_flow_policy_replay.validates_public_indirect_prompt_injection_information_flow_policy_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public indirect prompt injection information flow policy replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-indirect-prompt-injection-information-flow-policy-replay-validates-public-indirect-prompt-injection-information-flow-policy-replay"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-14",
"kind": "principle",
"kind_label": "principle",
"label": "Carry basis and provenance together",
"relation": "governed by",
"url": "doctrine.html#dcard-p-14"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.indirect_prompt_injection_information_flow_policy_replay",
"reader_packet": {
"abstract": "Validator-backed claim: a source-faithful trace refactor separated trusted instructions from untrusted web/tool/browser text before any privileged action.",
"authority_ceiling": "Source-faithful refactored fixtures and metadata-only result records only; not a live information-flow product, complete-security proof, or launch claim.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.indirect_prompt_injection_information_flow_policy_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-indirect-prompt-injection-information-flow-policy-replay-diagram-1",
"html_url": "docs/paper-module-indirect-prompt-injection-information-flow-policy-replay.html#paper-module-indirect-prompt-injection-information-flow-policy-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD sources[\"Source rows trusted and untrusted, with taint labels\"] flows[\"Source-to-sink flow rows declared taint and verdict\"] propagate[\"Propagate taint along the source-to-sink graph\"] derive[\"Derive verdict from taint, privilege, sanitizer, sink kind, action\"] compare{\"Declared labels and verdict match the derived ones?\"} blocked[\"Block the row\"] verdicts[\"Pre-action verdict allow / warn / review / block\"] outputs[\"Sanitized output no trusted context, no obeyed instruction\"] toolspan[\"One tool-call span as untrusted tool output\"] record[\"metadata-only result record\"] sources --> flows flows --> propagate toolspan --> propagate propagate --> derive derive --> compare compare -- no --> blocked compare -- yes --> verdicts verdicts --> outputs outputs --> record blocked --> record",
"source_format": "mermaid",
"source_hash": "sha256:125feca7e654a3d773a0010d368be00043c70e42563f4110f15dcdcd2be516f3",
"source_line_end": 121,
"source_line_start": 99
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "indirect_prompt_injection_information_flow_policy_replay",
"kind": "organ",
"kind_label": "component",
"label": "Prompt-Injection Flow-Policy Replay",
"relation": "explains",
"url": "component-indirect-prompt-injection-information-flow-policy-replay.html"
},
{
"id": "mechanism.indirect_prompt_injection_information_flow_policy_replay.validates_public_indirect_prompt_injection_information_flow_policy_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public indirect prompt injection information flow policy replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-indirect-prompt-injection-information-flow-policy-replay-validates-public-indirect-prompt-injection-information-flow-policy-replay"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-14",
"kind": "principle",
"kind_label": "principle",
"label": "Carry basis and provenance together",
"relation": "governed by",
"url": "doctrine.html#dcard-p-14"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.indirect_prompt_injection_information_flow_policy_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 228,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/indirect_prompt_injection_information_flow_policy_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Not what you've signed up for",
"section_id": "indirect-injection-and-information-flow-lineage",
"url": "https://arxiv.org/abs/2302.12173"
},
{
"kind": "external_reference",
"label": "AgentDojo",
"section_id": "indirect-injection-and-information-flow-lineage",
"url": "https://arxiv.org/abs/2406.13352"
},
{
"kind": "external_reference",
"label": "Simple Prompt Injection Attacks Can Leak Personal Data",
"section_id": "indirect-injection-and-information-flow-lineage",
"url": "https://arxiv.org/abs/2506.01055"
},
{
"kind": "external_reference",
"label": "Permissive Information-Flow Analysis for Large Language Models",
"section_id": "indirect-injection-and-information-flow-lineage",
"url": "https://arxiv.org/abs/2410.03055"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Indirect prompt injection is untrusted text an agent obeys as an instruction; this component replays one synthetic episode to check untrusted text stayed data and reached no privileged action before a policy gate ruled.** An agent that reads web pages, tool output, or retrieved documents takes in text from sources it does not control. Indirect prompt injection is the case where that untrusted text carries an instruction and the agent acts on it as if the operator had asked. This component replays one recorded synthetic agent episode and checks a single property: untrusted text stayed labelled as data, and no untrusted source reached a privileged action before a policy gate ruled on that f\u2026",
"level": 4,
"order": 1,
"section_id": "indirect-prompt-injection-is-untrusted-text-an-agent-obeys-as-an-instruction",
"source_line_end": 14,
"source_line_start": 3,
"title": "Indirect prompt injection is untrusted text an agent obeys as an instruction"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It runs over metadata-only rows of refs, labels, and counts rather than raw prompts or tool bodies, recomputing the trust decisions and recording which source fed which sink and what the policy decided.** It runs over metadata-only rows. Every source, flow, verdict, and output is a small JSON record of refs, labels, and counts, never a raw prompt, a real account, or a tool-output body. The runtime reads those rows, recomputes the trust decisions itself, and writes a result record a reader can inspect: which source fed which sink, what the policy decided before the action, and that no secret or trusted context leaked.",
"level": 4,
"order": 2,
"section_id": "it-runs-over-body-free-rows-of-refs",
"source_line_end": 25,
"source_line_start": 15,
"title": "It runs over metadata-only rows of refs"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component does not defend a live agent; it is a regression check over a fixed fixture and one exported bundle, and a pass means only that the recorded trace satisfies the flow-policy contract.** The component does not defend a live agent. It is a regression check over a fixed fixture and one exported example bundle. A pass means the recorded trace satisfies the flow-policy contract, and nothing beyond that.",
"level": 4,
"order": 3,
"section_id": "the-component-does-not-defend-a-live-agent",
"source_line_end": 33,
"source_line_start": 26,
"title": "The component does not defend a live agent"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 35,
"source_line_start": 34,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The guarded failure is a trace that looks safe because its labels were written to look safe, which a field-by-field check waves through while the boundary never held.** The failure this guards against is a trace that looks safe because the labels were written to look safe. A field-by-field check trusts whatever the fixture declares. If a row relabels an untrusted web source as clean, or marks a dangerous action allow, a loose check waves it through and the trace passes while the boundary never held.",
"level": 4,
"order": 5,
"section_id": "a-trace-can-be-labelled-to-look-safe",
"source_line_end": 45,
"source_line_start": 36,
"title": "A trace can be labelled to look safe"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The runtime ignores declared taint labels and verdicts and recomputes them by propagating taint along the source-to-sink graph, so a flow cannot quietly relabel an untrusted source or downgrade a dangerous action.** So the runtime does not trust the declared labels. Each flow row claims a set of taint labels and a policy verdict, and the runtime ignores both and recomputes them. It propagates taint along the source-to-sink graph from the labelled sources, so a sink inherits the taint of everything that fed it. It then derives the verdict from that propagated taint plus the sink's privilege, the sanitizer state, the sink kind, and the proposed action. When a declared label or verdict disag\u2026",
"level": 4,
"order": 6,
"section_id": "the-runtime-recomputes-never-trusts-labels",
"source_line_end": 59,
"source_line_start": 46,
"title": "The runtime recomputes, never trusts labels"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 61,
"source_line_start": 60,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "buildresult loads the seven inputs, runs a secret-exclusion scan, builds the public trace spans, and folds the ordered validators into one result, with run enabling negative cases and the bundle path leaving them out. build_result is the pipeline. It loads the seven input files, runs a secret-exclusion scan over the input paths, builds the public trace spans, then calls the validators in order and folds their statuses into one result. run drives fixture mode with the negative cases enabled. runpromptinjection_bundle drives the exported example bundle and leaves the fixture-only negative cases out. Both write metadata-only records through writereceipts and can emit a compact card through res\u2026",
"level": 4,
"order": 8,
"section_id": "build-result-runs-the-validators-in-order",
"source_line_end": 84,
"source_line_start": 62,
"title": "`_build_result` runs the validators in order"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "derivedpolicy_verdict is a small fixed ladder that blocks hidden or instruction-channel promotion and unsanitized untrusted taint into privileged sinks, reviews or warns the sanitized cases, and allows everything else.** The verdict ladder in derivedpolicy_verdict is small and fixed. A flow that promotes hidden or instruction-channel text is blocked. Untrusted taint into a privileged sink is blocked when unsanitized and set to review when sanitized. Untrusted taint into a non-privileged sink warns if sanitized and reviews if not. Everything else is allowed. validateprojectionprotocol and validateinjectionpolicy sit alongside these: the policy must define all four verdicts and deny every aut\u2026",
"level": 4,
"order": 9,
"section_id": "the-fixed-verdict-ladder",
"source_line_end": 122,
"source_line_start": 85,
"title": "The fixed verdict ladder"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The fixture carries five sources, five flows covering all four verdicts, five outputs, five cold replays, one promoted tool-call trace, and eight negatives, while the bundle adds a digest-checked source-module manifest.** The fixture carries five source documents, three untrusted and two trusted, five source-to-sink flows whose verdicts cover allow, warn, block, and review, five sanitized outputs, five cold replay passes, five public trace spans, one promoted tool-call trace, and eight negative cases. The exported bundle also ships a sourcemodulemanifest.json, and sourcemodulemanifestresult checks that its copied source-module bodies are classified as copiednonsecretmacrobody, keep body t\u2026",
"level": 4,
"order": 10,
"section_id": "what-the-fixture-and-bundle-carry",
"source_line_end": 136,
"source_line_start": 123,
"title": "What the fixture and bundle carry"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 11,
"section_id": "negative-cases",
"source_line_end": 138,
"source_line_start": 137,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The eight EXPECTEDNEGATIVE_CASES are the admission boundary, each a deliberately-bad flow the runtime must reject, and fixture mode is blocked if any goes unobserved. The eight fixtures in EXPECTEDNEGATIVE_CASES are the admission boundary. validatenegativecases must reject each one: real account material, secret or trusted-context exfiltration, raw prompt body export, untrusted tool output treated as instruction authority, hidden system-message promotion, account secret exfiltration, final-answer-only success, and ungated untrusted flow into a privileged sink. These are falsification fixtures, part of the contract, not samples of live exploit traffic. Fixture mode is blocked if any expected\u2026",
"level": 4,
"order": 12,
"section_id": "eight-falsification-fixtures-as-the-boundary",
"source_line_end": 150,
"source_line_start": 139,
"title": "Eight falsification fixtures as the boundary"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "prior-art-grounding",
"source_line_end": 152,
"source_line_start": 151,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The threat shape follows Greshake et al., AgentDojo, and tool-calling data-leakage work, with the policy mechanism drawing on dynamic information-flow and taint-tracking, turned here into a metadata-only replay with negative cases.** The threat shape follows the indirect-injection framing of Greshake et al. in Not what you've signed up for, the agentic evaluation setting of AgentDojo, and the tool-calling data-leakage results in Simple Prompt Injection Attacks Can Leak Personal Data. The policy mechanism draws on dynamic information-flow and taint-tracking, including Permissive Information-Flow Analysis for Large Language Models. The engineering lesson taken from that work is narrow: untr\u2026",
"level": 4,
"order": 14,
"section_id": "indirect-injection-and-information-flow-lineage",
"source_line_end": 169,
"source_line_start": 153,
"title": "Indirect-injection and information-flow lineage"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "validation-receipt-path",
"source_line_end": 171,
"source_line_start": 170,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run the fixture validator from the repo root and write its record outside the working tree: Run the fixture validator from the repo root and write its record outside the working tree: Then run the exported bundle validator: The focused regression test and corpus projection check are:",
"level": 4,
"order": 16,
"section_id": "fixture-validation-outside-the-working-tree",
"source_line_end": 193,
"source_line_start": 172,
"title": "Fixture validation outside the working tree"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the synthetic trace reproduced its verdicts and sanitized outputs, observed the expected negative cases, and wrote no source body, not general robustness or live account safety.** A pass means the synthetic trace reproduced its verdicts and sanitized outputs, observed the expected negative cases, and wrote no source body. It does not mean general prompt-injection robustness or live account safety.",
"level": 4,
"order": 17,
"section_id": "what-a-pass-means-here",
"source_line_end": 202,
"source_line_start": 194,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "claim-ceiling",
"source_line_end": 204,
"source_line_start": 203,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Over declared synthetic metadata-only rows the replay recomputes source trust, taint propagation, pre-action verdicts, sanitized-output constraints, cold replay, and copied source-module digests, exposing the results without leaking a body.** The strongest claim the evidence supports is this: over declared synthetic metadata-only rows, the replay recomputes source trust, taint propagation, pre-action verdicts, sanitized-output constraints, cold replay, and the copied source-module digests, and it exposes those results without leaking a body. The copied-body digest proves byte parity for the named files only; it does not widen the replay into live source authority.",
"level": 4,
"order": 19,
"section_id": "the-strongest-supported-claim",
"source_line_end": 216,
"source_line_start": 205,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The proof boundary is that fixed input set and the result records, not general prompt-injection robustness, live account or tool safety, raw body export, benchmark claims, or launch-scope decision.** The proof boundary is that fixed input set and the validation result records listed above, not this prose. Its scope limit stops there. It does not claim general prompt-injection robustness, live account safety, live tool or provider behaviour, raw prompt or system or tool body export, account secret-bearing data, hidden-message handling in production, benchmark claims, source-file changes, public sharing, hosting, launch-scope decision, or complete security.",
"level": 4,
"order": 20,
"section_id": "what-the-replay-does-not-claim",
"source_line_end": 227,
"source_line_start": 217,
"title": "What the replay does not claim"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/indirect_prompt_injection_information_flow_policy_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Validator-backed claim: a source-faithful trace refactor separated trusted instructions from untrusted web/tool/browser text before any privileged action."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "indirect_prompt_injection_information_flow_policy_replay.py",
"path": "src/microcosm_core/organs/indirect_prompt_injection_information_flow_policy_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for the public replay/regression validator.",
"symbols": [
"run",
"run_prompt_injection_bundle",
"_build_result",
"_write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/indirect_prompt_injection_information_flow_policy_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/indirect_prompt_injection_information_flow_policy_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/indirect_prompt_injection_information_flow_policy_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Validator-backed public claim contract for indirect prompt-injection information-flow policy. Admits one narrow claim: a source-faithful trace refactor separated trusted instructions from untrusted web/tool/browser text before any privileged action or answer, checked against negative cases and scope limits.",
"summary_line": "Validator-backed claim: a source-faithful trace refactor separated trusted instructions from untrusted web/tool/browser text before any privileged action.",
"title": "Indirect Prompt-Injection Information-Flow Policy Replay"
},
{
"anchor": "paper-module-mcp-tool-authority-replay",
"authority_ceiling": "Synthetic MCP-like replay fixture only; not a live MCP account test, external model access, account secret-handling certification, benchmark security result, or launch claim.",
"body_html": "
This component replays a recorded MCP tool trace and calls the tool use safe only when every action shows a narrow scope, prior approval, a rollback path, and untrusted output kept out of instructions. When an agent uses outside tools through a protocol like MCP, the sentence "the agent used the tool safely" is cheap to write. This component makes it expensive. It takes a recorded trace of tool use and checks that every action in it was fenced: bound to a narrow permission, approved before it changed anything, given a way to undo it, and stopped from letting untrusted tool output give the agent orders.
\n
The check runs on a built-in example of labels and references only, scope names, argument hashes, and token ids, so tool-use safety is inspectable without touching a live account. The check runs against a small built-in example, not a live account. The example carries only labels and references: capability scope names, argument hashes, approval token ids, ledger and rollback references. No secrets, no account secrets, no tool contents. So the safety properties an agent's tool use is meant to have become inspectable without anything touching a real provider or account.
\n
It reads a directory of input JSON and emits four metadata-only records carrying counts, status bits, and digests, but no payload, account secret, or tool body. The component reads a directory of input JSON and emits a set of metadata-only result records: a result, a board, a validation record, and a fixture sign-off record. Each records counts, status bits, and reference digests. None of them carries a payload, a account secret, or a tool body.
\n
Purpose
\n
What a loose tool trace hides
\n
Describing a tool call is easy and backing it is hard, and loose prose hides the two failure modes that matter for tool-using agents. A tool call is easy to describe and hard to back. A trace can say a write happened and that it was fine. Loose prose like that hides the two failure modes that matter for tool-using agents.
\n
Confused deputy and injected instructions
\n
The two modes are the confused deputy, where a call borrows a scope wider than its task, and tool-output-as-instruction, where untrusted output is obeyed as a command. The first is the confused deputy: a call asks for a scope wider than its task needs, so a tool quietly borrows more authority than it was granted. The second is tool-output-as-instruction, the prompt-injection shape where text returned by an untrusted tool is obeyed as a command. Both pass a reader who only sees a summary. Both fail once you require the trace to show its working.
\n
Each call is an evidence-citing transaction
\n
Each tool call is treated as a transaction that must cite its evidence, and every reference is cross-checked against the accepted calls by call id rather than taken on trust. So the component treats each tool call as a small transaction that must cite its evidence. A call names a narrow capability scope and an argument hash. If it writes, it must also name an approval token, a side-effect ledger entry, and a rollback record. Those references are not taken on trust. The side-effect ledger and the cold replay records are cross-checked against the accepted call records by call id, so a rollback record no call refers to, or a write that skips approval, is caught rather than waved through. A reference string is not authority until something downstream resolves it.
\n
How it works
\n
Two entrypoints over one pipeline
\n
run validates the first-wave fixture with its eight negative cases and run_tool_authority_bundle validates an exported bundle without them, and both call the shared _build_result pipeline. The two entrypoints are run and run_tool_authority_bundle. run validates the first-wave fixture with the eight negative cases included, then writes the four fixture records through _write_receipts. run_tool_authority_bundle validates an exported bundle without the negative cases and writes a single bundle result record. Both call _build_result, which is the pipeline.
\n
Secret scan runs before plane validation
\n
_build_result runs in a fixed order: load the input JSON, scan inputs and copied source for a leaked secret first, then validate each plane and pass the accepted records forward._build_result runs in a fixed order. It loads the input JSON with _load_payloads, then runs load_forbidden_classes and scan_paths over the inputs and any copied source files to catch a leaked secret before anything else. It then validates each plane on its own and passes the accepted records forward:
\n
Function
What it checks
validate_tool_policy
the declared policy id and allowed tool classes
validate_tool_manifest
declared tool ids and their classes; produces the set of accepted tool ids
validate_tool_calls
each call, using _validate_call_row; requires three calls with one approved write and one untrusted result
validate_tool_results
each result binds to an accepted call id; one untrusted output is kept as data and ignored as instruction
validate_side_effect_ledger
one approved write side effect whose approval, ledger, and rollback references match its call record
validate_cold_replay
three passing replays whose evidence references bind to the manifest and to runtime records
validate_projection_protocol
the protocol id and source-to-target projection references
\n
Where a call is admitted or blocked
\n
_validate_call_row rejects undeclared tools, overbroad scopes, exported account secrets, unapproved writes, missing rollbacks, and in-record payloads, returning accepted_tool_call_metadata only when no reason fires._validate_call_row is where a call is admitted or rejected. It rejects an undeclared tool id, an overbroad scope (*, account_full_access, and the like), an exported account secret, untrusted output flagged as instruction, live account access, final-answer-only grading, a write that lacks approval or a ledger reference, a missing rollback record, and any call whose payload sits in the record instead of staying metadata-only. It also checks that every field in REQUIRED_CALL_FIELDS is present. A clean call returns accepted_tool_call_metadata; any reason at all returns blocked.
\n
Trace spans and the body floor
\n
After the planes validate, three public trace spans are contributed and the body floor verifies each copied source digest against its live reference with body_in_receipt: false on every row. After the planes validate, build_public_mcp_tool_authority_trace contributes three public trace spans for the calls, and _source_module_manifest_result runs the body floor for the bundle path: it verifies the digest of each copied source file against its live source reference, checks required anchors, and requires body_in_receipt: false on every row. _body_import_verification binds that public refactor back to the source and target digests.
\n
The final status is pass only when every plane passes, no expected negative case is missing, the secret scan has zero blocking hits, and the source-module floor is present when the bundle requires it. The final status is pass only when every plane passes, no expected negative case is missing, the secret scan has zero blocking hits, and the source-module floor is present when the bundle requires it. _write_receipts then emits the result, board, validation, and sign-off records, and result_card projects a compact card that carries counts and status bits while omitting the tool rows, call rows, findings, and source bodies listed in CARD_OMITTED_FULL_PAYLOAD_KEYS.
\nDiagram of the mechanism (7 steps).Diagram source & refs
flowchart TD input["input JSON directory"] scan["secret scan load_forbidden_classes + scan_paths"] planes["plane validators policy, manifest, calls, results, side effects, cold replay"] trace["three public trace spans"] floor["source-module body floor"] status["pass only if all clear"] records["metadata-only result records"] input --> scan scan --> planes planes --> trace planes --> floor trace --> status floor --> status status --> records
\n
Negative cases
\n
Eight negatives as a falsification set
\n
The first-wave fixture ships eight negative inputs, and a run passes only if it observes the error code for every one, making the fixture a falsification set. The first-wave fixture ships eight negative inputs, one per case in EXPECTED_NEGATIVE_CASES. A first-wave run passes only if it observes the error code for every one of them, so the fixture is a falsification set the validators have to trip on:
A negative case that never trips is recorded under missing_negative_cases, and its absence blocks the run rather than letting it pass. If any expected case goes unobserved, _build_result records it under missing_negative_cases and the run is blocked.
\n
Prior Art Grounding
\n
Capability security made replayable
\n
The design follows capability-security and least-privilege lineage from Saltzer and Schroeder through Hardy and the MCP authorization spec, then makes it replayable so scope and approval references must precede any write authority. The component follows capability security, least privilege, and current MCP authorization guidance. The classic lineage is Saltzer and Schroeder's Protection of Information in Computer Systems and Hardy's Confused Deputy: authority should be narrow, mediated, and bound to the action being requested. The MCP-specific lineage is the official MCP authorization and security best practices guidance on least-privilege scopes and token audience boundaries. The contribution here is making that shape replayable: a tool story has to expose scope, approval, side-effect, rollback, and instruction/data references before its write authority counts as evidence.
\n
Validation Result record Path
\n
Run the first-wave fixture into disposable records from the Microcosm root:
\n
Run the exported bundle through the same component:
A green run certifies only internal consistency: every plane validated, all eight negatives observed, the secret scan clean, and the records metadata-only. A pass means every plane validated, all eight negative cases were observed, the secret scan found nothing, and the records stayed metadata-only.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The strongest claim is narrow: a synthetic MCP-like replay preserved its tool-authority boundaries over metadata records, with every reference and negative case lined up or the run blocked. The strongest claim is this: a synthetic MCP-like replay preserved its tool-authority boundaries over metadata records. Capability scopes, argument hashes, approval references, side-effect and rollback references, instruction/data splits, cold replay references, source digests, the eight negative cases, and metadata-only validation records all lined up, or the run blocked.
\n
What stays false on a pass
\n
Even on a pass the component opens no account, calls no provider, and handles no account secret, so it claims no live MCP safety and certifies no real integration secure. That is the scope limit and the proof boundary. The component never opens an MCP account, calls a provider, or handles a account secret, and it keeps every payload, tool result, and account secret out of what it writes. So it does not claim live MCP account safety, account secret-handling certification, live tool or provider behavior, benchmark security, source-file changes, complete security, publishing-scope decision, or launch-scope decision. It checks that a tool-authority story is internally consistent and metadata-only. It does not certify that any real tool integration is secure.
\n",
"components": [
{
"anchor": "component-mcp_tool_authority_replay",
"component_id": "mcp_tool_authority_replay",
"public_label": "MCP Tool-Authority Policy Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 2
},
"link_count": 7,
"links": [
{
"id": "mcp_tool_authority_replay",
"kind": "organ",
"kind_label": "component",
"label": "MCP Tool-Authority Policy Replay",
"relation": "explains",
"url": "component-mcp-tool-authority-replay.html"
},
{
"id": "mechanism.mcp_tool_authority_replay.validates_public_mcp_tool_authority_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public mcp tool authority replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-mcp-tool-authority-replay-validates-public-mcp-tool-authority-replay"
},
{
"id": "P-4",
"kind": "principle",
"kind_label": "principle",
"label": "Possession is not permission",
"relation": "governed by",
"url": "doctrine.html#dcard-p-4"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "AX-3",
"kind": "axiom",
"kind_label": "axiom",
"label": "Authority by derivation, not possession",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-3"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.agent_sandbox_policy_escape_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Sandbox Policy-Escape Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[35:paper_module.agent_sandbox_policy_escape_replay].depends_on",
"url": "paper-module-agent-sandbox-policy-escape-replay.html"
},
{
"id": "paper_module.sleeper_memory_poisoning_quarantine_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Sleeper Memory Poisoning Quarantine Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[37:paper_module.sleeper_memory_poisoning_quarantine_replay].depends_on",
"url": "paper-module-sleeper-memory-poisoning-quarantine-replay.html"
}
],
"paper_module_id": "paper_module.mcp_tool_authority_replay",
"reader_packet": {
"abstract": "Synthetic MCP-like replay fixture for a tool-authority claim contract: replay result records, negative cases, scope limits; no live MCP/provider/account secret",
"authority_ceiling": "Synthetic MCP-like replay fixture only; not a live MCP account test, external model access, account secret-handling certification, benchmark security result, or launch claim.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.mcp_tool_authority_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-mcp-tool-authority-replay-diagram-1",
"html_url": "docs/paper-module-mcp-tool-authority-replay.html#paper-module-mcp-tool-authority-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD input[\"input JSON directory\"] scan[\"secret scan load_forbidden_classes + scan_paths\"] planes[\"plane validators policy, manifest, calls, results, side effects, cold replay\"] trace[\"three public trace spans\"] floor[\"source-module body floor\"] status[\"pass only if all clear\"] records[\"metadata-only result records\"] input --> scan scan --> planes planes --> trace planes --> floor trace --> status floor --> status status --> records",
"source_format": "mermaid",
"source_hash": "sha256:3bb2b7755223a68b417eb336f8695cd1700263c8a30d7e213a255071df7ece27",
"source_line_end": 156,
"source_line_start": 139
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-3",
"kind": "axiom",
"kind_label": "axiom",
"label": "Authority by derivation, not possession",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-3"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "mcp_tool_authority_replay",
"kind": "organ",
"kind_label": "component",
"label": "MCP Tool-Authority Policy Replay",
"relation": "explains",
"url": "component-mcp-tool-authority-replay.html"
},
{
"id": "mechanism.mcp_tool_authority_replay.validates_public_mcp_tool_authority_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public mcp tool authority replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-mcp-tool-authority-replay-validates-public-mcp-tool-authority-replay"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-4",
"kind": "principle",
"kind_label": "principle",
"label": "Possession is not permission",
"relation": "governed by",
"url": "doctrine.html#dcard-p-4"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.mcp_tool_authority_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 265,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/mcp_tool_authority_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component replays a recorded MCP tool trace and calls the tool use safe only when every action shows a narrow scope, prior approval, a rollback path, and untrusted output kept out of instructions.** When an agent uses outside tools through a protocol like MCP, the sentence \"the agent used the tool safely\" is cheap to write. This component makes it expensive. It takes a recorded trace of tool use and checks that every action in it was fenced: bound to a narrow permission, approved before it changed anything, given a way to undo it, and stopped from letting untrusted tool output give the agent orders.",
"level": 4,
"order": 1,
"section_id": "this-component-replays-a-recorded-mcp-tool-trace",
"source_line_end": 14,
"source_line_start": 3,
"title": "This component replays a recorded MCP tool trace"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The check runs on a built-in example of labels and references only, scope names, argument hashes, and token ids, so tool-use safety is inspectable without touching a live account.** The check runs against a small built-in example, not a live account. The example carries only labels and references: capability scope names, argument hashes, approval token ids, ledger and rollback references. No secrets, no account secrets, no tool contents. So the safety properties an agent's tool use is meant to have become inspectable without anything touching a real provider or account.",
"level": 4,
"order": 2,
"section_id": "the-check-runs-on-a-built-in-example-of-labels-and-references-only",
"source_line_end": 25,
"source_line_start": 15,
"title": "The check runs on a built-in example of labels and references only"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It reads a directory of input JSON and emits four metadata-only records carrying counts, status bits, and digests, but no payload, account secret, or tool body.** The component reads a directory of input JSON and emits a set of metadata-only result records: a result, a board, a validation record, and a fixture sign-off record. Each records counts, status bits, and reference digests. None of them carries a payload, a account secret, or a tool body.",
"level": 4,
"order": 3,
"section_id": "it-reads-a-directory-of-input-json",
"source_line_end": 34,
"source_line_start": 26,
"title": "It reads a directory of input JSON"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 36,
"source_line_start": 35,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Describing a tool call is easy and backing it is hard, and loose prose hides the two failure modes that matter for tool-using agents.** A tool call is easy to describe and hard to back. A trace can say a write happened and that it was fine. Loose prose like that hides the two failure modes that matter for tool-using agents.",
"level": 4,
"order": 5,
"section_id": "what-a-loose-tool-trace-hides",
"source_line_end": 44,
"source_line_start": 37,
"title": "What a loose tool trace hides"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The two modes are the confused deputy, where a call borrows a scope wider than its task, and tool-output-as-instruction, where untrusted output is obeyed as a command.** The first is the confused deputy: a call asks for a scope wider than its task needs, so a tool quietly borrows more authority than it was granted. The second is tool-output-as-instruction, the prompt-injection shape where text returned by an untrusted tool is obeyed as a command. Both pass a reader who only sees a summary. Both fail once you require the trace to show its working.",
"level": 4,
"order": 6,
"section_id": "confused-deputy-and-injected-instructions",
"source_line_end": 55,
"source_line_start": 45,
"title": "Confused deputy and injected instructions"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Each tool call is treated as a transaction that must cite its evidence, and every reference is cross-checked against the accepted calls by call id rather than taken on trust.** So the component treats each tool call as a small transaction that must cite its evidence. A call names a narrow capability scope and an argument hash. If it writes, it must also name an approval token, a side-effect ledger entry, and a rollback record. Those references are not taken on trust. The side-effect ledger and the cold replay records are cross-checked against the accepted call records by call id, so a rollback record no call refers to, or a write that skips approval, is caught rather than waved through. A\u2026",
"level": 4,
"order": 7,
"section_id": "each-call-is-an-evidence-citing-transaction",
"source_line_end": 69,
"source_line_start": 56,
"title": "Each call is an evidence-citing transaction"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 71,
"source_line_start": 70,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**run validates the first-wave fixture with its eight negative cases and runtoolauthority_bundle validates an exported bundle without them, and both call the shared buildresult pipeline.** The two entrypoints are run and runtoolauthority_bundle. run validates the first-wave fixture with the eight negative cases included, then writes the four fixture records through writereceipts. runtoolauthority_bundle validates an exported bundle without the negative cases and writes a single bundle result record. Both call buildresult, which is the pipeline.",
"level": 4,
"order": 9,
"section_id": "two-entrypoints-over-one-pipeline",
"source_line_end": 82,
"source_line_start": 72,
"title": "Two entrypoints over one pipeline"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "buildresult runs in a fixed order: load the input JSON, scan inputs and copied source for a leaked secret first, then validate each plane and pass the accepted records forward. build_result runs in a fixed order. It loads the input JSON with loadpayloads, then runs loadforbiddenclasses and scan_paths over the inputs and any copied source files to catch a leaked secret before anything else. It then validates each plane on its own and passes the accepted records forward:",
"level": 4,
"order": 10,
"section_id": "secret-scan-runs-before-plane-validation",
"source_line_end": 102,
"source_line_start": 83,
"title": "Secret scan runs before plane validation"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatecall_row rejects undeclared tools, overbroad scopes, exported account secrets, unapproved writes, missing rollbacks, and in-record payloads, returning acceptedtoolcallmetadata only when no reason fires. validatecall_row is where a call is admitted or rejected. It rejects an undeclared tool id, an overbroad scope (, accountfull_access, and the like), an exported account secret, untrusted output flagged as instruction, live account access, final-answer-only grading, a write that lacks approval or a ledger reference, a missing rollback record, and any call whose payload sits in the record instead of staying metadata-only. It also checks that every field in REQUIREDCALLFIELDS is present\u2026",
"level": 4,
"order": 11,
"section_id": "where-a-call-is-admitted-or-blocked",
"source_line_end": 116,
"source_line_start": 103,
"title": "Where a call is admitted or blocked"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**After the planes validate, three public trace spans are contributed and the body floor verifies each copied source digest against its live reference with bodyinreceipt: false on every row.** After the planes validate, buildpublicmcptoolauthority_trace contributes three public trace spans for the calls, and sourcemodulemanifestresult runs the body floor for the bundle path: it verifies the digest of each copied source file against its live source reference, checks required anchors, and requires bodyinreceipt: false on every row. bodyimport_verification binds that public refactor back to the source and target digests.",
"level": 4,
"order": 12,
"section_id": "trace-spans-and-the-body-floor",
"source_line_end": 128,
"source_line_start": 117,
"title": "Trace spans and the body floor"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "The final status is pass only when every plane passes, no expected negative case is missing, the secret scan has zero blocking hits, and the source-module floor is present when the bundle requires it. The final status is pass only when every plane passes, no expected negative case is missing, the secret scan has zero blocking hits, and the source-module floor is present when the bundle requires it. writereceipts then emits the result, board, validation, and sign-off records, and result_card projects a compact card that carries counts and status bits while omitting the tool rows, call rows, findings, and source bodies listed in CARDOMITTEDFULLPAYLOADKEYS.",
"level": 4,
"order": 13,
"section_id": "the-final-status-is-pass-only-when-every-plane-passes",
"source_line_end": 157,
"source_line_start": 129,
"title": "The final status is `pass` only when every plane passes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 159,
"source_line_start": 158,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**The first-wave fixture ships eight negative inputs, and a run passes only if it observes the error code for every one, making the fixture a falsification set.** The first-wave fixture ships eight negative inputs, one per case in EXPECTEDNEGATIVECASES. A first-wave run passes only if it observes the error code for every one of them, so the fixture is a falsification set the validators have to trip on: overbroadscope raises MCPTOOLOVERBROADSCOPE. hiddencredentialexport raises MCPTOOLHIDDENCREDENTIALEXPORT. tooloutputasinstruction raises MCPTOOLOUTPUTAS_INSTRUCTION. unapprovedsideeffect raises MCPTOOLUNAPPROVEDSIDEEFFECT. liveaccountaccess raises MCPTOOLLIVEACCOUNTACCESS. finalansweronlygrad\u2026",
"level": 4,
"order": 15,
"section_id": "eight-negatives-as-a-falsification-set",
"source_line_end": 177,
"source_line_start": 160,
"title": "Eight negatives as a falsification set"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A negative case that never trips is recorded under missingnegative_cases, and its absence blocks the run rather than letting it pass.** If any expected case goes unobserved, buildresult records it under missingnegativecases and the run is blocked.",
"level": 4,
"order": 16,
"section_id": "an-unobserved-case-blocks-the-run",
"source_line_end": 184,
"source_line_start": 178,
"title": "An unobserved case blocks the run"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 186,
"source_line_start": 185,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The design follows capability-security and least-privilege lineage from Saltzer and Schroeder through Hardy and the MCP authorization spec, then makes it replayable so scope and approval references must precede any write authority.** The component follows capability security, least privilege, and current MCP authorization guidance. The classic lineage is Saltzer and Schroeder's [Protection of Information in Computer Systems](https://www.cs.virginia.edu/~evans/cs551/saltzer/) and Hardy's [Confused Deputy](https://web.cs.wpi.edu/~cs557/f14/papers/confused_deputy-hardy.pdf): authority should be narrow, mediated, and bound to the action being requested. The MCP-specific lineage is the officia\u2026",
"level": 4,
"order": 18,
"section_id": "capability-security-made-replayable",
"source_line_end": 207,
"source_line_start": 187,
"title": "Capability security made replayable"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run the first-wave fixture into disposable records from the Microcosm root: Run the exported bundle through the same component: Then run the focused test and the corpus parity check:",
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 231,
"source_line_start": 208,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A green run certifies only internal consistency: every plane validated, all eight negatives observed, the secret scan clean, and the records metadata-only.** A pass means every plane validated, all eight negative cases were observed, the secret scan found nothing, and the records stayed metadata-only.",
"level": 4,
"order": 20,
"section_id": "what-a-pass-certifies",
"source_line_end": 238,
"source_line_start": 232,
"title": "What a pass certifies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 240,
"source_line_start": 239,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is narrow: a synthetic MCP-like replay preserved its tool-authority boundaries over metadata records, with every reference and negative case lined up or the run blocked.** The strongest claim is this: a synthetic MCP-like replay preserved its tool-authority boundaries over metadata records. Capability scopes, argument hashes, approval references, side-effect and rollback references, instruction/data splits, cold replay references, source digests, the eight negative cases, and metadata-only validation records all lined up, or the run blocked.",
"level": 4,
"order": 22,
"section_id": "the-strongest-supported-claim",
"source_line_end": 251,
"source_line_start": 241,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Even on a pass the component opens no account, calls no provider, and handles no account secret, so it claims no live MCP safety and certifies no real integration secure.** That is the scope limit and the proof boundary. The component never opens an MCP account, calls a provider, or handles a account secret, and it keeps every payload, tool result, and account secret out of what it writes. So it does not claim live MCP account safety, account secret-handling certification, live tool or provider behavior, benchmark security, source-file changes, complete security, public sharing decision, or launch authority. It checks that a tool-authority story is internally consistent and metadata-only.\u2026",
"level": 4,
"order": 23,
"section_id": "what-stays-false-on-a-pass",
"source_line_end": 264,
"source_line_start": 252,
"title": "What stays false on a pass"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/mcp_tool_authority_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Synthetic MCP-like replay fixture for a tool-authority claim contract: replay result records, negative cases, scope limits; no live MCP/provider/account secret"
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "mcp_tool_authority_replay.py",
"path": "src/microcosm_core/organs/mcp_tool_authority_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for the public replay/regression validator.",
"symbols": [
"run",
"run_tool_authority_bundle",
"_build_result",
"_write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/mcp_tool_authority_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/mcp_tool_authority_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/mcp_tool_authority_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Public Plectis projection of a tool-authority claim contract. Replays a synthetic MCP-like tool-authority scenario and validates replay result records, negative cases, and scope limits with metadata-only result records.",
"summary_line": "Synthetic MCP-like replay fixture for a tool-authority claim contract: replay result records, negative cases, scope limits; no live MCP/provider/account secret",
"title": "MCP Tool Authority Replay"
},
{
"anchor": "paper-module-tactic-portfolio-availability",
"authority_ceiling": "Copied tactic affordance probe rows and public fixture/exported-bundle result records only; no Lean/Lake rerun, theorem proof, benchmark performance, external model access, Mathlib-dependent proof authority, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
It turns one captured Lean smoke run into an explicit list of callable tactics, checking pre-recorded status rows for honesty without running Lean itself.tactic_portfolio_availability_probe turns one captured Lean run into an explicit list of which proof tactics were callable, before any router or proof search treats a tactic as usable. It reads pre-recorded status rows from a real smoke run and checks them for honesty. It does not run Lean itself.
\n
The copied 2026-05-11 fixture covers eight tactics, seven compiling and aesop failed on a missing Mathlib import, reproduced into a metadata-only board and result records. The fixture is copied from a 2026-05-11 tactic affordance probe. It covers eight tactics: rfl, decide, omega, simp, simp_all, grind, native_decide, and aesop. Seven are recorded as compiling. aesop is recorded as failed because the paired environment probe reports that the Mathlib import it needs was absent. The component reproduces that verdict from the copied rows and writes a metadata-only availability board plus result records.
\n
Purpose
\n
A tactic name is not availability
\n
A tactic name alone does not mean it can run; a router trusting the bare name reaches for tactics the environment cannot run and misreads the failure as a fact about the goal. A tactic name is not a usable tactic. aesop runs only if the surrounding Lean and Std environment carries the imports it needs. omega is callable in one project layout and not in another. A router that trusts a bare tactic name will reach for tactics the current environment cannot run, then misread the failure as a fact about the goal rather than a fact about the environment.
\n
Missing import kept distinct from broken tactic
\n
It answers which tactics were callable and on what evidence, keeping a missing import distinct from a broken tactic so a router can re-attempt elsewhere. The component answers one question: in the observed environment, which tactics were callable, and on what evidence? The interesting part is how it treats failure. A recorded Lean FAIL is not flattened into one "unavailable" verdict. A missing import is kept distinct from a broken tactic, so a router can re-attempt the same tactic in a different environment instead of striking it off for good.
\n
How it works
\n
Two entrypoints over _build_result
\n
run validates the first-wave fixture with negatives on, run_availability_bundle validates an exported bundle with negatives off and reuses an unchanged prior result. Two entrypoints drive the pipeline. run validates the first-wave fixture with negative cases enabled. run_availability_bundle validates an exported bundle with negative cases disabled, and reuses a prior result when _fresh_availability_bundle_receipt finds a matching record whose inputs have not changed since it was written. Both delegate the work to _build_result and then persist records.
\n
_build_result runs these steps in order.
\n
Digest-checking the copied source payloads
\n
It loads the three copied source payloads and digest-checks each body against a recorded SHA-256, so a silently edited body becomes a finding rather than trusted input. It loads the copied source payloads with _load_source_payloads: the run-level affordance probe, the portfolio_core_v0 availability artifact, and the paired corpus-readiness file. Each copied body is digest-checked by _source_body_imports and _probe_source_imports against a recorded SHA-256, so a silently edited body becomes a finding rather than trusted input.
\n
Mathlib probe gates dependent tactics
\n
It derives the environment state from the Mathlib probe, whose single fact of whether the Lake import was available gates every Mathlib-dependent tactic. It derives the environment state with _authoritative_environment_state, which reads the Mathlib probe and reports whether the Mathlib Lake import was available. That single fact gates every Mathlib-dependent tactic.
\n
Deriving each tactic's public status
\n
It builds one row per tactic, mapping a source PASS to compile_pass and a FAIL to environment_fail when a required Mathlib import was absent, else compile_fail. It builds one availability row per tactic with _availability_rows_from_source. The public status comes from _derived_public_status_from_source: a source PASS becomes compile_pass; a source FAIL becomes environment_fail when the tactic requires Mathlib and Mathlib was absent, and compile_fail otherwise. When the status is a failure, _failure_classifier_from_source labels a Mathlib gap as MATHLIB_IMPORT_MISSING and otherwise carries the source error class through. This is why aesop reads as an environment failure, not a broken tactic.
\n
Portfolio summary and latency banding
\n
It summarises available and unavailable ids and bands each available tactic's copied duration as fast, moderate, or slow, stamped environment-scoped rather than a benchmark. It summarises the rows with _portfolio_summary: available and unavailable tactic ids, Mathlib-dependent ids, and a latency profile. _latency_band bands each available tactic by its copied duration as fast at 2500ms or less, moderate up to 5000ms, and slow above that. The latency profile is stamped environment-scoped, so a router can prefer a cheaper available tactic without reading speed as a benchmark result.
\n
Negatives, secret scan, and result records
\n
It evaluates the negative cases, scans every input path for forbidden material, and passes only with no finding, all negatives observed, and zero secret hits before writing result records. It evaluates the declared negative cases with _negative_findings, scans all input paths for forbidden material with scan_paths, and sets the overall status. The result is pass only when no category produced a finding, every expected negative case was observed, and the secret scan had zero blocking hits. _write_receipts then writes the result record, the availability board, the validation record, and the sign-off record, each carrying result record paths but no tactic bodies.
\nDiagram of the mechanism (9 steps).Diagram source & refs
flowchart TD A["Copied affordance probe rows compile_status, requires_mathlib, duration_ms"] --> B["_build_result"] C["Mathlib environment probe"] --> B B --> D{"source compile_status"} D -->|PASS| E["compile_pass band duration fast / moderate / slow"] D -->|FAIL + requires Mathlib + Mathlib absent| F["environment_fail reason MATHLIB_IMPORT_MISSING"] D -->|FAIL otherwise| G["compile_fail"] E --> H["Availability board"] F --> H G --> H B --> I["metadata-only result and validation records"]
\n
Negative cases
\n
Six perturbed fixtures and their codes
\n
Six perturbed fixtures must each be rejected with a specific error code, spanning missing status, unbacked Mathlib claims, unprobed tactics, missing durations, leaked bodies, and authority overclaim._negative_findings checks six perturbed fixtures and requires each to be rejected with a specific error code. A tactic row with no compile_status gives TACTIC_PORTFOLIO_MISSING_COMPILE_STATUS. A Mathlib-dependent tactic marked available with no passing Mathlib probe gives TACTIC_PORTFOLIO_MATHLIB_CLAIM_WITHOUT_PROBE. A downstream request for a tactic not in the probed portfolio gives TACTIC_PORTFOLIO_UNPROBED_TACTIC_REFERENCED. An available tactic with no copied duration gives TACTIC_PORTFOLIO_AVAILABLE_DURATION_MISSING. A fixture carrying a proof or provider body gives TACTIC_PORTFOLIO_PROOF_BODY_FORBIDDEN. A payload asserting proof, provider, benchmark, or launch-scope decision gives TACTIC_PORTFOLIO_AUTHORITY_OVERCLAIM. A missing expected case blocks the run.
\n
Prior Art Grounding
\n
Feature-detection discipline from Autoconf and Lean
\n
It follows the feature-detection habit of probing a local capability before relying on it, applying Autoconf and Lean's environment-sensitive discipline to one recorded run. The component follows the feature-detection habit of testing a local capability before relying on it, the same discipline GNU Autoconf's configure workflow established for build environments. Lean's own tactic documentation notes that tactic use is environment- and goal-sensitive, so a tactic name alone does not justify downstream routing. This component applies that probe discipline to one recorded run and keeps a Mathlib-dependent absence as evidence rather than discarding it.
\n
GNU Autoconf feature probing: https://ftp.gnu.org/old-gnu/Manuals/autoconf-2.57/html_chapter/autoconf.html
From microcosm-substrate/, reproduce the evidence with temporary records:
\n
What a pass certifies
\n
A pass means every copied body matched its digest, the eight rows produced the expected board with aesop as an environment failure, all six negatives were rejected, and the secret scan was clean. A pass means every copied source body matched its recorded digest, the eight tactic rows produced the expected availability board with aesop as an environment failure, all six negative cases were rejected, and the secret scan found no forbidden material.
\n
Scope boundary
\n
Scope limit
\n
Callability held as callability
\n
The strongest claim is that in one observed environment these tactics were callable on digest-matched probe rows; it proves no goal, reruns no Lean, and authorizes nothing. The strongest claim the evidence supports is narrow: in one observed Lean and Std environment, these tactics were callable, on the evidence of copied probe rows whose bodies match recorded digests, with Mathlib absence preserved as a recorded fact. The proof boundary is those copied rows, the derived availability board, the digest and secret-exclusion checks, the six negative cases, and the validation result records. The scope limit holds callability as callability. It does not establish any goal, rerun Lean or Lake, widen Lean or Lake authority, use external model services, claim benchmark performance, export non-public paths, include launch operations, or treat tactic callability as proof quality.
\n",
"components": [
{
"anchor": "component-tactic_portfolio_availability_probe",
"component_id": "tactic_portfolio_availability_probe",
"public_label": "Tactic Portfolio Availability Probe"
}
],
"doctrine_feed": {
"counts": {
"axiom": 5,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 6
},
"link_count": 17,
"links": [
{
"id": "tactic_portfolio_availability_probe",
"kind": "organ",
"kind_label": "component",
"label": "Tactic Portfolio Availability Probe",
"relation": "explains",
"url": "component-tactic-portfolio-availability-probe.html"
},
{
"id": "mechanism.tactic_portfolio_availability_probe.validates_public_tactic_availability_projection",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public tactic availability projection",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-tactic-portfolio-availability-probe-validates-public-tactic-availability-projection"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.corpus_readiness_mathlib_absence_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Corpus Readiness Mathlib Absence Gate",
"relation": "depends on",
"url": "paper-module-corpus-readiness-mathlib-absence-gate.html"
},
{
"id": "paper_module.formal_math_readiness_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Readiness Gate",
"relation": "depends on",
"url": "paper-module-formal-math-readiness-gate.html"
},
{
"id": "paper_module.target_shape_tactic_routing",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Target Shape Tactic Routing",
"relation": "depends on",
"url": "paper-module-target-shape-tactic-routing.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.corpus_readiness_mathlib_absence_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Corpus Readiness Mathlib Absence Gate",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[8:paper_module.corpus_readiness_mathlib_absence_gate].depends_on",
"url": "paper-module-corpus-readiness-mathlib-absence-gate.html"
},
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[23:paper_module.formal_math_verifier_trace_repair_loop].depends_on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
},
{
"id": "paper_module.target_shape_tactic_routing",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Target Shape Tactic Routing",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[41:paper_module.target_shape_tactic_routing].depends_on",
"url": "paper-module-target-shape-tactic-routing.html"
},
{
"id": "paper_module.undeclared_library_prior_classifier",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Undeclared Library Prior Classifier",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[55:paper_module.undeclared_library_prior_classifier].depends_on",
"url": "paper-module-undeclared-library-prior-classifier.html"
},
{
"id": "paper_module.verifier_lab_kernel",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Kernel",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[0:paper_module.verifier_lab_kernel].depends_on",
"url": "paper-module-verifier-lab-kernel.html"
}
],
"paper_module_id": "paper_module.tactic_portfolio_availability",
"reader_packet": {
"abstract": "Environment-scoped tactic availability rows gate downstream tactic routing without becoming proof, benchmark, or launch-scope decision.",
"authority_ceiling": "Copied tactic affordance probe rows and public fixture/exported-bundle result records only; no Lean/Lake rerun, theorem proof, benchmark performance, external model access, Mathlib-dependent proof authority, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.tactic_portfolio_availability.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-tactic-portfolio-availability-diagram-1",
"html_url": "docs/paper-module-tactic-portfolio-availability.html#paper-module-tactic-portfolio-availability-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Copied affordance probe rows compile_status, requires_mathlib, duration_ms\"] --> B[\"_build_result\"] C[\"Mathlib environment probe\"] --> B B --> D{\"source compile_status\"} D -->|PASS| E[\"compile_pass band duration fast / moderate / slow\"] D -->|FAIL + requires Mathlib + Mathlib absent| F[\"environment_fail reason MATHLIB_IMPORT_MISSING\"] D -->|FAIL otherwise| G[\"compile_fail\"] E --> H[\"Availability board\"] F --> H G --> H B --> I[\"metadata-only result and validation records\"]",
"source_format": "mermaid",
"source_hash": "sha256:a1bb4c41e734a019a0c145f2a496ef95d233ea1fce1ab95391ce8a44926e9732",
"source_line_end": 127,
"source_line_start": 115
}
],
"grouped_relation_feeds": [
{
"count": 5,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.corpus_readiness_mathlib_absence_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Corpus Readiness Mathlib Absence Gate",
"relation": "depends on",
"url": "paper-module-corpus-readiness-mathlib-absence-gate.html"
},
{
"id": "paper_module.formal_math_readiness_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Readiness Gate",
"relation": "depends on",
"url": "paper-module-formal-math-readiness-gate.html"
},
{
"id": "paper_module.target_shape_tactic_routing",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Target Shape Tactic Routing",
"relation": "depends on",
"url": "paper-module-target-shape-tactic-routing.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "tactic_portfolio_availability_probe",
"kind": "organ",
"kind_label": "component",
"label": "Tactic Portfolio Availability Probe",
"relation": "explains",
"url": "component-tactic-portfolio-availability-probe.html"
},
{
"id": "mechanism.tactic_portfolio_availability_probe.validates_public_tactic_availability_projection",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public tactic availability projection",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-tactic-portfolio-availability-probe-validates-public-tactic-availability-projection"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.tactic_portfolio_availability",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 201,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/tactic_portfolio_availability.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It turns one captured Lean smoke run into an explicit list of callable tactics, checking pre-recorded status rows for honesty without running Lean itself.** tacticportfolioavailability_probe turns one captured Lean run into an explicit list of which proof tactics were callable, before any router or proof search treats a tactic as usable. It reads pre-recorded status rows from a real smoke run and checks them for honesty. It does not run Lean itself.",
"level": 4,
"order": 1,
"section_id": "it-turns-one-captured-lean-smoke-run-into-an-explicit-list",
"source_line_end": 11,
"source_line_start": 3,
"title": "It turns one captured Lean smoke run into an explicit list"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The copied 2026-05-11 fixture covers eight tactics, seven compiling and aesop failed on a missing Mathlib import, reproduced into a metadata-only board and result records.** The fixture is copied from a 2026-05-11 tactic affordance probe. It covers eight tactics: rfl, decide, omega, simp, simp_all, grind, native_decide, and aesop. Seven are recorded as compiling. aesop is recorded as failed because the paired environment probe reports that the Mathlib import it needs was absent. The component reproduces that verdict from the copied rows and writes a metadata-only availability board plus result records.",
"level": 4,
"order": 2,
"section_id": "the-copied-2026-05-11-fixture-covers-eight-tactics",
"source_line_end": 22,
"source_line_start": 12,
"title": "The copied 2026-05-11 fixture covers eight tactics"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 3,
"section_id": "purpose",
"source_line_end": 24,
"source_line_start": 23,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A tactic name alone does not mean it can run; a router trusting the bare name reaches for tactics the environment cannot run and misreads the failure as a fact about the goal.** A tactic name is not a usable tactic. aesop runs only if the surrounding Lean and Std environment carries the imports it needs. omega is callable in one project layout and not in another. A router that trusts a bare tactic name will reach for tactics the current environment cannot run, then misread the failure as a fact about the goal rather than a fact about the environment.",
"level": 4,
"order": 4,
"section_id": "a-tactic-name-is-not-availability",
"source_line_end": 34,
"source_line_start": 25,
"title": "A tactic name is not availability"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It answers which tactics were callable and on what evidence, keeping a missing import distinct from a broken tactic so a router can re-attempt elsewhere.** The component answers one question: in the observed environment, which tactics were callable, and on what evidence? The interesting part is how it treats failure. A recorded Lean FAIL is not flattened into one \"unavailable\" verdict. A missing import is kept distinct from a broken tactic, so a router can re-attempt the same tactic in a different environment instead of striking it off for good.",
"level": 4,
"order": 5,
"section_id": "missing-import-kept-distinct-from-broken-tactic",
"source_line_end": 45,
"source_line_start": 35,
"title": "Missing import kept distinct from broken tactic"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 6,
"section_id": "how-it-works",
"source_line_end": 47,
"source_line_start": 46,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**run validates the first-wave fixture with negatives on, runavailabilitybundle validates an exported bundle with negatives off and reuses an unchanged prior result.** Two entrypoints drive the pipeline. run validates the first-wave fixture with negative cases enabled. runavailabilitybundle validates an exported bundle with negative cases disabled, and reuses a prior result when freshavailabilitybundlereceipt finds a matching record whose inputs have not changed since it was written. Both delegate the work to buildresult and then persist records. buildresult runs these steps in order.",
"level": 4,
"order": 7,
"section_id": "two-entrypoints-over-build-result",
"source_line_end": 60,
"source_line_start": 48,
"title": "Two entrypoints over `_build_result`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It loads the three copied source payloads and digest-checks each body against a recorded SHA-256, so a silently edited body becomes a finding rather than trusted input. It loads the copied source payloads with loadsourcepayloads: the run-level affordance probe, the portfoliocorev0 availability artifact, and the paired corpus-readiness file. Each copied body is digest-checked by sourcebodyimports and probesourceimports against a recorded SHA-256, so a silently edited body becomes a finding rather than trusted input.",
"level": 4,
"order": 8,
"section_id": "digest-checking-the-copied-source-payloads",
"source_line_end": 70,
"source_line_start": 61,
"title": "Digest-checking the copied source payloads"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It derives the environment state from the Mathlib probe, whose single fact of whether the Lake import was available gates every Mathlib-dependent tactic.** It derives the environment state with authoritativeenvironment_state, which reads the Mathlib probe and reports whether the Mathlib Lake import was available. That single fact gates every Mathlib-dependent tactic.",
"level": 4,
"order": 9,
"section_id": "mathlib-probe-gates-dependent-tactics",
"source_line_end": 78,
"source_line_start": 71,
"title": "Mathlib probe gates dependent tactics"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It builds one row per tactic, mapping a source PASS to compilepass and a FAIL to environment_fail when a required Mathlib import was absent, else compilefail. It builds one availability row per tactic with availabilityrowsfromsource. The public status comes from derivedpublicstatusfrom_source: a source PASS becomes compilepass; a source FAIL becomes environmentfail when the tactic requires Mathlib and Mathlib was absent, and compile_fail otherwise. When the status is a failure, failureclassifierfromsource labels a Mathlib gap as MATHLIBIMPORTMISSING and otherwise carries the source error class through. This is why aesop reads as an environment failure, not a broken tactic.",
"level": 4,
"order": 10,
"section_id": "deriving-each-tactic-s-public-status",
"source_line_end": 91,
"source_line_start": 79,
"title": "Deriving each tactic's public status"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It summarises available and unavailable ids and bands each available tactic's copied duration as fast, moderate, or slow, stamped environment-scoped rather than a benchmark. It summarises the rows with portfolio_summary: available and unavailable tactic ids, Mathlib-dependent ids, and a latency profile. latencyband bands each available tactic by its copied duration as fast at 2500ms or less, moderate up to 5000ms, and slow above that. The latency profile is stamped environment-scoped, so a router can prefer a cheaper available tactic without reading speed as a benchmark result.",
"level": 4,
"order": 11,
"section_id": "portfolio-summary-and-latency-banding",
"source_line_end": 102,
"source_line_start": 92,
"title": "Portfolio summary and latency banding"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "**It evaluates the negative cases, scans every input path for forbidden material, and passes only with no finding, all negatives observed, and zero secret hits before writing result records. It evaluates the declared negative cases with negative_findings, scans all input paths for forbidden material with scan_paths, and sets the overall status. The result is pass only when no category produced a finding, every expected negative case was observed, and the secret scan had zero blocking hits. writereceipts then writes the result record, the availability board, the validation record, and the sign-off record, each carrying result record paths but no tactic bodies.",
"level": 4,
"order": 12,
"section_id": "negatives-secret-scan-and-receipts",
"source_line_end": 128,
"source_line_start": 103,
"title": "Negatives, secret scan, and result records"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 130,
"source_line_start": 129,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Six perturbed fixtures must each be rejected with a specific error code, spanning missing status, unbacked Mathlib claims, unprobed tactics, missing durations, leaked bodies, and authority overclaim. negative_findings checks six perturbed fixtures and requires each to be rejected with a specific error code. A tactic row with no compile_status gives TACTICPORTFOLIOMISSINGCOMPILESTATUS. A Mathlib-dependent tactic marked available with no passing Mathlib probe gives TACTICPORTFOLIOMATHLIBCLAIMWITHOUT_PROBE. A downstream request for a tactic not in the probed portfolio gives TACTICPORTFOLIOUNPROBEDTACTICREFERENCED. An available tactic with no copied duration gives TACTICPORTFOLIOAVAILABLEDURA\u2026",
"level": 4,
"order": 14,
"section_id": "six-perturbed-fixtures-and-their-codes",
"source_line_end": 146,
"source_line_start": 131,
"title": "Six perturbed fixtures and their codes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 148,
"source_line_start": 147,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**It follows the feature-detection habit of probing a local capability before relying on it, applying Autoconf and Lean's environment-sensitive discipline to one recorded run.** The component follows the feature-detection habit of testing a local capability before relying on it, the same discipline GNU Autoconf's configure workflow established for build environments. Lean's own tactic documentation notes that tactic use is environment- and goal-sensitive, so a tactic name alone does not justify downstream routing. This component applies that probe discipline to one recorded run and keeps a Mathlib-dependent absence as evidence rather than discarding it. GNU Autoconf feature probing: https:/\u2026",
"level": 4,
"order": 16,
"section_id": "feature-detection-discipline-from-autoconf-and-lean",
"source_line_end": 165,
"source_line_start": 149,
"title": "Feature-detection discipline from Autoconf and Lean"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate/, reproduce the evidence with temporary records:",
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 175,
"source_line_start": 166,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means every copied body matched its digest, the eight rows produced the expected board with aesop as an environment failure, all six negatives were rejected, and the secret scan was clean.** A pass means every copied source body matched its recorded digest, the eight tactic rows produced the expected availability board with aesop as an environment failure, all six negative cases were rejected, and the secret scan found no forbidden material.",
"level": 4,
"order": 18,
"section_id": "what-a-pass-certifies",
"source_line_end": 184,
"source_line_start": 176,
"title": "What a pass certifies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "claim-ceiling",
"source_line_end": 186,
"source_line_start": 185,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is that in one observed environment these tactics were callable on digest-matched probe rows; it proves no goal, reruns no Lean, and authorizes nothing.** The strongest claim the evidence supports is narrow: in one observed Lean and Std environment, these tactics were callable, on the evidence of copied probe rows whose bodies match recorded digests, with Mathlib absence preserved as a recorded fact. The proof boundary is those copied rows, the derived availability board, the digest and secret-exclusion checks, the six negative cases, and the validation result records. The scope limit holds callability as callability. It does not establish any goal, rerun Lean or Lake,\u2026",
"level": 4,
"order": 20,
"section_id": "callability-held-as-callability",
"source_line_end": 200,
"source_line_start": 187,
"title": "Callability held as callability"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/tactic_portfolio_availability.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Environment-scoped tactic availability rows gate downstream tactic routing without becoming proof, benchmark, or launch-scope decision."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "tactic_portfolio_availability_probe.py",
"path": "src/microcosm_core/organs/tactic_portfolio_availability_probe.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for tactic availability fixture validation, exported bundle validation, Mathlib absence gating, negative cases, and scope limits.",
"symbols": [
"run",
"run_availability_bundle",
"_build_result",
"_write_receipts",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/tactic_portfolio_availability_probe.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/tactic_portfolio_availability.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/tactic_portfolio_availability.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Tactic Portfolio Availability Probe validates copied Lean/Std tactic affordance rows before downstream routing can treat a tactic as usable. It checks compile status, Mathlib absence handling, probe portfolio membership, negative cases, source digests, and metadata-only result records while keeping proof bodies, model-output data, benchmark claims, and launch-scope decision out of scope.",
"summary_line": "Environment-scoped tactic availability rows gate downstream tactic routing without becoming proof, benchmark, or launch-scope decision.",
"title": "Tactic Portfolio Availability"
},
{
"anchor": "paper-module-target-shape-tactic-routing",
"authority_ceiling": "Public route-reference fixture and exported-bundle result records only; no Lean/Lake execution, formal-result correctness, private target material disclosure, post-execution routing, external model access, proof body export, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
Before any proof attempt it reads the candidate tactics for one goal and records an allow-or-reject decision for each. Before a proof is attempted, this component reads a list of candidate proof tactics for one goal and writes down an allow-or-reject decision for each. It runs in src/microcosm_core/organs/target_shape_tactic_routing_gate.py.
\n
It decides from goal shape and an environment probe without calling Lean, emitting a tactic-by-tactic record of what was admitted or blocked and why. The decision is made from evidence that already exists: the shape of the goal and a probe that says which tactics the declared environment can run. Nothing here calls Lean, runs a prover, or proves anything. The output is a record that shows, tactic by tactic, what was admitted or blocked and the plain reason, in place of an opaque "we tried these" claim.
\n
A rejection is one of three kinds: unprobed, known but unavailable, or available but wrong for the goal shape, kept separate so a reader sees which limit each hit. A rejection has one of three kinds. The tactic is not in the tested probe set, so it was never measured. It is known but marked unavailable in the declared environment. Or it is available but does not fit the goal shape. The record separates those three so a reader can see which limit each blocked tactic hit.
\n
Purpose
\n
Most proof cost is wasted on doomed tactics
\n
Most proof cost goes to tactics that were never going to work, and a loose gate that trusts the route's own choice cannot catch an unavailable or unprobed one. A proof attempt is expensive, and most of that cost is spent on tactics that were never going to work. A loose gate that trusts the route's own declared choice cannot catch this: the route is a claim about what should run, and a route that admits an unavailable or unprobed tactic passes a loose check while still wasting the later prover call.
\n
The gate recomputes the tactic choice
\n
The gate recomputes the selection from a per-shape preferred order, skipping and recording unselectable tactics, and flags a route whose declared choice disagrees. So the gate recomputes the choice rather than accepting the declared one. Each target shape carries a short preferred-tactic order, held in PREFERRED_TACTICS_BY_TARGET_SHAPE (for example omega for int_linear_arithmetic, decide for closed_nat_mod_decision). The gate walks that order, skips any preferred tactic it cannot select, records why it skipped, and falls back to the next allowed candidate or to DEFAULT_PREFERRED_TACTICS for a shape it does not recognise. A route whose declared selection disagrees with the computed one is flagged, not honoured.
\n
How it works
\n
_build_result orders the pipeline
\n
_build_result loads the two inputs, scans for forbidden material, builds availability sets, scores each route, folds in negatives, and sets one pass-or-blocked status._build_result runs the pipeline in order. It loads the two input files through _load_payloads, scans every input path for forbidden material with scan_paths, builds the availability sets with _portfolio, sanitizes and scores each route case, checks the copied source artifacts, folds in the negative-case fixtures, then sets a single pass-or-blocked status and emits the result and board.
\n
_portfolio splits the probe into three sets
\n
_portfolio splits the probe into known, available, and unavailable ids, counting a tactic available only when its status is one of four passing values._portfolio reads the tactic probe file and splits it into three sets: known_tactic_ids (every tactic listed), available_tactic_ids, and unavailable_tactic_ids. A tactic counts as available only when its status is one of available, pass, compiled, or compile_pass.
\n
_score_case compares declared against computed
\n
_score_case reads a case's allowed and candidate lists, gets a per-tactic verdict, computes the selection, and compares the declared tactic against the computed one._score_case scores one route case. It reads the case's allowed_tactic_ids and candidate list, asks _decision_for_tactic for a verdict on each candidate, computes the selection with _shape_preferred_selection, and compares the case's declared selected_tactic_id against the computed one.
\n
_decision_for_tactic rejects in a fixed order
\n
_decision_for_tactic rejects in order as unprobed, unavailable, or shape-inadmissible, admitting only a tactic that is probed, available, and allowed for the shape._decision_for_tactic is the per-tactic gate. It rejects in a fixed order: a tactic absent from known is UNPROBED_TACTIC, a known tactic absent from available is UNAVAILABLE_TACTIC, an available tactic absent from allowed is TARGET_SHAPE_ADMISSIBILITY_REJECTED. Only a tactic that is probed, available, and allowed for the shape gets TARGET_SHAPE_ADMISSIBLE.
\n
_shape_preferred_selection picks the first allowed tactic
\n
_shape_preferred_selection walks the shape's preferred order and returns the first allowed tactic, recording skips, the default-used flag, and a preferred-unavailable case._shape_preferred_selection walks the preferred order for the shape and returns the first tactic that is allowed. It records the skipped preferred tactics and their reasons, sets unknown_shape_default_used when the shape has no specific map, and marks the preferred-unavailable case when the first preferred tactic is known but not usable.
\n
_route_integrity_findings types each per-case problem
\n
_route_integrity_findings turns any per-case problem into a typed finding: an admitted unavailable or unprobed tactic, a post-execution stage, or a selection mismatch._route_integrity_findings turns any per-case problem into a typed finding: an admitted unavailable tactic, an admitted unprobed tactic, a post-execution route stage, or a declared selection that does not match the computed preference.
\n
Sanitizing private goal material
\n
The sanitizer keeps the public route shape but drops forbidden private-target keys and masks private-looking strings before any public row, reporting only the field path and class._private_target_material_findings and _sanitize_private_target_material handle the scope limit on private goals. The sanitizer keeps the public route shape but drops any key in FORBIDDEN_PRIVATE_TARGET_KEYS (private target ids, formulas, target statuses, solver and derivation traces, proof-packet and certificate-search refs, submission packages) and replaces private-looking strings with private_target_material_omitted before any public row is built. The findings function reports the rejected field path and class, never the matched value.
\n
Source-digest checks and the two entry points
\n
_source_artifact_imports re-hashes the four copied source files and marks each copy, rewrite, or mismatch, while run includes the negative cases and run_routing_bundle omits them._source_artifact_imports re-hashes the four copied source files named in REAL_SUBSTRATE_REFS and marks each as an exact copy, a path rewrite, or a digest mismatch. write_receipts writes the result, board, validation, and sign-off records. run drives the fixture with the negative cases included; run_routing_bundle drives the exported bundle without them.
\n
Function
Role
_build_result
Orders the pipeline and sets the pass-or-blocked status
_portfolio
Splits the probe into known, available, unavailable sets
_score_case
Scores one route case and compares declared against computed
_decision_for_tactic
Rejects a tactic as unprobed, unavailable, or not allowed
_shape_preferred_selection
Computes the preferred selection and records skips
_route_integrity_findings
Turns per-case problems into typed findings
_sanitize_private_target_material
Drops private goal keys before public rows
write_receipts
Writes the result, board, validation, sign-off records
\nDiagram of the mechanism (8 steps).Diagram source & refs
flowchart TD Inputs["Probe file + route cases"] Portfolio["_portfolio known / available / unavailable"] Sanitize["_sanitize_private_target_material drop private goal keys"] Decide["_decision_for_tactic per-tactic verdict"] Select["_shape_preferred_selection compute the selection"] Score["_score_case declared vs computed"] Findings["_route_integrity_findings typed rejections"] Result["_build_result status + board"] Inputs --> Portfolio Inputs --> Sanitize Portfolio --> Decide Sanitize --> Decide Decide --> Select Select --> Score Score --> Findings Findings --> Result
\n
What the exported bundle scores to
\n
On the exported bundle four tactics are available and aesop unavailable, all seven route cases are pre-execution, and four source files are copied in with digest checks. On the exported bundle the probe marks decide, omega, simp_all, and rfl available and aesop unavailable. All seven route cases are pre_execution, and four source files are copied in with digest checks.
\n
Negative cases
\n
The gate must reject each of these, listed in EXPECTED_NEGATIVE_CASES: The gate must reject each of these, listed in EXPECTED_NEGATIVE_CASES:
\n
unavailable_tactic_admitted: a route admits a tactic the probe marks unavailable, here aesop while Mathlib is absent.
unprobed_tactic_allowed: a route admits a tactic missing from the probe.
shape_preferred_tactic_overridden: a declared tactic bypasses the computed shape preference.
unknown_shape_fallback: a shape bypasses the default safe order.
preferred_unavailable_fallback: a route bypasses the next available tactic after the first preferred one is unavailable.
proof_body_leakage: a route carries a proof, provider, or Lean body field.
private_target_material: a route carries a private target id, formula, status, trace, or challenge path, rejected without echoing the value.
post_execution_route: a route selection carries execution evidence.
release_overclaim: a route claims proof, provider, Lean, or launch-scope decision.
\n
Prior Art Grounding
\n
A pre-execution proof-search admissibility filter
\n
It follows the proof-search habit of matching a goal shape to available methods before spending runtime, narrowed to a pre-execution admissibility filter over shape, references, and availability. The routing layer follows a known proof-search pattern: match a goal shape to methods that are known to be available before spending runtime on them. Lean's tactic documentation gives the proof-assistant context for goal-directed tactic choice, and Isabelle's Sledgehammer is a mature example of selecting external provers and relevant facts from a goal. This component narrows the idea to a pre-execution admissibility filter: shape, allowed references, and current availability must line up before a route is exported.
Isabelle Sledgehammer user guide: https://isabelle.in.tum.de/doc/sledgehammer.pdf
\n
Validation Result record Path
\n
From microcosm-substrate/, reproduce the result with temporary result records:
\n
What a pass certifies
\n
A pass means the seven cases scored as expected, the four source copies matched digests, every negative was rejected, and the secret scan was clean, without widening the ceiling. A pass means the seven route cases scored as expected, the four source copies matched their digests, every negative case was rejected, and the secret scan found no blocking hit. It does not widen the no-Lean, no-proof ceiling.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
For public pre-execution cases the gate admits only probed, available tactics before any proof attempt and recomputes the selection so a route cannot smuggle in an inadmissible one. The strongest claim the evidence supports: for public pre-execution route cases, this gate admits only tactics that were both probed and available before any proof attempt, and it recomputes the selection so a route cannot smuggle in an unavailable, unprobed, or shape-inadmissible tactic by declaring it.
\n
The proof boundary and scope limit
\n
The proof boundary is the route references it reads, and AUTHORITY_CEILING stops it from running Lean, proving a goal, exposing private material, or granting launch control. The proof boundary is the route references it reads: tactic probe availability, target-shape cases, selected tactic ids, source-digest evidence, the negative-case records, and the statement-scope omission records. The scope limit is set in AUTHORITY_CEILING. The gate does not run Lean or Lake, prove a goal, export proof or provider bodies, expose private target material, authorize post-execution route selection, use external model services, claim Mathlib-dependent proof authority, or authorize public sharing or launch.
\n",
"components": [
{
"anchor": "component-target_shape_tactic_routing_gate",
"component_id": "target_shape_tactic_routing_gate",
"public_label": "Target Shape Tactic Routing Gate"
}
],
"doctrine_feed": {
"counts": {
"axiom": 5,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 5,
"principle": 6
},
"link_count": 19,
"links": [
{
"id": "target_shape_tactic_routing_gate",
"kind": "organ",
"kind_label": "component",
"label": "Target Shape Tactic Routing Gate",
"relation": "explains",
"url": "component-target-shape-tactic-routing-gate.html"
},
{
"id": "mechanism.target_shape_tactic_routing_gate.validates_public_tactic_routing_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public tactic routing boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-target-shape-tactic-routing-gate-validates-public-tactic-routing-boundary"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depends on",
"url": "paper-module-tactic-portfolio-availability.html"
},
{
"id": "paper_module.formal_math_readiness_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Readiness Gate",
"relation": "depends on",
"url": "paper-module-formal-math-readiness-gate.html"
},
{
"id": "paper_module.proof_diagnostic_evidence_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof Diagnostic Evidence Spine",
"relation": "depends on",
"url": "paper-module-proof-diagnostic-evidence-spine.html"
},
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depends on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
},
{
"id": "paper_module.formal_evidence_cell_anchor_resolver",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Evidence Cell Anchor Resolver",
"relation": "depends on",
"url": "paper-module-formal-evidence-cell-anchor-resolver.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[23:paper_module.formal_math_verifier_trace_repair_loop].depends_on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
},
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[40:paper_module.tactic_portfolio_availability].depends_on",
"url": "paper-module-tactic-portfolio-availability.html"
},
{
"id": "paper_module.verifier_lab_kernel",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Kernel",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[0:paper_module.verifier_lab_kernel].depends_on",
"url": "paper-module-verifier-lab-kernel.html"
}
],
"paper_module_id": "paper_module.target_shape_tactic_routing",
"reader_packet": {
"abstract": "Pre-execution tactic routing admits or rejects tactics from target shape and probe evidence while excluding private target material.",
"authority_ceiling": "Public route-reference fixture and exported-bundle result records only; no Lean/Lake execution, formal-result correctness, private target material disclosure, post-execution routing, external model access, proof body export, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.target_shape_tactic_routing.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-target-shape-tactic-routing-diagram-1",
"html_url": "docs/paper-module-target-shape-tactic-routing.html#paper-module-target-shape-tactic-routing-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Inputs[\"Probe file + route cases\"] Portfolio[\"_portfolio known / available / unavailable\"] Sanitize[\"_sanitize_private_target_material drop private goal keys\"] Decide[\"_decision_for_tactic per-tactic verdict\"] Select[\"_shape_preferred_selection compute the selection\"] Score[\"_score_case declared vs computed\"] Findings[\"_route_integrity_findings typed rejections\"] Result[\"_build_result status + board\"] Inputs --> Portfolio Inputs --> Sanitize Portfolio --> Decide Sanitize --> Decide Decide --> Select Select --> Score Score --> Findings Findings --> Result",
"source_format": "mermaid",
"source_hash": "sha256:16e4db6a299fc1cbe62c5c3db3006af51516304dc4f7546eab6c7eb94db1f9a7",
"source_line_end": 167,
"source_line_start": 149
}
],
"grouped_relation_feeds": [
{
"count": 5,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 5,
"links": [
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depends on",
"url": "paper-module-tactic-portfolio-availability.html"
},
{
"id": "paper_module.formal_math_readiness_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Readiness Gate",
"relation": "depends on",
"url": "paper-module-formal-math-readiness-gate.html"
},
{
"id": "paper_module.proof_diagnostic_evidence_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof Diagnostic Evidence Spine",
"relation": "depends on",
"url": "paper-module-proof-diagnostic-evidence-spine.html"
},
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depends on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
},
{
"id": "paper_module.formal_evidence_cell_anchor_resolver",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Evidence Cell Anchor Resolver",
"relation": "depends on",
"url": "paper-module-formal-evidence-cell-anchor-resolver.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "target_shape_tactic_routing_gate",
"kind": "organ",
"kind_label": "component",
"label": "Target Shape Tactic Routing Gate",
"relation": "explains",
"url": "component-target-shape-tactic-routing-gate.html"
},
{
"id": "mechanism.target_shape_tactic_routing_gate.validates_public_tactic_routing_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public tactic routing boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-target-shape-tactic-routing-gate-validates-public-tactic-routing-boundary"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.target_shape_tactic_routing",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 260,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/target_shape_tactic_routing.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Before any proof attempt it reads the candidate tactics for one goal and records an allow-or-reject decision for each.** Before a proof is attempted, this component reads a list of candidate proof tactics for one goal and writes down an allow-or-reject decision for each. It runs in src/microcosmcore/components/targetshapetacticrouting_gate.py.",
"level": 4,
"order": 1,
"section_id": "before-any-proof-attempt-it-reads-the-candidate-tactics-for-one-goal",
"source_line_end": 10,
"source_line_start": 3,
"title": "Before any proof attempt it reads the candidate tactics for one goal"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It decides from goal shape and an environment probe without calling Lean, emitting a tactic-by-tactic record of what was admitted or blocked and why.** The decision is made from evidence that already exists: the shape of the goal and a probe that says which tactics the declared environment can run. Nothing here calls Lean, runs a prover, or proves anything. The output is a record that shows, tactic by tactic, what was admitted or blocked and the plain reason, in place of an opaque \"we tried these\" claim.",
"level": 4,
"order": 2,
"section_id": "it-decides-from-goal-shape-and-an-environment-probe-without-calling-lean",
"source_line_end": 20,
"source_line_start": 11,
"title": "It decides from goal shape and an environment probe without calling Lean"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A rejection is one of three kinds: unprobed, known but unavailable, or available but wrong for the goal shape, kept separate so a reader sees which limit each hit.** A rejection has one of three kinds. The tactic is not in the tested probe set, so it was never measured. It is known but marked unavailable in the declared environment. Or it is available but does not fit the goal shape. The record separates those three so a reader can see which limit each blocked tactic hit.",
"level": 4,
"order": 3,
"section_id": "a-rejection-is-one-of-three-kinds",
"source_line_end": 29,
"source_line_start": 21,
"title": "A rejection is one of three kinds"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 31,
"source_line_start": 30,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Most proof cost goes to tactics that were never going to work, and a loose gate that trusts the route's own choice cannot catch an unavailable or unprobed one.** A proof attempt is expensive, and most of that cost is spent on tactics that were never going to work. A loose gate that trusts the route's own declared choice cannot catch this: the route is a claim about what should run, and a route that admits an unavailable or unprobed tactic passes a loose check while still wasting the later prover call.",
"level": 4,
"order": 5,
"section_id": "most-proof-cost-is-wasted-on-doomed-tactics",
"source_line_end": 41,
"source_line_start": 32,
"title": "Most proof cost is wasted on doomed tactics"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The gate recomputes the selection from a per-shape preferred order, skipping and recording unselectable tactics, and flags a route whose declared choice disagrees.** So the gate recomputes the choice rather than accepting the declared one. Each target shape carries a short preferred-tactic order, held in PREFERREDTACTICSBYTARGETSHAPE (for example omega for intlineararithmetic, decide for closednatmod_decision). The gate walks that order, skips any preferred tactic it cannot select, records why it skipped, and falls back to the next allowed candidate or to DEFAULTPREFERREDTACTICS for a shape it does not recognise. A route whose declared selection disagrees with the computed one is flagged,\u2026",
"level": 4,
"order": 6,
"section_id": "the-gate-recomputes-the-tactic-choice",
"source_line_end": 54,
"source_line_start": 42,
"title": "The gate recomputes the tactic choice"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 56,
"source_line_start": 55,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult loads the two inputs, scans for forbidden material, builds availability sets, scores each route, folds in negatives, and sets one pass-or-blocked status. build_result runs the pipeline in order. It loads the two input files through loadpayloads, scans every input path for forbidden material with scanpaths, builds the availability sets with portfolio, sanitizes and scores each route case, checks the copied source artifacts, folds in the negative-case fixtures, then sets a single pass-or-blocked status and emits the result and board.",
"level": 4,
"order": 8,
"section_id": "build-result-orders-the-pipeline",
"source_line_end": 67,
"source_line_start": 57,
"title": "`_build_result` orders the pipeline"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**_portfolio splits the probe into known, available, and unavailable ids, counting a tactic available only when its status is one of four passing values. portfolio reads the tactic probe file and splits it into three sets: knowntacticids (every tactic listed), availabletacticids, and unavailabletacticids. A tactic counts as available only when its status is one of available, pass, compiled, or compile_pass.",
"level": 4,
"order": 9,
"section_id": "portfolio-splits-the-probe-into-three-sets",
"source_line_end": 76,
"source_line_start": 68,
"title": "`_portfolio` splits the probe into three sets"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "scorecase reads a case's allowed and candidate lists, gets a per-tactic verdict, computes the selection, and compares the declared tactic against the computed one. scorecase scores one route case. It reads the case's allowedtactic_ids and candidate list, asks decisionfor_tactic for a verdict on each candidate, computes the selection with shapepreferred_selection, and compares the case's declared selectedtacticid against the computed one.",
"level": 4,
"order": 10,
"section_id": "score-case-compares-declared-against-computed",
"source_line_end": 85,
"source_line_start": 77,
"title": "`_score_case` compares declared against computed"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "decisionfor_tactic rejects in order as unprobed, unavailable, or shape-inadmissible, admitting only a tactic that is probed, available, and allowed for the shape. decisionfortactic is the per-tactic gate. It rejects in a fixed order: a tactic absent from known is UNPROBED_TACTIC, a known tactic absent from available is UNAVAILABLE_TACTIC, an available tactic absent from allowed is TARGETSHAPEADMISSIBILITY_REJECTED. Only a tactic that is probed, available, and allowed for the shape gets TARGETSHAPEADMISSIBLE.",
"level": 4,
"order": 11,
"section_id": "decision-for-tactic-rejects-in-a-fixed-order",
"source_line_end": 95,
"source_line_start": 86,
"title": "`_decision_for_tactic` rejects in a fixed order"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "shapepreferred_selection walks the shape's preferred order and returns the first allowed tactic, recording skips, the default-used flag, and a preferred-unavailable case. shapepreferredselection walks the preferred order for the shape and returns the first tactic that is allowed. It records the skipped preferred tactics and their reasons, sets unknownshapedefault_used when the shape has no specific map, and marks the preferred-unavailable case when the first preferred tactic is known but not usable.",
"level": 4,
"order": 12,
"section_id": "shape-preferred-selection-picks-the-first-allowed-tactic",
"source_line_end": 105,
"source_line_start": 96,
"title": "`_shape_preferred_selection` picks the first allowed tactic"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "routeintegrity_findings turns any per-case problem into a typed finding: an admitted unavailable or unprobed tactic, a post-execution stage, or a selection mismatch. routeintegrityfindings turns any per-case problem into a typed finding: an admitted unavailable tactic, an admitted unprobed tactic, a post-execution route stage, or a declared selection that does not match the computed preference.",
"level": 4,
"order": 13,
"section_id": "route-integrity-findings-types-each-per-case-problem",
"source_line_end": 114,
"source_line_start": 106,
"title": "`_route_integrity_findings` types each per-case problem"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The sanitizer keeps the public route shape but drops forbidden private-target keys and masks private-looking strings before any public row, reporting only the field path and class. privatetargetmaterialfindings and sanitizeprivatetarget_material handle the scope limit on private goals. The sanitizer keeps the public route shape but drops any key in FORBIDDENPRIVATETARGET_KEYS (private target ids, formulas, target statuses, solver and derivation traces, proof-packet and certificate-search refs, submission packages) and replaces private-looking strings with privatetargetmaterial_omitted before any public row is built. The findings function reports the rejected field path and class, never th\u2026",
"level": 4,
"order": 14,
"section_id": "sanitizing-private-goal-material",
"source_line_end": 127,
"source_line_start": 115,
"title": "Sanitizing private goal material"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "sourceartifact_imports re-hashes the four copied source files and marks each copy, rewrite, or mismatch, while run includes the negative cases and runroutingbundle omits them. sourceartifactimports re-hashes the four copied source files named in REALSUBSTRATEREFS and marks each as an exact copy, a path rewrite, or a digest mismatch. write_receipts writes the result, board, validation, and sign-off records. run drives the fixture with the negative cases included; runroutingbundle drives the exported bundle without them.",
"level": 4,
"order": 15,
"section_id": "source-digest-checks-and-the-two-entry-points",
"source_line_end": 168,
"source_line_start": 128,
"title": "Source-digest checks and the two entry points"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**On the exported bundle four tactics are available and aesop unavailable, all seven route cases are pre-execution, and four source files are copied in with digest checks. On the exported bundle the probe marks decide, omega, simpall, and rfl available and aesop unavailable. All seven route cases are pre_execution, and four source files are copied in with digest checks.",
"level": 4,
"order": 16,
"section_id": "what-the-exported-bundle-scores-to",
"source_line_end": 176,
"source_line_start": 169,
"title": "What the exported bundle scores to"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "negative-cases",
"source_line_end": 178,
"source_line_start": 177,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The gate must reject each of these, listed in EXPECTEDNEGATIVECASES: The gate must reject each of these, listed in EXPECTEDNEGATIVECASES: unavailabletacticadmitted: a route admits a tactic the probe marks unavailable, here aesop while Mathlib is absent. unprobedtacticallowed: a route admits a tactic missing from the probe. shapepreferredtactic_overridden: a declared tactic bypasses the computed shape preference. unknownshapefallback: a shape bypasses the default safe order. preferredunavailablefallback: a route bypasses the next available tactic after the first preferred one is unavailable. proofbodyleakage: a route carries a proof, provider, or Lean body field. privatetargetmaterial: a rou\u2026",
"level": 4,
"order": 18,
"section_id": "the-gate-must-reject-each-of-these-listed-in",
"source_line_end": 198,
"source_line_start": 179,
"title": "The gate must reject each of these, listed in"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "prior-art-grounding",
"source_line_end": 200,
"source_line_start": 199,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**It follows the proof-search habit of matching a goal shape to available methods before spending runtime, narrowed to a pre-execution admissibility filter over shape, references, and availability.** The routing layer follows a known proof-search pattern: match a goal shape to methods that are known to be available before spending runtime on them. Lean's tactic documentation gives the proof-assistant context for goal-directed tactic choice, and Isabelle's Sledgehammer is a mature example of selecting external provers and relevant facts from a goal. This component narrows the idea to a pre-execution admissibility filter: shape, allowed references, and current availability must line up before\u2026",
"level": 4,
"order": 20,
"section_id": "a-pre-execution-proof-search-admissibility-filter",
"source_line_end": 217,
"source_line_start": 201,
"title": "A pre-execution proof-search admissibility filter"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate/, reproduce the result with temporary result records:",
"level": 2,
"order": 21,
"section_id": "validation-receipt-path",
"source_line_end": 228,
"source_line_start": 218,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the seven cases scored as expected, the four source copies matched digests, every negative was rejected, and the secret scan was clean, without widening the ceiling.** A pass means the seven route cases scored as expected, the four source copies matched their digests, every negative case was rejected, and the secret scan found no blocking hit. It does not widen the no-Lean, no-proof ceiling.",
"level": 4,
"order": 22,
"section_id": "what-a-pass-certifies",
"source_line_end": 236,
"source_line_start": 229,
"title": "What a pass certifies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 23,
"section_id": "claim-ceiling",
"source_line_end": 238,
"source_line_start": 237,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**For public pre-execution cases the gate admits only probed, available tactics before any proof attempt and recomputes the selection so a route cannot smuggle in an inadmissible one.** The strongest claim the evidence supports: for public pre-execution route cases, this gate admits only tactics that were both probed and available before any proof attempt, and it recomputes the selection so a route cannot smuggle in an unavailable, unprobed, or shape-inadmissible tactic by declaring it.",
"level": 4,
"order": 24,
"section_id": "the-strongest-supported-claim",
"source_line_end": 247,
"source_line_start": 239,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The proof boundary is the route references it reads, and scope_limit stops it from running Lean, proving a goal, exposing private material, or granting launch control.** The proof boundary is the route references it reads: tactic probe availability, target-shape cases, selected tactic ids, source-digest evidence, the negative-case records, and the statement-scope omission records. The scope limit is set in scope_limit. The gate does not run Lean or Lake, prove a goal, export proof or provider bodies, expose private target material, authorize post-execution route selection, use external model services, claim Mathlib-dependent proof authority, or authorize public sharing or launch.",
"level": 4,
"order": 25,
"section_id": "the-proof-boundary-and-authority-ceiling",
"source_line_end": 259,
"source_line_start": 248,
"title": "The proof boundary and scope limit"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/target_shape_tactic_routing.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Pre-execution tactic routing admits or rejects tactics from target shape and probe evidence while excluding private target material."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "target_shape_tactic_routing_gate.py",
"path": "src/microcosm_core/organs/target_shape_tactic_routing_gate.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for target-shape routing fixture validation, exported bundle validation, unavailable/unprobed tactic refusals, statement-scope private-target exclusion, negative cases, and scope limits.",
"symbols": [
"run",
"run_routing_bundle",
"_build_result",
"write_receipts",
"EXPECTED_NEGATIVE_CASES",
"STATEMENT_SCOPE_BOUNDARY",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/target_shape_tactic_routing_gate.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/target_shape_tactic_routing.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/target_shape_tactic_routing.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Target Shape Tactic Routing is the pre-execution gate between tactic availability and proof attempts. It validates Ring2 problem-domain, failure-class, graph-update, and tactic-probe references, rejects unavailable or unprobed tactics, rejects private target ids/formulas/statuses/traces without echoing values, and records metadata-only route decisions before any Lean/Lake proof authority is claimed.",
"summary_line": "Pre-execution tactic routing admits or rejects tactics from target shape and probe evidence while excluding private target material.",
"title": "Target Shape Tactic Routing"
},
{
"anchor": "paper-module-ring2-premise-precision-recall",
"authority_ceiling": "Copied Ring2 retrieval records and public fixture/exported-bundle result records only; no Lean/Lake execution, theorem proof, benchmark performance, external model access, provider-context label flow, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
Premise-retrieval precision and recall replay
\n
It replays copied premise-retrieval records after the fact and reports, problem by problem, how much of the needed support the search actually ranked.ring2_premise_retrieval_precision_recall_harness scores saved premise-retrieval runs after the fact. A proof search first looks for the supporting facts a proof will need, the premises. This component replays copied records of that search and reports, problem by problem, how much of the needed support the search actually ranked.
\n
It keeps two failures apart, a missed premise versus a proof that had everything and still failed, by computing precision and recall and labelling each problem one of four outcomes. The point is to keep two failures apart. A proof can fail because the search never surfaced a premise it needed. A proof can also fail even though every needed premise was ranked. Those call for different fixes. The component computes precision and recall from the copied records, then labels each problem as one of four outcomes so a reader can tell which case they are looking at.
\n
It emits measurement result records, bounded evidence authority, and the labels may never flow back into the ranking or into a provider recipe. It emits result records, bounded evidence authority. The labels are measurement labels. They score a finished run and they are never allowed to flow back into the ranking or into a provider recipe.
\n
Purpose
\n
Blaming the prover hides a missed premise
\n
A failed search invites blaming the prover, but a single reported precision figure cannot say whether the fault was a missed premise or a proof that had everything and still failed. When a proof search fails it is easy to blame the prover and miss the simpler cause: the right facts were never put in front of it. A single reported precision figure hides this. It cannot say whether a low score came from a missed premise or from a proof that had everything it needed and still did not go through.
\n
Recomputed overlap read against proof outcome
\n
It recomputes the numbers by intersecting ranked ids with needed ids and reading the proof outcome, so full recall with a failing proof surfaces as a downstream fault. So the component recomputes the numbers rather than trusting a reported one. For each problem it intersects the ranked premise ids with the labelled needed-premise ids and reads the proof outcome alongside that overlap. Full recall with a passing proof is one outcome; full recall with a failing proof is a different one. That second case is the signal that the fault lies downstream of retrieval, and a loose metric that only averages scores would erase it.
\n
How it works
\n
Two entrypoints over one _build_result
\n
run validates the fixture with five negative cases and run_precision_recall_bundle validates the export without them, both calling _build_result to set one PASS or blocked status. Two entrypoints drive the same core. run validates a fixture directory with the five negative cases included. run_precision_recall_bundle validates an exported bundle with the negative cases turned off. Both call _build_result, which runs the stages below in order and decides a single status of PASS or blocked.
\n
Function
Role
_load_payloads
reads the four input files (retrieval_runs, problem_labels, retrieval_rankings, evaluation_policy) plus the negative inputs when they are included
_evaluate
computes precision and recall per problem and assigns the four-way failure mode
_validate_run_material
checks that copied run material carries its source ref, digest, target refs, and provenance status
_validate_source_artifacts
checks the four copied source files against a known digest
_negative_findings
turns the five adversarial inputs into typed refusal codes
_write_receipts
writes the result, board, validation, and sign-off records with proof bodies left out
\n
Precision and recall at a top-k cutoff
\n
_evaluate truncates each ranking to top_k, intersects ranked with needed ids for hits, and computes precision and recall against the policy's expected aggregates. The scoring core is _evaluate. It indexes the after-the-fact labels by problem_id, reads default_top_k from the policy (4 if unset) or a per-ranking top_k, and truncates each ranking to that cutoff. It intersects the truncated ranked ids with the labelled needed ids to get the hits, then computes precision = hits / top_k and recall = hits / needed. Aggregate precision and recall come from summed hit, candidate, and needed counts, and are compared against the policy's expected_aggregate_metrics.
\n
The four-way label is mechanical, driven by recall and the recorded proof_outcome: The four-way label is mechanical, driven by recall and the recorded proof_outcome:
\n
retrieval_hit: recall reached 1.0 and the proof passed.
proof_failure_despite_hit: recall reached 1.0 but the proof did not pass. Retrieval did its job; the fault is elsewhere.
partial_retrieval_miss: some needed premises were ranked, not all.
retrieval_miss: none of the needed premises were ranked.
\n
Enforcing the shape of the evaluation set
\n
_evaluate also requires every expected failure mode, an adversarial decoy problem, and the minimum problem count, turning any gap into a finding._evaluate also enforces the shape of the evaluation set, not just the happy path. It requires every mode named in the policy's expected_failure_modes to appear, requires an adversarial decoy problem whose needed premise is absent or missed, and requires the policy's minimum problem count. A gap in any of these becomes a finding.
\n
Provenance and digest checks gate the status
\n
Before any score is trusted, _validate_run_material and _validate_source_artifacts check provenance and digests, and status is PASS only when every stage and the secret scan clear. Before any score is trusted, the provenance stages run. _validate_run_material rejects copied material that is missing a source_ref, source_sha256, target_refs, validation_refs, or the expected copied_non_secret_macro_body_with_provenance status. _validate_source_artifacts walks the four files in SOURCE_REFS, hashes each copied artifact under source_artifacts/, and accepts it only if the hash matches the recorded exact digest in SOURCE_DIGESTS or a known path-rewrite digest. A missing file or a hash mismatch is a finding. _build_result also runs a secret-exclusion scan over the inputs. The overall status is PASS only when there are no floor findings, no run-material findings, the source artifacts pass, the evaluation passes, no expected negative case is missing, and the secret scan has no blocking hit. Otherwise it stays blocked.
\n
Result records expose metrics, never proof bodies
\n
_write_receipts emits import ids, target refs, digest status, and metric counts, but writes no proof bodies, model-output data, or non-public paths._write_receipts then emits the records. It exposes import ids, target refs, digest status, aggregate metrics, and failure-mode counts. It does not write proof bodies, model-output data, or non-public paths.
\nDiagram of the mechanism (8 steps).Diagram source & refs
flowchart TD Inputs["four input files runs, labels, rankings, policy"] --> Build["_build_result orchestrates stages, sets status"] Build --> Evaluate["_evaluate precision/recall + four-way label"] Build --> RunMat["_validate_run_material copied-material provenance"] Build --> Artifacts["_validate_source_artifacts digest match per file"] Build --> Negatives["_negative_findings five typed refusals"] Evaluate --> Status["status: PASS or blocked"] RunMat --> Status Artifacts --> Status Negatives --> Status Status --> Result records["_write_receipts metrics only, no proof bodies"]
\n
Negative cases
\n
EXPECTED_NEGATIVE_CASES names five adversarial inputs that run must reject, each mapped to one refusal code:EXPECTED_NEGATIVE_CASES names five adversarial inputs that run must reject, each mapped to one refusal code:
\n
oracle_labels_in_ranking plants the known-correct premise ids inside a ranking. _inspect_forbidden_bodies catches the oracle_needed_premise_ids field and emits RING2_RETRIEVAL_ORACLE_LABELS_IN_RANKING.
proof_body_leakage carries a proof, provider, or private body field and emits RING2_RETRIEVAL_PROOF_BODY_FORBIDDEN.
test_split_tuning_attempt marks a run as tuned on the test labels and emits RING2_RETRIEVAL_TEST_SPLIT_TUNING_FORBIDDEN.
metric_overclaim asserts proof, benchmark, provider, or launch-scope decision and emits RING2_RETRIEVAL_METRIC_OVERCLAIM.
missing_adversarial_decoy omits the required decoy case and emits RING2_RETRIEVAL_ADVERSARIAL_DECOY_REQUIRED.
\n
Missing negative cases block the run
\n
If any of the five is not observed, _build_result records it as missing and the run is blocked. If any of the five is not observed, _build_result records it as missing and the run is blocked.
\n
Prior Art Grounding
\n
Grounded in IR metrics and premise selection
\n
The design draws on TREC and scikit-learn precision/recall evaluation and on hammer-style premise selection, where finding the right facts is a distinct step from replaying a proof. This component is grounded in information-retrieval evaluation. NIST's TREC evaluation measures give the older precision/recall frame for judging retrieval systems, and scikit-learn's precision/recall metric API is a common machine-learning interface for reporting those labels. On the theorem-proving side, premise selection sits next to hammer workflows such as Isabelle Sledgehammer, where finding the right facts is a distinct step from replaying a proof. This component keeps that distinction explicit and measurable.
\n
Validation Result record Path
\n
From microcosm-substrate/, reproduce this page's proof boundary with temporary records:
\n
What a pass over the records means
\n
A pass means the records scored to expectation, the four digests matched, and all five negative cases were rejected, without anything running Lean or a prover. A pass means the copied records scored to their expected metrics, the four source artifacts matched their digests, and all five negative cases were rejected. It does not mean anything ran Lean or a prover.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The strongest claim is that copied public records can be scored, sorted into the four labels, and guarded by five refusals, all reproducible from the fixture. The strongest claim this component supports is that copied public premise-retrieval records can be scored for precision and recall, sorted into the four retrieval-versus-proof-failure labels, and guarded by the five typed refusals, all reproducible from the fixture. The AUTHORITY_CEILING constant fixes the rest.
\n
What the labels never claim
\n
It runs no Lean, no providers, and no launch, and a retrieval_hit label says only that needed ids appeared in the ranking, nothing about proof soundness. It does not run Lean or Lake, use external model services, emit proof bodies, tune retrieval on test answers, claim benchmark performance, prove that a theorem is true, or include launch operations. A retrieval_hit label means the needed-premise ids appeared in the ranking under this fixture's policy, nothing about whether the downstream proof is sound. The reported numbers are bounded by the declared fixture and bundle; a different corpus, cutoff, or premise labelling has to be rerun. The proof boundary is the copied records and public result records only, and the scope limit holds even where site projections make this page walkable. Its labels are measurement labels and they may not enter a provider-context recipe.
\n",
"components": [
{
"anchor": "component-ring2_premise_retrieval_precision_recall_harness",
"component_id": "ring2_premise_retrieval_precision_recall_harness",
"public_label": "Ring2 Premise Retrieval Precision Recall Harness"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 2,
"principle": 6
},
"link_count": 17,
"links": [
{
"id": "ring2_premise_retrieval_precision_recall_harness",
"kind": "organ",
"kind_label": "component",
"label": "Ring2 Premise Retrieval Precision Recall Harness",
"relation": "explains",
"url": "component-ring2-premise-retrieval-precision-recall-harness.html"
},
{
"id": "mechanism.ring2_premise_retrieval_precision_recall_harness.validates_public_premise_retrieval_attribution",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public premise retrieval attribution",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-ring2-premise-retrieval-precision-recall-harness-validates-public-premise-retrieval-attribution"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.formal_math_premise_retrieval",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Premise Retrieval",
"relation": "depends on",
"url": "paper-module-formal-math-premise-retrieval.html"
},
{
"id": "paper_module.proof_diagnostic_evidence_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof Diagnostic Evidence Spine",
"relation": "depends on",
"url": "paper-module-proof-diagnostic-evidence-spine.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.mathematical_strategy_atlas",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mathematical Strategy Atlas",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[43:paper_module.mathematical_strategy_atlas].depends_on",
"url": "paper-module-mathematical-strategy-atlas.html"
},
{
"id": "paper_module.verifier_lab_kernel",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Kernel",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[0:paper_module.verifier_lab_kernel].depends_on",
"url": "paper-module-verifier-lab-kernel.html"
}
],
"paper_module_id": "paper_module.ring2_premise_precision_recall",
"reader_packet": {
"abstract": "After-the-fact premise retrieval metrics separate retrieval misses from proof failures without becoming theorem authority.",
"authority_ceiling": "Copied Ring2 retrieval records and public fixture/exported-bundle result records only; no Lean/Lake execution, theorem proof, benchmark performance, external model access, provider-context label flow, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.ring2_premise_precision_recall.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-ring2-premise-precision-recall-diagram-1",
"html_url": "docs/paper-module-ring2-premise-precision-recall.html#paper-module-ring2-premise-precision-recall-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Inputs[\"four input files runs, labels, rankings, policy\"] --> Build[\"_build_result orchestrates stages, sets status\"] Build --> Evaluate[\"_evaluate precision/recall + four-way label\"] Build --> RunMat[\"_validate_run_material copied-material provenance\"] Build --> Artifacts[\"_validate_source_artifacts digest match per file\"] Build --> Negatives[\"_negative_findings five typed refusals\"] Evaluate --> Status[\"status: PASS or blocked\"] RunMat --> Status Artifacts --> Status Negatives --> Status Status --> Result records[\"_write_receipts metrics only, no proof bodies\"]",
"source_format": "mermaid",
"source_hash": "sha256:b5025c0635ecf14d5ceb6d0db4e2686d6d661259dde6d97fe67c07af8e7b66fa",
"source_line_end": 78,
"source_line_start": 66
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 2,
"links": [
{
"id": "paper_module.formal_math_premise_retrieval",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Premise Retrieval",
"relation": "depends on",
"url": "paper-module-formal-math-premise-retrieval.html"
},
{
"id": "paper_module.proof_diagnostic_evidence_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof Diagnostic Evidence Spine",
"relation": "depends on",
"url": "paper-module-proof-diagnostic-evidence-spine.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "ring2_premise_retrieval_precision_recall_harness",
"kind": "organ",
"kind_label": "component",
"label": "Ring2 Premise Retrieval Precision Recall Harness",
"relation": "explains",
"url": "component-ring2-premise-retrieval-precision-recall-harness.html"
},
{
"id": "mechanism.ring2_premise_retrieval_precision_recall_harness.validates_public_premise_retrieval_attribution",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public premise retrieval attribution",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-ring2-premise-retrieval-precision-recall-harness-validates-public-premise-retrieval-attribution"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.ring2_premise_precision_recall",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 129,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/ring2_premise_precision_recall.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "TREC evaluation measures",
"section_id": "grounded-in-ir-metrics-and-premise-selection",
"url": "https://trec.nist.gov/pubs/trec20/appendices/measures.pdf"
},
{
"kind": "external_reference",
"label": "precision/recall metric API",
"section_id": "grounded-in-ir-metrics-and-premise-selection",
"url": "https://scikit-learn.org/stable/modules/generated/sklearn.metrics.precision_recall_fscore_support.html"
},
{
"kind": "external_reference",
"label": "Sledgehammer",
"section_id": "grounded-in-ir-metrics-and-premise-selection",
"url": "https://isabelle.in.tum.de/doc/sledgehammer.pdf"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "It replays copied premise-retrieval records after the fact and reports, problem by problem, how much of the needed support the search actually ranked. ring2premiseretrievalprecisionrecall_harness scores saved premise-retrieval runs after the fact. A proof search first looks for the supporting facts a proof will need, the premises. This component replays copied records of that search and reports, problem by problem, how much of the needed support the search actually ranked.",
"level": 4,
"order": 1,
"section_id": "premise-retrieval-precision-and-recall-replay",
"source_line_end": 6,
"source_line_start": 3,
"title": "Premise-retrieval precision and recall replay"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It keeps two failures apart, a missed premise versus a proof that had everything and still failed, by computing precision and recall and labelling each problem one of four outcomes. The point is to keep two failures apart. A proof can fail because the search never surfaced a premise it needed. A proof can also fail even though every needed premise was ranked. Those call for different fixes. The component computes precision and recall from the copied records, then labels each problem as one of four outcomes so a reader can tell which case they are looking at.",
"level": 4,
"order": 2,
"section_id": "it-keeps-two-failures-apart",
"source_line_end": 10,
"source_line_start": 7,
"title": "It keeps two failures apart"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It emits measurement result records, bounded evidence authority, and the labels may never flow back into the ranking or into a provider recipe. It emits result records, bounded evidence authority. The labels are measurement labels. They score a finished run and they are never allowed to flow back into the ranking or into a provider recipe.",
"level": 4,
"order": 3,
"section_id": "it-emits-measurement-result-records",
"source_line_end": 14,
"source_line_start": 11,
"title": "It emits measurement result records"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A failed search invites blaming the prover, but a single reported precision figure cannot say whether the fault was a missed premise or a proof that had everything and still failed. When a proof search fails it is easy to blame the prover and miss the simpler cause: the right facts were never put in front of it. A single reported precision figure hides this. It cannot say whether a low score came from a missed premise or from a proof that had everything it needed and still did not go through.",
"level": 4,
"order": 5,
"section_id": "blaming-the-prover-hides-a-missed-premise",
"source_line_end": 20,
"source_line_start": 17,
"title": "Blaming the prover hides a missed premise"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It recomputes the numbers by intersecting ranked ids with needed ids and reading the proof outcome, so full recall with a failing proof surfaces as a downstream fault. So the component recomputes the numbers rather than trusting a reported one. For each problem it intersects the ranked premise ids with the labelled needed-premise ids and reads the proof outcome alongside that overlap. Full recall with a passing proof is one outcome; full recall with a failing proof is a different one. That second case is the signal that the fault lies downstream of retrieval, and a loose metric that only averages scores would erase it.",
"level": 4,
"order": 6,
"section_id": "recomputed-overlap-read-against-proof-outcome",
"source_line_end": 24,
"source_line_start": 21,
"title": "Recomputed overlap read against proof outcome"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "run validates the fixture with five negative cases and runprecisionrecallbundle validates the export without them, both calling buildresult to set one PASS or blocked status. Two entrypoints drive the same core. run validates a fixture directory with the five negative cases included. runprecisionrecallbundle validates an exported bundle with the negative cases turned off. Both call buildresult, which runs the stages below in order and decides a single status of PASS or blocked.",
"level": 4,
"order": 8,
"section_id": "two-entrypoints-over-one-build-result",
"source_line_end": 39,
"source_line_start": 27,
"title": "Two entrypoints over one `_build_result`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluate truncates each ranking to topk, intersects ranked with needed ids for hits, and computes precision and recall against the policy's expected aggregates. The scoring core is evaluate. It indexes the after-the-fact labels by problemid, reads defaulttopk from the policy (4 if unset) or a per-ranking topk, and truncates each ranking to that cutoff. It intersects the truncated ranked ids with the labelled needed ids to get the hits, then computes precision = hits / topk and recall = hits / needed. Aggregate precision and recall come from summed hit, candidate, and needed counts, and are compared against the policy's expectedaggregate_metrics.",
"level": 4,
"order": 9,
"section_id": "precision-and-recall-at-a-top-k-cutoff",
"source_line_end": 43,
"source_line_start": 40,
"title": "Precision and recall at a top-k cutoff"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The four-way label is mechanical, driven by recall and the recorded proofoutcome:** The four-way label is mechanical, driven by recall and the recorded proof_outcome: retrieval_hit: recall reached 1.0 and the proof passed. prooffailuredespite_hit: recall reached 1.0 but the proof did not pass. Retrieval did its job; the fault is elsewhere. partialretrievalmiss: some needed premises were ranked, not all. retrieval_miss: none of the needed premises were ranked.",
"level": 4,
"order": 10,
"section_id": "the-four-way-label-is-mechanical-driven-by-recall-and-the-recorded",
"source_line_end": 53,
"source_line_start": 44,
"title": "The four-way label is mechanical, driven by recall and the recorded"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluate also requires every expected failure mode, an adversarial decoy problem, and the minimum problem count, turning any gap into a finding. evaluate also enforces the shape of the evaluation set, not just the happy path. It requires every mode named in the policy's expectedfailure_modes to appear, requires an adversarial decoy problem whose needed premise is absent or missed, and requires the policy's minimum problem count. A gap in any of these becomes a finding.",
"level": 4,
"order": 11,
"section_id": "enforcing-the-shape-of-the-evaluation-set",
"source_line_end": 57,
"source_line_start": 54,
"title": "Enforcing the shape of the evaluation set"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Before any score is trusted, validaterunmaterial and validatesourceartifacts check provenance and digests, and status is PASS only when every stage and the secret scan clear. Before any score is trusted, the provenance stages run. validaterunmaterial rejects copied material that is missing a sourceref, sourcesha256, targetrefs, validationrefs, or the expected copiednonsecretmacrobodywithprovenance status. validatesourceartifacts walks the four files in SOURCEREFS, hashes each copied artifact under sourceartifacts/, and accepts it only if the hash matches the recorded exact digest in SOURCEDIGESTS or a known path-rewrite digest. A missing file or a hash mismatch is a finding. build_result al\u2026",
"level": 4,
"order": 12,
"section_id": "provenance-and-digest-checks-gate-the-status",
"source_line_end": 61,
"source_line_start": 58,
"title": "Provenance and digest checks gate the status"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "writereceipts emits import ids, target refs, digest status, and metric counts, but writes no proof bodies, model-output data, or non-public paths. write_receipts then emits the records. It exposes import ids, target refs, digest status, aggregate metrics, and failure-mode counts. It does not write proof bodies, model-output data, or non-public paths.",
"level": 4,
"order": 13,
"section_id": "receipts-expose-metrics-never-proof-bodies",
"source_line_end": 79,
"source_line_start": 62,
"title": "Result records expose metrics, never proof bodies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 81,
"source_line_start": 80,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVECASES names five adversarial inputs that run must reject, each mapped to one refusal code: EXPECTEDNEGATIVECASES names five adversarial inputs that run must reject, each mapped to one refusal code: oraclelabelsinranking plants the known-correct premise ids inside a ranking. inspectforbiddenbodies catches the oracleneededpremiseids field and emits RING2RETRIEVALORACLELABELSINRANKING. proofbodyleakage carries a proof, provider, or private body field and emits RING2RETRIEVALPROOFBODYFORBIDDEN. testsplittuningattempt marks a run as tuned on the test labels and emits RING2RETRIEVALTESTSPLITTUNINGFORBIDDEN. metricoverclaim asserts proof, benchmark, provider, or launch-scope decisi\u2026",
"level": 4,
"order": 15,
"section_id": "expected-negative-cases-names-five-adversarial-inputs-that-run-must-reject",
"source_line_end": 92,
"source_line_start": 82,
"title": "`EXPECTED_NEGATIVE_CASES` names five adversarial inputs that `run` must reject"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "If any of the five is not observed, buildresult records it as missing and the run is blocked. If any of the five is not observed, buildresult records it as missing and the run is blocked.",
"level": 4,
"order": 16,
"section_id": "missing-negative-cases-block-the-run",
"source_line_end": 97,
"source_line_start": 93,
"title": "Missing negative cases block the run"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 99,
"source_line_start": 98,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The design draws on TREC and scikit-learn precision/recall evaluation and on hammer-style premise selection, where finding the right facts is a distinct step from replaying a proof. This component is grounded in information-retrieval evaluation. NIST's TREC evaluation measures give the older precision/recall frame for judging retrieval systems, and scikit-learn's precision/recall metric API is a common machine-learning interface for reporting those labels. On the theorem-proving side, premise selection sits next to hammer workflows such as Isabelle Sledgehammer, where finding the right facts is a distinct step from replaying a proof. This component keeps that distinction explicit and measur\u2026",
"level": 4,
"order": 18,
"section_id": "grounded-in-ir-metrics-and-premise-selection",
"source_line_end": 103,
"source_line_start": 100,
"title": "Grounded in IR metrics and premise selection"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate/, reproduce this page's proof boundary with temporary records:",
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 114,
"source_line_start": 104,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means the records scored to expectation, the four digests matched, and all five negative cases were rejected, without anything running Lean or a prover. A pass means the copied records scored to their expected metrics, the four source artifacts matched their digests, and all five negative cases were rejected. It does not mean anything ran Lean or a prover.",
"level": 4,
"order": 20,
"section_id": "what-a-pass-over-the-records-means",
"source_line_end": 118,
"source_line_start": 115,
"title": "What a pass over the records means"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 120,
"source_line_start": 119,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim is that copied public records can be scored, sorted into the four labels, and guarded by five refusals, all reproducible from the fixture. The strongest claim this component supports is that copied public premise-retrieval records can be scored for precision and recall, sorted into the four retrieval-versus-proof-failure labels, and guarded by the five typed refusals, all reproducible from the fixture. The scope_limit constant fixes the rest.",
"level": 4,
"order": 22,
"section_id": "the-strongest-supported-claim",
"source_line_end": 124,
"source_line_start": 121,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It runs no Lean, no providers, and no launch, and a retrievalhit label says only that needed ids appeared in the ranking, nothing about proof soundness. It does not run Lean or Lake, use external model services, emit proof bodies, tune retrieval on test answers, claim benchmark performance, prove that a theorem is true, or include launch operations. A retrievalhit label means the needed-premise ids appeared in the ranking under this fixture's policy, nothing about whether the downstream proof is sound. The reported numbers are bounded by the declared fixture and bundle; a different corpus, cutoff, or premise labelling has to be rerun. The proof boundary is the copied records and public resu\u2026",
"level": 4,
"order": 23,
"section_id": "what-the-labels-never-claim",
"source_line_end": 128,
"source_line_start": 125,
"title": "What the labels never claim"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/ring2_premise_precision_recall.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "After-the-fact premise retrieval metrics separate retrieval misses from proof failures without becoming theorem authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "ring2_premise_retrieval_precision_recall_harness.py",
"path": "src/microcosm_core/organs/ring2_premise_retrieval_precision_recall_harness.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for Ring2 precision/recall fixture validation, exported bundle validation, copied source artifact checks, negative cases, and scope limits.",
"symbols": [
"run",
"run_precision_recall_bundle",
"_build_result",
"_write_receipts",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/ring2_premise_retrieval_precision_recall_harness.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/ring2_premise_precision_recall.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/ring2_premise_precision_recall.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Ring-2 Premise Precision Recall validates copied Ring2 retrieval rankings against after-the-fact metric labels. It computes precision/recall classes, rejects oracle-label leakage and tuning shortcuts, and keeps proof bodies, model-output data, benchmark claims, and theorem-correctness claims outside the public result record boundary.",
"summary_line": "After-the-fact premise retrieval metrics separate retrieval misses from proof failures without becoming theorem authority.",
"title": "Ring-2 Premise Precision Recall"
},
{
"anchor": "paper-module-mathematical-strategy-atlas",
"authority_ceiling": "Copied public strategy metadata, public source tool bodies, and public fixture/exported-bundle result records only; no Lean/Lake execution, formal-result correctness, oracle-label visibility, external model access, benchmark performance, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
A scorer for formal-math proof-search hypotheses
\n
mathematical_strategy_atlas_hypothesis_scorer is a small component that records a formal-math proof search's first move before any proof runs.mathematical_strategy_atlas_hypothesis_scorer is a small component that records a formal-math proof search's first move before any proof runs. It reads public problem feature tags and a strategy atlas, scores each candidate opening strategy, and writes down which one it picked and why. It never proves anything.
\n
The output is a set of result records: the selected strategy id for each case, the integer score that chose it, the component breakdown that produced the score, and a list of cases where nothing matched. The output is a set of result records: the selected strategy id for each case, the integer score that chose it, the component breakdown that produced the score, and a list of cases where nothing matched. A cold reader can take the same three input files and re-derive every selection by hand.
\n
The shipped finite-sum example
\n
A worked example: a problem tagged iff_goal scores the iff_split strategy highest and selects it. A worked example: a problem tagged iff_goal scores the iff_split strategy highest and selects it. A problem whose only tag matches nothing scores no candidate above zero and resolves to a typed STRATEGY_SELECTION_MISS instead of a silent guess.
\n
Purpose
\n
A proof search has to start somewhere. Before a premise is retrieved or a tactic is run, an agent has already committed to a first move: a goal shape and a family of tactics it expects to use. A proof search has to start somewhere. Before a premise is retrieved or a tactic is run, an agent has already committed to a first move: a goal shape and a family of tactics it expects to use. That choice is usually implicit, buried inside a model call. This component pulls it into the open and answers one question: for a given problem shape, which strategy did the system pick first, and on what visible evidence?
\n
The constraint is what the answer is allowed to depend on. The scorer never sees the oracle's expected strategy, the ground-truth proof, or any provider output. The constraint is what the answer is allowed to depend on. The scorer never sees the oracle's expected strategy, the ground-truth proof, or any provider output. It works from public feature tags and a strategy atlas of trigger features, negative triggers, and retrieval-expansion terms. The selected strategy is a hypothesis recomputed from readable inputs, not a label copied from an answer key.
\n
A loose classifier fails here in two ways. It can bake the answer in by declaring the chosen strategy as a plain label, or it can score on shallow feature overlap that happens to line up with the known-good label. A loose classifier fails here in two ways. It can bake the answer in by declaring the chosen strategy as a plain label, or it can score on shallow feature overlap that happens to line up with the known-good label. This component rejects both. A declared selection must match the score the component recomputes, and a strategy chosen on overlap alone is a typed negative case rather than a pass.
\n
How it works
\n
The entrypoints are run for the fixture and run_strategy_bundle for the exported bundle. Both call _build_result, which runs the stages below in order and then hands the result to write_receipts. The entrypoints are run for the fixture and run_strategy_bundle for the exported bundle. Both call _build_result, which runs the stages below in order and then hands the result to write_receipts.
\n
_build_result loads three public inputs: strategy_atlas.json, problem_features.json, and hypothesis_cases.json._build_result loads three public inputs: strategy_atlas.json, problem_features.json, and hypothesis_cases.json. It scans every input path for non-public-state leakage first, so a blocking scan hit fails the run before scoring is trusted. Then it runs scoring, the copied-body floor check, and the source-artifact consistency check, and it derives a single status that is PASS only when all of them pass and no expected negative case is missing.
\n
validate_strategy_selection is the scoring core. For each case it calls _score_case, which calls _score_strategy once per candidate.validate_strategy_selection is the scoring core. For each case it calls _score_case, which calls _score_strategy once per candidate. The score is one integer from three terms. Each problem feature that matches a strategy's trigger_features adds four points (TRIGGER_FEATURE_WEIGHT). Each feature that matches its negative_triggers subtracts three (NEGATIVE_TRIGGER_PENALTY). Each retrieval-query term found in the strategy's expansion terms adds one point (RETRIEVAL_TERM_BONUS), capped at two (RETRIEVAL_BONUS_CAP). Plain feature overlap is recorded as a diagnostic count and kept out of the score.
\n
Selection is a deterministic sort. Only strategies with a positive score are eligible. Selection is a deterministic sort. Only strategies with a positive score are eligible. Among those, _score_case ranks by score highest first, breaks ties by the strategy's declared order and then its id, and takes the top row. If no candidate scores positive the case resolves to STRATEGY_SELECTION_MISS. Each case record carries the selected id, the score, the component breakdown, the ranked candidate scores, and the trigger, negative, and retrieval hits that produced them.
\n
The weights encode the design intent. Trigger matches are worth more than retrieval matches, so a strategy is chosen for the shape it claims to handle, not for how many search terms coincide. The weights encode the design intent. Trigger matches are worth more than retrieval matches, so a strategy is chosen for the shape it claims to handle, not for how many search terms coincide. A negative trigger can veto a strategy that looks apt. The retrieval cap stops a strategy from winning on keyword volume.
\n
The same recomputation enforces honesty. When a case declares its own selected_strategy_id, score, classifier, retrieval_bonus, or candidate_scores, _verify_declared_case_outcomes recomputes each from the evidence and reports a stale-declaration finding on any mismatch. The same recomputation enforces honesty. When a case declares its own selected_strategy_id, score, classifier, retrieval_bonus, or candidate_scores, _verify_declared_case_outcomes recomputes each from the evidence and reports a stale-declaration finding on any mismatch. A case that declares only a bare selected_strategy_id, with nothing to check it against, is rejected as label-only.
\n
Two further checks guard the copied inputs. validate_source_module_imports reads source_module_manifest.json and confirms each copied source body against its recorded SHA-256 digest, line count, byte count, and required anchors, and rejects any body that would land inside a generated result record. Two further checks guard the copied inputs. validate_source_module_imports reads source_module_manifest.json and confirms each copied source body against its recorded SHA-256 digest, line count, byte count, and required anchors, and rejects any body that would land inside a generated result record. validate_source_artifact_consistency opens the three copied runtime artifacts (strategy_cards.json, strategy_hypothesis_set.json, prover_skill_atlas.json) and checks them for oracle-label and proof-body leakage. write_receipts then emits four metadata-only result records: mathematical_strategy_atlas_result.json, mathematical_strategy_atlas_board.json, mathematical_strategy_atlas_validation_receipt.json, and a fixture-sign-off record. Each carries refs, counts, hits, and verdicts, never body text.
\n
Function
Role
run, run_strategy_bundle
Entrypoints for the fixture and exported bundle
_build_result
Runs the stages in order and derives status
validate_strategy_selection
Scores every case and checks the negative cases
_score_strategy, _score_case
Weighted per-strategy score and per-case selection
validate_source_module_imports
Digest, count, and anchor check for copied bodies
validate_source_artifact_consistency
Leakage check on the three copied runtime artifacts
write_receipts
Emits the four metadata-only result records
\nDiagram of the mechanism (8 steps).Diagram source & refs
flowchart TD inputs["strategy_atlas.json problem_features.json hypothesis_cases.json"] scan["scan inputs for non-public-state leakage"] score["_score_strategy trigger x4 - negative x3 + retrieval (cap 2)"] select{"any positive score?"} selected["selected_strategy_id + score breakdown"] miss["STRATEGY_SELECTION_MISS"] verify["recompute vs declared fields"] result records["four metadata-only result records"] inputs --> scan scan --> score score --> select select -- yes --> selected select -- no --> miss selected --> verify miss --> verify verify --> result records
\n
Negative cases
\n
The fixture supplies eight named negative cases in EXPECTED_NEGATIVE_CASES, and the run fails if any expected case is not observed. The fixture supplies eight named negative cases in EXPECTED_NEGATIVE_CASES, and the run fails if any expected case is not observed. unknown_strategy_id rejects a strategy id outside the public enum. proof_body_with_strategy rejects any proof, provider, or source body field on a case. oracle_strategy_label_leakage rejects a fixture that exposes an oracle strategy label. post_oracle_strategy_selection rejects a case marked as selected after oracle labels are visible. release_overclaim rejects a payload that tries to set release_authorized or any sibling authority flag. superficial_overlap_only_scoring rejects legacy unweighted overlap as the decision rule. missing_rich_strategy_card_fields rejects a strategy card that lacks trigger_features, negative_triggers, retrieval_expansion_terms, or proof_plan_template. retrieval_bonus_ceiling_overclaim rejects a claimed retrieval cap above two.
\n
Prior Art Grounding
\n
The atlas follows the formal-methods practice of separating problem-shape classification from proof execution. Lean's tactic model, described in Theorem Proving in Lean 4, is the immediate precedent: proof work is arrange around tactics chosen for a goal shape while the kernel checks the final proof state. The atlas follows the formal-methods practice of separating problem-shape classification from proof execution. Lean's tactic model, described in Theorem Proving in Lean 4, is the immediate precedent: proof work is arrange around tactics chosen for a goal shape while the kernel checks the final proof state. The mathlib overview motivates explicit retrieval terms and domain tags, because a large formal library is navigated by topic and reusable theorem families. The pattern is also adjacent to hammer-style method selection such as Isabelle Sledgehammer, where a front-end tool searches for useful facts before replay. This component keeps the pattern pre-oracle and metadata-only: it records why a first strategy was selected, not whether the proof can be completed.
A green run means the strategy scoring, the copied-body digests, the source-artifact consistency, and the eight negative cases all checked out. A green run means the strategy scoring, the copied-body digests, the source-artifact consistency, and the eight negative cases all checked out. It does not mean a proof ran.
\n
Scope boundary
\n
Scope limit
\n
The bounded strategy-selection claim
\n
The strongest claim the evidence supports is narrow: public pre-oracle strategy hypotheses, the retrieval-lens metadata that feeds them, the copied public source bodies, their source-artifact digests, and the eight negative cases are all inspectable and reproducible from the three input files. The strongest claim the evidence supports is narrow: public pre-oracle strategy hypotheses, the retrieval-lens metadata that feeds them, the copied public source bodies, their source-artifact digests, and the eight negative cases are all inspectable and reproducible from the three input files. The proof boundary is exactly that metadata and those metadata-only result records.
\n
The component does not run Lean or Lake, prove formal-result correctness, reveal oracle strategy labels, expose proof bodies, use external model services, tune on test answers, certify benchmark performance, or authorize public sharing or launch. The component does not run Lean or Lake, prove formal-result correctness, reveal oracle strategy labels, expose proof bodies, use external model services, tune on test answers, certify benchmark performance, or authorize public sharing or launch. The copied runtime artifacts are public strategy traces, not oracle labels or model-output data. That scope limit is set in AUTHORITY_CEILING, where every proof, provider, and launch flag is false.
\n",
"components": [
{
"anchor": "component-mathematical_strategy_atlas_hypothesis_scorer",
"component_id": "mathematical_strategy_atlas_hypothesis_scorer",
"public_label": "Mathematical Strategy Atlas Hypothesis Scorer"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 6
},
"link_count": 18,
"links": [
{
"id": "mathematical_strategy_atlas_hypothesis_scorer",
"kind": "organ",
"kind_label": "component",
"label": "Mathematical Strategy Atlas Hypothesis Scorer",
"relation": "explains",
"url": "component-mathematical-strategy-atlas-hypothesis-scorer.html"
},
{
"id": "mechanism.mathematical_strategy_atlas_hypothesis_scorer.validates_public_strategy_hypothesis_projection",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public strategy hypothesis projection",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-mathematical-strategy-atlas-hypothesis-scorer-validates-public-strategy-hypothesis-projection"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.lean_std_premise_index",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Lean/Std Premise Index",
"relation": "depends on",
"url": "paper-module-lean-std-premise-index.html"
},
{
"id": "paper_module.formal_math_premise_retrieval",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Premise Retrieval",
"relation": "depends on",
"url": "paper-module-formal-math-premise-retrieval.html"
},
{
"id": "paper_module.ring2_premise_precision_recall",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Ring-2 Premise Precision Recall",
"relation": "depends on",
"url": "paper-module-ring2-premise-precision-recall.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.mathematical_strategy_atlas",
"reader_packet": {
"abstract": "Pre-oracle strategy hypotheses make the first proof-search move inspectable without claiming proof or provider authority.",
"authority_ceiling": "Copied public strategy metadata, public source tool bodies, and public fixture/exported-bundle result records only; no Lean/Lake execution, formal-result correctness, oracle-label visibility, external model access, benchmark performance, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.mathematical_strategy_atlas.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-mathematical-strategy-atlas-diagram-1",
"html_url": "docs/paper-module-mathematical-strategy-atlas.html#paper-module-mathematical-strategy-atlas-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD inputs[\"strategy_atlas.json problem_features.json hypothesis_cases.json\"] scan[\"scan inputs for non-public-state leakage\"] score[\"_score_strategy trigger x4 - negative x3 + retrieval (cap 2)\"] select{\"any positive score?\"} selected[\"selected_strategy_id + score breakdown\"] miss[\"STRATEGY_SELECTION_MISS\"] verify[\"recompute vs declared fields\"] result records[\"four metadata-only result records\"] inputs --> scan scan --> score score --> select select -- yes --> selected select -- no --> miss selected --> verify miss --> verify verify --> result records",
"source_format": "mermaid",
"source_hash": "sha256:524cac1e5910a97f6c2d472878fb447833719e6b11887b9262d580ab123ef46c",
"source_line_end": 165,
"source_line_start": 147
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.lean_std_premise_index",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Lean/Std Premise Index",
"relation": "depends on",
"url": "paper-module-lean-std-premise-index.html"
},
{
"id": "paper_module.formal_math_premise_retrieval",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Premise Retrieval",
"relation": "depends on",
"url": "paper-module-formal-math-premise-retrieval.html"
},
{
"id": "paper_module.ring2_premise_precision_recall",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Ring-2 Premise Precision Recall",
"relation": "depends on",
"url": "paper-module-ring2-premise-precision-recall.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "mathematical_strategy_atlas_hypothesis_scorer",
"kind": "organ",
"kind_label": "component",
"label": "Mathematical Strategy Atlas Hypothesis Scorer",
"relation": "explains",
"url": "component-mathematical-strategy-atlas-hypothesis-scorer.html"
},
{
"id": "mechanism.mathematical_strategy_atlas_hypothesis_scorer.validates_public_strategy_hypothesis_projection",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public strategy hypothesis projection",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-mathematical-strategy-atlas-hypothesis-scorer-validates-public-strategy-hypothesis-projection"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.mathematical_strategy_atlas",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 250,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/mathematical_strategy_atlas.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Theorem Proving in Lean 4",
"section_id": "the-atlas-follows-the-formal-methods-practice-of-separating-problem-shape-classification-from-proof-execution",
"url": "https://lean-lang.org/theorem_proving_in_lean4/Tactics/"
},
{
"kind": "external_reference",
"label": "mathlib overview",
"section_id": "the-atlas-follows-the-formal-methods-practice-of-separating-problem-shape-classification-from-proof-execution",
"url": "https://leanprover-community.github.io/mathlib-overview.html"
},
{
"kind": "external_reference",
"label": "Sledgehammer",
"section_id": "the-atlas-follows-the-formal-methods-practice-of-separating-problem-shape-classification-from-proof-execution",
"url": "https://isabelle.in.tum.de/doc/sledgehammer.pdf"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "mathematicalstrategyatlashypothesisscorer is a small component that records a formal-math proof search's first move before any proof runs. mathematicalstrategyatlashypothesisscorer is a small component that records a formal-math proof search's first move before any proof runs. It reads public problem feature tags and a strategy atlas, scores each candidate opening strategy, and writes down which one it picked and why. It never proves anything.",
"level": 4,
"order": 1,
"section_id": "a-scorer-for-formal-math-proof-search-hypotheses",
"source_line_end": 11,
"source_line_start": 3,
"title": "A scorer for formal-math proof-search hypotheses"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The output is a set of result records: the selected strategy id for each case, the integer score that chose it, the component breakdown that produced the score, and a list of cases where nothing matched. The output is a set of result records: the selected strategy id for each case, the integer score that chose it, the component breakdown that produced the score, and a list of cases where nothing matched. A cold reader can take the same three input files and re-derive every selection by hand.",
"level": 4,
"order": 2,
"section_id": "the-output-is-a-set-of-result-records",
"source_line_end": 19,
"source_line_start": 12,
"title": "The output is a set of result records"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A worked example: a problem tagged iffgoal scores the iffsplit strategy highest and selects it. A worked example: a problem tagged iffgoal scores the iffsplit strategy highest and selects it. A problem whose only tag matches nothing scores no candidate above zero and resolves to a typed STRATEGYSELECTIONMISS instead of a silent guess.",
"level": 4,
"order": 3,
"section_id": "the-shipped-finite-sum-example",
"source_line_end": 27,
"source_line_start": 20,
"title": "The shipped finite-sum example"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 29,
"source_line_start": 28,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A proof search has to start somewhere. Before a premise is retrieved or a tactic is run, an agent has already committed to a first move: a goal shape and a family of tactics it expects to use. A proof search has to start somewhere. Before a premise is retrieved or a tactic is run, an agent has already committed to a first move: a goal shape and a family of tactics it expects to use. That choice is usually implicit, buried inside a model call. This component pulls it into the open and answers one question: for a given problem shape, which strategy did the system pick first, and on what visible evidence?",
"level": 4,
"order": 5,
"section_id": "a-proof-search-has-to-start-somewhere",
"source_line_end": 39,
"source_line_start": 30,
"title": "A proof search has to start somewhere"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The constraint is what the answer is allowed to depend on. The scorer never sees the oracle's expected strategy, the ground-truth proof, or any provider output. The constraint is what the answer is allowed to depend on. The scorer never sees the oracle's expected strategy, the ground-truth proof, or any provider output. It works from public feature tags and a strategy atlas of trigger features, negative triggers, and retrieval-expansion terms. The selected strategy is a hypothesis recomputed from readable inputs, not a label copied from an answer key.",
"level": 4,
"order": 6,
"section_id": "the-constraint-is-what-the-answer-is-allowed-to-depend-on",
"source_line_end": 49,
"source_line_start": 40,
"title": "The constraint is what the answer is allowed to depend on"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A loose classifier fails here in two ways. It can bake the answer in by declaring the chosen strategy as a plain label, or it can score on shallow feature overlap that happens to line up with the known-good label. A loose classifier fails here in two ways. It can bake the answer in by declaring the chosen strategy as a plain label, or it can score on shallow feature overlap that happens to line up with the known-good label. This component rejects both. A declared selection must match the score the component recomputes, and a strategy chosen on overlap alone is a typed negative case rather than a pass.",
"level": 4,
"order": 7,
"section_id": "a-loose-classifier-fails-here-in-two-ways",
"source_line_end": 59,
"source_line_start": 50,
"title": "A loose classifier fails here in two ways"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 61,
"source_line_start": 60,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The entrypoints are run for the fixture and runstrategybundle for the exported bundle. Both call buildresult, which runs the stages below in order and then hands the result to writereceipts.** The entrypoints are run for the fixture and runstrategybundle for the exported bundle. Both call buildresult, which runs the stages below in order and then hands the result to write_receipts.",
"level": 4,
"order": 9,
"section_id": "the-entrypoints-are-run-for-the-fixture-and-run-strategy-bundle-for-the-exported-bundle",
"source_line_end": 68,
"source_line_start": 62,
"title": "The entrypoints are `run` for the fixture and `run_strategy_bundle` for the exported bundle"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult loads three public inputs: strategyatlas.json, problemfeatures.json, and hypothesiscases.json. buildresult loads three public inputs: strategy_atlas.json, problemfeatures.json, and hypothesiscases.json. It scans every input path for non-public-state leakage first, so a blocking scan hit fails the run before scoring is trusted. Then it runs scoring, the copied-body floor check, and the source-artifact consistency check, and it derives a single status that is PASS only when all of them pass and no expected negative case is missing.",
"level": 4,
"order": 10,
"section_id": "build-result-loads-three-public-inputs",
"source_line_end": 78,
"source_line_start": 69,
"title": "`_build_result` loads three public inputs"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatestrategyselection is the scoring core. For each case it calls scorecase, which calls scorestrategy once per candidate. validatestrategyselection is the scoring core. For each case it calls scorecase, which calls scorestrategy once per candidate. The score is one integer from three terms. Each problem feature that matches a strategy's triggerfeatures adds four points (TRIGGERFEATURE_WEIGHT). Each feature that matches its negativetriggers subtracts three (NEGATIVETRIGGER_PENALTY). Each retrieval-query term found in the strategy's expansion terms adds one point (RETRIEVALTERMBONUS), capped at two (RETRIEVALBONUSCAP). Plain feature overlap is recorded as a diagnostic count and kept out\u2026",
"level": 4,
"order": 11,
"section_id": "validate-strategy-selection-is-the-scoring-core",
"source_line_end": 90,
"source_line_start": 79,
"title": "`validate_strategy_selection` is the scoring core"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Selection is a deterministic sort. Only strategies with a positive score are eligible. Selection is a deterministic sort. Only strategies with a positive score are eligible. Among those, scorecase ranks by score highest first, breaks ties by the strategy's declared order and then its id, and takes the top row. If no candidate scores positive the case resolves to STRATEGYSELECTIONMISS. Each case record carries the selected id, the score, the component breakdown, the ranked candidate scores, and the trigger, negative, and retrieval hits that produced them.",
"level": 4,
"order": 12,
"section_id": "selection-is-a-deterministic-sort",
"source_line_end": 101,
"source_line_start": 91,
"title": "Selection is a deterministic sort"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The weights encode the design intent. Trigger matches are worth more than retrieval matches, so a strategy is chosen for the shape it claims to handle, not for how many search terms coincide. The weights encode the design intent. Trigger matches are worth more than retrieval matches, so a strategy is chosen for the shape it claims to handle, not for how many search terms coincide. A negative trigger can veto a strategy that looks apt. The retrieval cap stops a strategy from winning on keyword volume.",
"level": 4,
"order": 13,
"section_id": "the-weights-encode-the-design-intent",
"source_line_end": 110,
"source_line_start": 102,
"title": "The weights encode the design intent"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The same recomputation enforces honesty. When a case declares its own selectedstrategyid, score, classifier, retrievalbonus, or candidatescores, verifydeclaredcaseoutcomes recomputes each from the evidence and reports a stale-declaration finding on any mismatch. The same recomputation enforces honesty. When a case declares its own selectedstrategyid, score, classifier, retrieval_bonus, or candidatescores, verifydeclaredcase_outcomes recomputes each from the evidence and reports a stale-declaration finding on any mismatch. A case that declares only a bare selectedstrategyid, with nothing to check it against, is rejected as label-only.",
"level": 4,
"order": 14,
"section_id": "the-same-recomputation-enforces-honesty",
"source_line_end": 120,
"source_line_start": 111,
"title": "The same recomputation enforces honesty"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "Two further checks guard the copied inputs. validatesourcemoduleimports reads sourcemodulemanifest.json and confirms each copied source body against its recorded SHA-256 digest, line count, byte count, and required anchors, and rejects any body that would land inside a generated result record.** Two further checks guard the copied inputs. validatesourcemodule_imports reads sourcemodulemanifest.json and confirms each copied source body against its recorded SHA-256 digest, line count, byte count, and required anchors, and rejects any body that would land inside a generated result record. validatesourceartifact_consistency opens the three copied runtime artifacts (strategycards.json, strategyh\u2026",
"level": 4,
"order": 15,
"section_id": "two-further-checks-guard-the-copied-inputs",
"source_line_end": 166,
"source_line_start": 121,
"title": "Two further checks guard the copied inputs"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "negative-cases",
"source_line_end": 168,
"source_line_start": 167,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The fixture supplies eight named negative cases in EXPECTEDNEGATIVECASES, and the run fails if any expected case is not observed. The fixture supplies eight named negative cases in EXPECTEDNEGATIVECASES, and the run fails if any expected case is not observed. unknownstrategyid rejects a strategy id outside the public enum. proofbodywith_strategy rejects any proof, provider, or source body field on a case. oraclestrategylabel_leakage rejects a fixture that exposes an oracle strategy label. postoraclestrategy_selection rejects a case marked as selected after oracle labels are visible. release_overclaim rejects a payload that tries to set release_authorized or any sibling authority flag. super\u2026",
"level": 4,
"order": 17,
"section_id": "the-fixture-supplies-eight-named-negative-cases-in-expected-negative-cases",
"source_line_end": 185,
"source_line_start": 169,
"title": "The fixture supplies eight named negative cases in `EXPECTED_NEGATIVE_CASES`"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "prior-art-grounding",
"source_line_end": 187,
"source_line_start": 186,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The atlas follows the formal-methods practice of separating problem-shape classification from proof execution. Lean's tactic model, described in Theorem Proving in Lean 4, is the immediate precedent: proof work is arrange around tactics chosen for a goal shape while the kernel checks the final proof state. The atlas follows the formal-methods practice of separating problem-shape classification from proof execution. Lean's tactic model, described in Theorem Proving in Lean 4, is the immediate precedent: proof work is arrange around tactics chosen for a goal shape while the kernel checks the final proof state. The mathlib overview motivates explicit retrieval terms and domain tags, because a\u2026",
"level": 4,
"order": 19,
"section_id": "the-atlas-follows-the-formal-methods-practice-of-separating-problem-shape-classification-from-proof-execution",
"source_line_end": 204,
"source_line_start": 188,
"title": "The atlas follows the formal-methods practice of separating problem-shape classification from proof execution"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from microcosm-substrate:",
"level": 2,
"order": 20,
"section_id": "validation-receipt-path",
"source_line_end": 221,
"source_line_start": 205,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A green run means the strategy scoring, the copied-body digests, the source-artifact consistency, and the eight negative cases all checked out. A green run means the strategy scoring, the copied-body digests, the source-artifact consistency, and the eight negative cases all checked out. It does not mean a proof ran.",
"level": 4,
"order": 21,
"section_id": "what-a-green-strategy-run-establishes",
"source_line_end": 228,
"source_line_start": 222,
"title": "What a green strategy run establishes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 230,
"source_line_start": 229,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim the evidence supports is narrow: public pre-oracle strategy hypotheses, the retrieval-lens metadata that feeds them, the copied public source bodies, their source-artifact digests, and the eight negative cases are all inspectable and reproducible from the three input files. The strongest claim the evidence supports is narrow: public pre-oracle strategy hypotheses, the retrieval-lens metadata that feeds them, the copied public source bodies, their source-artifact digests, and the eight negative cases are all inspectable and reproducible from the three input files. The proof boundary is exactly that metadata and those metadata-only result records.",
"level": 4,
"order": 23,
"section_id": "the-bounded-strategy-selection-claim",
"source_line_end": 239,
"source_line_start": 231,
"title": "The bounded strategy-selection claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component does not run Lean or Lake, prove formal-result correctness, reveal oracle strategy labels, expose proof bodies, use external model services, tune on test answers, certify benchmark performance, or authorize public sharing or launch. The component does not run Lean or Lake, prove formal-result correctness, reveal oracle strategy labels, expose proof bodies, use external model services, tune on test answers, certify benchmark performance, or authorize public sharing or launch. The copied runtime artifacts are public strategy traces, not oracle labels or model-output data. That scope limit is set in scope_limit, where every proof, provider, and launch flag is false.",
"level": 4,
"order": 24,
"section_id": "the-component-does-not-run-lean-or-lake",
"source_line_end": 249,
"source_line_start": 240,
"title": "The component does not run Lean or Lake"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/mathematical_strategy_atlas.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Pre-oracle strategy hypotheses make the first proof-search move inspectable without claiming proof or provider authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "mathematical_strategy_atlas_hypothesis_scorer.py",
"path": "src/microcosm_core/organs/mathematical_strategy_atlas_hypothesis_scorer.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for strategy-atlas fixture validation, exported bundle validation, copied body-floor checks, negative cases, and scope limits.",
"symbols": [
"run",
"run_strategy_bundle",
"validate_strategy_selection",
"validate_source_module_imports",
"validate_source_artifact_consistency",
"write_receipts",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/mathematical_strategy_atlas_hypothesis_scorer.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/mathematical_strategy_atlas.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/mathematical_strategy_atlas.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Mathematical Strategy Atlas scores public problem features into explicit pre-oracle strategy hypotheses before retrieval or proof execution. It validates strategy ids, copied public source tool bodies, retrieval-term effects, oracle-label exclusion, negative cases, and metadata-only result records while keeping proof and provider authority out of scope.",
"summary_line": "Pre-oracle strategy hypotheses make the first proof-search move inspectable without claiming proof or provider authority.",
"title": "Mathematical Strategy Atlas"
},
{
"anchor": "paper-module-verifier-lab-execution-spine",
"authority_ceiling": "Bounded public fixture and exported-bundle execution result records only; no general formal-result correctness, benchmark solve-rate claim, external model access, oracle-to-proof authority, proof body export, source-file changes, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
It runs bounded Lean proof-step candidates in a throwaway copy of a public project and records each checker return code while keeping proof text and payloads out.verifier_lab_execution_spine is the component that actually runs bounded Lean proof-step attempts and writes down what the checker said. It copies a small public Lean project into a throwaway workspace, builds it once, writes one .lean file per candidate, runs the Lean checker on each, and records the return code. It keeps the proof text, the tool output, and every provider or oracle payload out of the result. A reader gets the pass or fail facts and the safety counts, and nothing that would leak.
\n
Narrower than the kernel, the spine executes each transition, turns a zero exit into accepted, and rejects contaminated candidates before the checker ever runs. The component is narrower than verifier_lab_kernel. The kernel reasons about the shape of a transition; this spine executes it. A zero exit code from the Lean checker becomes accepted. Anything else becomes a residual with a failure class. Candidates that carry a forbidden field, or that mark an oracle or provider as visible, are rejected before the checker is ever invoked, so a contaminated run cannot be dressed up as a clean one.
\n
It emits four records, full result, board summary, validation, and fixture sign-off, each stating the same scope limit and carrying counts without bodies. It emits four result records: a full result, a board summary, a validation record, and a fixture-sign-off record. All four state the same scope limit, carry the return codes and counts, and omit the bodies.
\n
Purpose
\n
Keeping hidden help out of the record
\n
Proof systems can report success while hiding oracle or provider help, so this component answers one question: did a bounded Lean candidate pass the checker unaided? Automated proof systems can blur how a result was reached. A model can be handed the answer by an oracle, or shown the proof by a provider, and still report the result as if it had found the proof alone. A loose result record hides that distinction: it prints a success without saying what help was in the room. This component exists to keep that blur out of the record. It answers one question. Did a bounded Lean candidate pass the checker, with no help the record is hiding?
\n
Authority separation holds counters at zero
\n
Authority separation keeps verified candidates, oracle references, and provider hypotheses as distinct classes, pinning the forward-success and export counters at zero by construction. The discipline is authority separation. A verified candidate is one class. An oracle reference and a provider hypothesis are separate classes that are never counted as forward success. The counters oracle_forward_success_increment_count and provider_results_counted are held at zero by construction, alongside proof_body_export_count and source_mutation_count. Real execution and a clean record are not in tension: the leak checks run first, so the Lean run that follows is already free of the fields that would compromise it.
\n
How it works
\n
Two entry points over _build_result
\n
run executes Lean for real on the fixture while run_execution_bundle validates an exported group offline, and both flow through _build_result then write_receipts. Two entry points drive the component. run handles the first-wave fixture and executes Lean for real. run_execution_bundle validates an exported group without a Lean toolchain, so a third party can inspect a shipped bundle offline. Both call _build_result, then write_receipts.
\n
_build_result scans, copies, and builds once
\n
_build_result loads the packet, secret-scans every input, and builds the copied Lean project once, executing candidates only when that build returns zero._build_result is the spine. It loads the execution packet, scans every input path for forbidden content with the shared secret-exclusion scan, and reads the tool versions. In fixture mode it opens a temporary directory, copies the Lean project into it, and calls _build_lake_project, which runs lake build MicrocosmProofWitness once with a sixty-second limit. Only if that build returns zero does it execute the transition candidates.
\n
Per-candidate contract check in _execute_transition
\n
_execute_transition first runs the contract check for forbidden keys, visible oracle or provider, and unknown action class, rejecting the candidate or else running lake env lean and recording the return code. Each candidate goes through _execute_transition. That function first calls _validate_transition_contract. The contract check walks the candidate for any key in FORBIDDEN_TRANSITION_KEYS (proof_body, raw_tactic_script, oracle_template, provider_output_body, and similar) using _walk_forbidden_keys, checks whether oracle_visible or provider_visible is set, and checks that the action_class is one of the nine names in ALLOWED_ACTION_CLASSES. If any check fires, the candidate becomes a CONTRACT_REJECTED record with the error codes attached and no Lean run. If the contract is clean, _execute_transition writes a .lean source built by _lean_source_for_transition, runs lake env lean on it, and sets accepted from the return code. The record carries the return code, the action class, and the failure class. It never carries the proof text or the stdout and stderr bodies.
\n
CP2 translation and evolve mutation lanes
\n
Two smaller lanes also run: CP2 translation emits typed next-action candidates from a fixed vocabulary, and evolve mutation touches only bounded policy artifacts, each rejecting out-of-scope requests. Two smaller lanes run on the same packet. CP2 translation reads residual cases and emits typed next-action candidates drawn from a fixed vocabulary; a request that smuggles a proof body is rejected. Evolve mutation adjusts only the bounded policy artifacts named in ALLOWED_EVOLVE_ARTIFACTS; a mutation aimed at anything else is rejected.
\n
Digest checks on the exported group
\n
validate_source_module_imports recomputes each copied file's SHA-256 against the manifest and blocks the group on any digest mismatch, missing anchor, or out-of-prefix source.validate_source_module_imports handles the exported group. It reads source_module_manifest.json, confirms the declared module count matches the rows, and for each row recomputes the SHA-256 digest of the copied file and compares it to the manifest. A digest mismatch, a missing anchor, or a source reference outside the allowed prefixes blocks the group. The copied bodies live in the bundle, never in the record.
\n
write_receipts pins the zero-authority facts
\n
write_receipts composes the four records with every authority flag false, and the status passes only when the scan is clean, the build returned zero, and the expected transition counts are met.write_receipts composes the four records and pins the safety facts: receipts_include_proof_bodies false, provider_calls_authorized false, source_mutation_authorized false, release_authorized false. The overall status in _build_result passes only when the secret scan is clean, Lean and Lake are available, the project build returned zero, no expected negative case is missing, and the run produced at least four transitions, two accepted, one residual, one CP2 effect, and one accepted mutation.
\n
Function
Role
run
First-wave fixture entry; runs Lean for real, then writes records
run_execution_bundle
Exported-group entry; validates a shipped bundle without a toolchain
_build_result
Loads the packet, scans inputs, builds, executes, gates the status
_build_lake_project
Runs lake build MicrocosmProofWitness once
_execute_transition
Contract-checks, then runs lake env lean on one candidate
Recomputes copied-file digests for the exported group
write_receipts
Writes the four records and pins the zero-authority facts
\nDiagram of the mechanism (9 steps).Diagram source & refs
flowchart TD Packet["Execution packet transition candidates, CP2 requests, evolve mutations, oracle/provider refs"] Scan["Secret-exclusion scan over every input path"] Gate["Contract check forbidden field? oracle/provider visible? action class out of vocabulary?"] Rejected["CONTRACT_REJECTED no Lean run"] Build["lake build MicrocosmProofWitness once, in a temporary copy"] Run["Write .lean, run lake env lean return code 0 = accepted"] Verified["Accepted transition"] Residual["Residual non-zero return code"] Records["Four result records return codes and counts kept, bodies and proof text omitted"] Packet --> Scan Scan --> Gate Gate -->|leak found| Rejected Gate -->|clean| Build Build -->|return 0| Run Run -->|exit 0| Verified Run -->|non-zero| Residual Rejected --> Records Verified --> Records Residual --> Records
\n
Negative cases
\n
Four negative inputs the run must observe
\n
The fixture ships four negative inputs listed in EXPECTED_NEGATIVE_CASES, each expected to raise one error code, and a run is blocked unless all four are observed. The fixture ships four negative inputs, each expected to raise one error code, declared in EXPECTED_NEGATIVE_CASES. A run that fails to observe all four is blocked.
\n
transition_leaks_candidate_body must raise VERIFIER_LAB_EXECUTION_TRANSITION_FIELD_FORBIDDEN: a candidate carrying a proof body or raw tactic script.
provider_oracle_visible_transition must raise VERIFIER_LAB_EXECUTION_PROVIDER_OR_ORACLE_VISIBLE: a candidate that exposes an oracle structured source record or provider text to forward execution.
cp2_candidate_contains_proof_body must raise VERIFIER_LAB_EXECUTION_CP2_PROOF_BODY_FORBIDDEN: a translation request that smuggles a proof body into the typed-action lane.
evolve_mutates_unbounded_source must raise VERIFIER_LAB_EXECUTION_EVOLVE_SCOPE_FORBIDDEN: a mutation aimed outside the bounded policy artifacts.
\n
Prior Art Grounding
\n
Small-kernel trust and artifact evaluation
\n
Execution follows the small-kernel proof-assistant tradition where trust rests on the checker not the tactic, and the record shape follows artifact-evaluation practice. The Lean and Lake execution inherits from the small-kernel proof-assistant tradition, the Lean theorem prover and LCF-style systems such as HOL Light, where trust rests on a small checker rather than the tactic that produced the term. The record shape follows artifact-evaluation practice: record the command identity, the tool facts, and the pass or fail counts separately from the claim they support, and keep the underlying artifact out of the summary.
\n
Validation Result record Path
\n
Run from microcosm-substrate:
\n
What a green run proves
\n
A green run counts as bounded execution-spine evidence only and never as general proof certification. A green run proves only bounded execution-spine evidence: command intent, tool facts, return-code pass or fail per candidate, the four observed negative cases, and the zero-held authority counters.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
A bounded set of public Lean candidates ran in a throwaway copy of a public project, and each checker return code was recorded without exposing proof text. The strongest claim the evidence supports is narrow. A bounded set of public Lean candidates was run in a throwaway copy of a public project, and the checker's return code for each was recorded without exposing the proof text. That is the whole claim.
\n
What it refuses past that line
\n
Past that it certifies no general proofs, carries no Mathlib authority, exports no bodies, calls no provider, mutates no source, and excludes launch. It refuses everything past that. It does not certify proofs in general, does not carry Mathlib-dependent proof authority, does not export proof bodies, does not certify a benchmark solve rate, does not call any provider, does not change source files, and excludes launch or hosted deployment. An oracle match is never forward success. Provider text is never a proof. The diagram and Atlas views are navigation aids and do not promote a tool run into proof certification.
\n",
"components": [
{
"anchor": "component-verifier_lab_execution_spine",
"component_id": "verifier_lab_execution_spine",
"public_label": "Verifier Lab Execution Spine"
}
],
"doctrine_feed": {
"counts": {
"axiom": 5,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 7
},
"link_count": 18,
"links": [
{
"id": "verifier_lab_execution_spine",
"kind": "organ",
"kind_label": "component",
"label": "Verifier Lab Execution Spine",
"relation": "explains",
"url": "component-verifier-lab-execution-spine.html"
},
{
"id": "mechanism.verifier_lab_execution_spine.validates_public_verifier_transition_witness",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public verifier transition witness",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-verifier-lab-execution-spine-validates-public-verifier-transition-witness"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.verifier_lab_kernel",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Kernel",
"relation": "depends on",
"url": "paper-module-verifier-lab-kernel.html"
},
{
"id": "paper_module.formal_math_lean_proof_witness",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Lean Proof Witness",
"relation": "depends on",
"url": "paper-module-formal-math-lean-proof-witness.html"
},
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depends on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.certificate_kernel_execution_lab",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Certificate Kernel Execution Lab",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[7:paper_module.certificate_kernel_execution_lab].depends_on",
"url": "paper-module-certificate-kernel-execution-lab.html"
},
{
"id": "paper_module.engine_room_lean_proof_search_lab",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Lean Proof Search Lab",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[87:paper_module.engine_room_lean_proof_search_lab].depends_on",
"url": "paper-module-engine-room-lean-proof-search-lab.html"
}
],
"paper_module_id": "paper_module.verifier_lab_execution_spine",
"reader_packet": {
"abstract": "Bounded public verifier execution result records witness command execution without upgrading output into theorem authority.",
"authority_ceiling": "Bounded public fixture and exported-bundle execution result records only; no general formal-result correctness, benchmark solve-rate claim, external model access, oracle-to-proof authority, proof body export, source-file changes, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.verifier_lab_execution_spine.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-verifier-lab-execution-spine-diagram-1",
"html_url": "docs/paper-module-verifier-lab-execution-spine.html#paper-module-verifier-lab-execution-spine-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Packet[\"Execution packet transition candidates, CP2 requests, evolve mutations, oracle/provider refs\"] Scan[\"Secret-exclusion scan over every input path\"] Gate[\"Contract check forbidden field? oracle/provider visible? action class out of vocabulary?\"] Rejected[\"CONTRACT_REJECTED no Lean run\"] Build[\"lake build MicrocosmProofWitness once, in a temporary copy\"] Run[\"Write .lean, run lake env lean return code 0 = accepted\"] Verified[\"Accepted transition\"] Residual[\"Residual non-zero return code\"] Records[\"Four result records return codes and counts kept, bodies and proof text omitted\"] Packet --> Scan Scan --> Gate Gate -->|leak found| Rejected Gate -->|clean| Build Build -->|return 0| Run Run -->|exit 0| Verified Run -->|non-zero| Residual Rejected --> Records Verified --> Records Residual --> Records",
"source_format": "mermaid",
"source_hash": "sha256:e880f88f8675449927c9e1c501b3b65ccabfac29c883b9c9bc0089b1c76065fd",
"source_line_end": 165,
"source_line_start": 143
}
],
"grouped_relation_feeds": [
{
"count": 5,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.verifier_lab_kernel",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Kernel",
"relation": "depends on",
"url": "paper-module-verifier-lab-kernel.html"
},
{
"id": "paper_module.formal_math_lean_proof_witness",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Lean Proof Witness",
"relation": "depends on",
"url": "paper-module-formal-math-lean-proof-witness.html"
},
{
"id": "paper_module.formal_math_verifier_trace_repair_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Verifier Trace Repair Loop",
"relation": "depends on",
"url": "paper-module-formal-math-verifier-trace-repair-loop.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "verifier_lab_execution_spine",
"kind": "organ",
"kind_label": "component",
"label": "Verifier Lab Execution Spine",
"relation": "explains",
"url": "component-verifier-lab-execution-spine.html"
},
{
"id": "mechanism.verifier_lab_execution_spine.validates_public_verifier_transition_witness",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public verifier transition witness",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-verifier-lab-execution-spine-validates-public-verifier-transition-witness"
}
],
"relation": "explains"
},
{
"count": 8,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.verifier_lab_execution_spine",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 252,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/verifier_lab_execution_spine.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Lean theorem prover",
"section_id": "small-kernel-trust-and-artifact-evaluation",
"url": "https://www.microsoft.com/en-us/research/publication/the-lean-theorem-prover-system-description/"
},
{
"kind": "external_reference",
"label": "HOL Light",
"section_id": "small-kernel-trust-and-artifact-evaluation",
"url": "https://www.cl.cam.ac.uk/~jrh13/papers/hollight.html"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It runs bounded Lean proof-step candidates in a throwaway copy of a public project and records each checker return code while keeping proof text and payloads out. verifierlabexecutionspine is the component that actually runs bounded Lean proof-step attempts and writes down what the checker said. It copies a small public Lean project into a throwaway workspace, builds it once, writes one .lean file per candidate, runs the Lean checker on each, and records the return code. It keeps the proof text, the tool output, and every provider or oracle payload out of the result. A reader gets the pass or fail facts and the safety counts, and nothing that would leak.",
"level": 4,
"order": 1,
"section_id": "it-runs-bounded-lean-proof-step-candidates-in-a-throwaway-copy-of-a-public-project",
"source_line_end": 14,
"source_line_start": 3,
"title": "It runs bounded Lean proof-step candidates in a throwaway copy of a public project"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Narrower than the kernel, the spine executes each transition, turns a zero exit into accepted, and rejects contaminated candidates before the checker ever runs. The component is narrower than verifierlab_kernel. The kernel reasons about the shape of a transition; this spine executes it. A zero exit code from the Lean checker becomes accepted. Anything else becomes a residual with a failure class. Candidates that carry a forbidden field, or that mark an oracle or provider as visible, are rejected before the checker is ever invoked, so a contaminated run cannot be dressed up as a clean one.",
"level": 4,
"order": 2,
"section_id": "narrower-than-the-kernel",
"source_line_end": 25,
"source_line_start": 15,
"title": "Narrower than the kernel"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It emits four records, full result, board summary, validation, and fixture sign-off, each stating the same scope limit and carrying counts without bodies.** It emits four result records: a full result, a board summary, a validation record, and a fixture-sign-off record. All four state the same scope limit, carry the return codes and counts, and omit the bodies.",
"level": 4,
"order": 3,
"section_id": "it-emits-four-records",
"source_line_end": 33,
"source_line_start": 26,
"title": "It emits four records"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 35,
"source_line_start": 34,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Proof systems can report success while hiding oracle or provider help, so this component answers one question: did a bounded Lean candidate pass the checker unaided?** Automated proof systems can blur how a result was reached. A model can be handed the answer by an oracle, or shown the proof by a provider, and still report the result as if it had found the proof alone. A loose result record hides that distinction: it prints a success without saying what help was in the room. This component exists to keep that blur out of the record. It answers one question. Did a bounded Lean candidate pass the checker, with no help the record is hiding?",
"level": 4,
"order": 5,
"section_id": "keeping-hidden-help-out-of-the-record",
"source_line_end": 46,
"source_line_start": 36,
"title": "Keeping hidden help out of the record"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Authority separation keeps verified candidates, oracle references, and provider hypotheses as distinct classes, pinning the forward-success and export counters at zero by construction.** The discipline is authority separation. A verified candidate is one class. An oracle reference and a provider hypothesis are separate classes that are never counted as forward success. The counters oracleforwardsuccessincrementcount and providerresultscounted are held at zero by construction, alongside proofbodyexportcount and sourcemutation_count. Real execution and a clean record are not in tension: the leak checks run first, so the Lean run that follows is already free of the fields that would compromi\u2026",
"level": 4,
"order": 6,
"section_id": "authority-separation-holds-counters-at-zero",
"source_line_end": 58,
"source_line_start": 47,
"title": "Authority separation holds counters at zero"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 60,
"source_line_start": 59,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run executes Lean for real on the fixture while runexecution_bundle validates an exported group offline, and both flow through buildresult then writereceipts. Two entry points drive the component. run handles the first-wave fixture and executes Lean for real. runexecutionbundle validates an exported group without a Lean toolchain, so a third party can inspect a shipped bundle offline. Both call buildresult, then write_receipts.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-over-build-result",
"source_line_end": 69,
"source_line_start": 61,
"title": "Two entry points over `_build_result`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult loads the packet, secret-scans every input, and builds the copied Lean project once, executing candidates only when that build returns zero. build_result is the spine. It loads the execution packet, scans every input path for forbidden content with the shared secret-exclusion scan, and reads the tool versions. In fixture mode it opens a temporary directory, copies the Lean project into it, and calls buildlake_project, which runs lake build MicrocosmProofWitness once with a sixty-second limit. Only if that build returns zero does it execute the transition candidates.",
"level": 4,
"order": 9,
"section_id": "build-result-scans-copies-and-builds-once",
"source_line_end": 80,
"source_line_start": 70,
"title": "`_build_result` scans, copies, and builds once"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "executetransition first runs the contract check for forbidden keys, visible oracle or provider, and unknown action class, rejecting the candidate or else running lake env lean and recording the return code. Each candidate goes through execute_transition. That function first calls validatetransition_contract. The contract check walks the candidate for any key in FORBIDDENTRANSITIONKEYS (proofbody, rawtactic_script, oracletemplate, provideroutput_body, and similar) using walkforbiddenkeys, checks whether oraclevisible or provider_visible is set, and checks that the action_class is one of the nine names in ALLOWEDACTIONCLASSES. If any check fires, the candidate becomes a CONTRACT_REJECTED reco\u2026",
"level": 4,
"order": 10,
"section_id": "per-candidate-contract-check-in-execute-transition",
"source_line_end": 98,
"source_line_start": 81,
"title": "Per-candidate contract check in `_execute_transition`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Two smaller lanes also run: CP2 translation emits typed next-action candidates from a fixed vocabulary, and evolve mutation touches only bounded policy artifacts, each rejecting out-of-scope requests.** Two smaller lanes run on the same packet. CP2 translation reads residual cases and emits typed next-action candidates drawn from a fixed vocabulary; a request that smuggles a proof body is rejected. Evolve mutation adjusts only the bounded policy artifacts named in ALLOWEDEVOLVEARTIFACTS; a mutation aimed at anything else is rejected.",
"level": 4,
"order": 11,
"section_id": "cp2-translation-and-evolve-mutation-lanes",
"source_line_end": 108,
"source_line_start": 99,
"title": "CP2 translation and evolve mutation lanes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatesourcemoduleimports recomputes each copied file's SHA-256 against the manifest and blocks the group on any digest mismatch, missing anchor, or out-of-prefix source. validatesourcemoduleimports handles the exported group. It reads sourcemodulemanifest.json, confirms the declared module count matches the rows, and for each row recomputes the SHA-256 digest of the copied file and compares it to the manifest. A digest mismatch, a missing anchor, or a source reference outside the allowed prefixes blocks the group. The copied bodies live in the bundle, never in the record.",
"level": 4,
"order": 12,
"section_id": "digest-checks-on-the-exported-group",
"source_line_end": 119,
"source_line_start": 109,
"title": "Digest checks on the exported group"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "writereceipts composes the four records with every authority flag false, and the status passes only when the scan is clean, the build returned zero, and the expected transition counts are met. writereceipts composes the four records and pins the safety facts: receiptsincludeproofbodies false, providercalls_authorized false, sourcemutationauthorized false, release_authorized false. The overall status in buildresult passes only when the secret scan is clean, Lean and Lake are available, the project build returned zero, no expected negative case is missing, and the run produced at least four transitions, two accepted, one residual, one CP2 effect, and one accepted mutation.",
"level": 4,
"order": 13,
"section_id": "write-receipts-pins-the-zero-authority-facts",
"source_line_end": 166,
"source_line_start": 120,
"title": "`write_receipts` pins the zero-authority facts"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 168,
"source_line_start": 167,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The fixture ships four negative inputs listed in EXPECTEDNEGATIVE_CASES, each expected to raise one error code, and a run is blocked unless all four are observed.** The fixture ships four negative inputs, each expected to raise one error code, declared in EXPECTEDNEGATIVECASES. A run that fails to observe all four is blocked. transitionleakscandidate_body must raise VERIFIERLABEXECUTIONTRANSITIONFIELD_FORBIDDEN: a candidate carrying a proof body or raw tactic script. provideroraclevisible_transition must raise VERIFIERLABEXECUTIONPROVIDERORORACLEVISIBLE: a candidate that exposes an oracle structured source record or provider text to forward execution. cp2candidatecontainsproofbody must rais\u2026",
"level": 4,
"order": 15,
"section_id": "four-negative-inputs-the-run-must-observe",
"source_line_end": 189,
"source_line_start": 169,
"title": "Four negative inputs the run must observe"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 191,
"source_line_start": 190,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Execution follows the small-kernel proof-assistant tradition where trust rests on the checker not the tactic, and the record shape follows artifact-evaluation practice.** The Lean and Lake execution inherits from the small-kernel proof-assistant tradition, the Lean theorem prover and LCF-style systems such as HOL Light, where trust rests on a small checker rather than the tactic that produced the term. The record shape follows artifact-evaluation practice: record the command identity, the tool facts, and the pass or fail counts separately from the claim they support, and keep the underlying artifact out of the summary.",
"level": 4,
"order": 17,
"section_id": "small-kernel-trust-and-artifact-evaluation",
"source_line_end": 203,
"source_line_start": 192,
"title": "Small-kernel trust and artifact evaluation"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from microcosm-substrate:",
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 222,
"source_line_start": 204,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A green run counts as bounded execution-spine evidence only and never as general proof certification.** A green run proves only bounded execution-spine evidence: command intent, tool facts, return-code pass or fail per candidate, the four observed negative cases, and the zero-held authority counters.",
"level": 4,
"order": 19,
"section_id": "what-a-green-run-proves",
"source_line_end": 229,
"source_line_start": 223,
"title": "What a green run proves"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 231,
"source_line_start": 230,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A bounded set of public Lean candidates ran in a throwaway copy of a public project, and each checker return code was recorded without exposing proof text.** The strongest claim the evidence supports is narrow. A bounded set of public Lean candidates was run in a throwaway copy of a public project, and the checker's return code for each was recorded without exposing the proof text. That is the whole claim.",
"level": 4,
"order": 21,
"section_id": "the-strongest-supported-claim",
"source_line_end": 240,
"source_line_start": 232,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Past that it certifies no general proofs, carries no Mathlib authority, exports no bodies, calls no provider, mutates no source, and excludes launch.** It refuses everything past that. It does not certify proofs in general, does not carry Mathlib-dependent proof authority, does not export proof bodies, does not certify a benchmark solve rate, does not call any provider, does not change source files, and excludes launch or hosted deployment. An oracle match is never forward success. Provider text is never a proof. The diagram and Atlas views are navigation aids and do not promote a tool run into proof certification.",
"level": 4,
"order": 22,
"section_id": "what-it-refuses-past-that-line",
"source_line_end": 251,
"source_line_start": 241,
"title": "What it refuses past that line"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/verifier_lab_execution_spine.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Bounded public verifier execution result records witness command execution without upgrading output into theorem authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "verifier_lab_execution_spine.py",
"path": "src/microcosm_core/organs/verifier_lab_execution_spine.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for bounded verifier execution fixtures, source-module import checks, temporary Lake-project execution, negative cases, and scope limits.",
"symbols": [
"run",
"run_execution_bundle",
"validate_source_module_imports",
"_build_lake_project",
"_build_result",
"write_receipts",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/verifier_lab_execution_spine.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/verifier_lab_execution_spine.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/verifier_lab_execution_spine.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Verifier Lab Execution Spine records bounded public Lean transition execution evidence: command intent, tool facts, return codes, result record refs, omitted dangerous payload fields, negative cases, and source-open body imports. It separates real execution evidence from formal-result correctness, provider text, oracle answers, proof-body exposure, source-file changes, and launch-scope decision.",
"summary_line": "Bounded public verifier execution result records witness command execution without upgrading output into theorem authority.",
"title": "Verifier Lab Execution Spine"
},
{
"anchor": "paper-module-bounded-autonomy-campaign-packet",
"authority_ceiling": "Self-proposal campaign packet fixture and exported-bundle result record evidence only; no self-repair authority, unsupervised source-file changes, live scheduler authority, external model access, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
Coverage gaps become proposals without action authority
\n
This component lets an agent draft a plan to close its own coverage gaps while stamping every row as a proposal that carries no authority to act.bounded_autonomy_campaign_packet is a component that lets an agent draft a plan to close its own coverage gaps without giving that plan any authority to act. It reads synthetic public inputs, watches a read-only campaign builder, and emits a draft candidate packet where every row is stamped as a proposal, not a change.
\n
The packet carries hashes and refs, never writes
\n
The result record carries only refs, digests, line counts, and stdout and stderr hashes, never a write surface, a true source-file changes flag, or the builder's output bodies. The result record it writes carries refs, digests, line counts, and stdout and stderr hashes. It never carries a write surface, a source-file changes flag set to true, or the builder's output bodies.
\n
The demonstration separates proposing from doing
\n
What it demonstrates is narrow: an agent can propose work drawn from a real surface capable of doing it, observed in a mode where it did not. What it demonstrates is narrow and deliberate: an agent can propose work drawn from a real surface that could do that work, observed in a mode where it did not.
\n
Purpose
\n
The risk of sketching sliding into doing
\n
An agent can usefully sketch a plan to fix its under-coverage, but the risk is that sketching quietly becomes doing as a proposal grows a write surface. An agent can usefully notice where it is under-covered and sketch a plan to fix it. The risk is that "sketch a plan" quietly becomes "do the work". A proposal grows a write surface, and a system meant to suggest starts editing its own source unsupervised.
\n
Keeping proposing and acting apart
\n
This component keeps proposing and acting apart, answering whether an agent can emit a draft campaign proposal from real coverage gaps without that proposal carrying any authority to act. This component keeps those two steps apart. It answers one question: can an agent emit a draft campaign proposal from real coverage gaps without that proposal carrying any authority to act on them?
\n
Candidate count comes from a witnessed builder
\n
The candidate count comes not from an invented list but from a check-only builder witness, and each candidate is stamped with no write surface, no source-file changes, and required review. The design choice that matters is where the candidate count comes from. The component does not invent a plausible list of work. It watches a real campaign builder run in a check-only mode and accepts its witness only when the builder reports candidate targets and leaves wrote_packet unset. Each drafted candidate is then stamped write_surface: none, source_mutation_authorized: false, and requires_human_review: true, so proposing can never be read as authorising.
\n
How it works
\n
Two entry points share the same checks
\n
run and run_bounded_autonomy_bundle both route through run_crown_jewel_organ with the same evaluate and evaluate_negative_case evaluators, so the identical checks guard both command surfaces.run and run_bounded_autonomy_bundle both hand the fixture and the exported bundle to run_crown_jewel_organ, passing evaluate as the evaluator and evaluate_negative_case as the negative-case evaluator. The bundle entry point adds the bundle input mode. The same checks guard both command surfaces.
\n
evaluate runs four ordered refusal checks
\n
evaluate loads three inputs and runs four ordered checks, refusing a source-write policy, a failed witness, an empty list, or an authorising candidate, and passing only when no finding remains.evaluate loads three declared inputs: coverage_gaps.json, campaign_policy.json, and failed_campaign_digests.json. It then runs four checks in order and collects a finding for each failure. First, if write_source appears in the policy's allowed_actions, it records BOUNDED_AUTONOMY_SOURCE_WRITE_FORBIDDEN before any candidate is drafted. Second, it builds the candidate packet and refuses it if the builder witness did not pass or the packet is not marked drafted. Third, it refuses an empty candidate list. Fourth, it refuses any candidate that sets source_mutation_authorized to true or writes to the source surface. It also counts repeated entries in the failed digest ledger and records BOUNDED_AUTONOMY_REPEATED_FAILED_DIGEST if any digest appears more than once. A run passes only when the finding list is empty.
\n
Turning the witnessed count into stamped rows
\n
_candidate_packet_subprocess caps the witnessed count at the policy's max_candidate_count and turns it into draft rows, each tied to a coverage gap and stamped with no write surface and required human review._candidate_packet_subprocess turns the witnessed target count into draft rows. It reads max_candidate_count from the policy (default 2), asks the witness for that many targets, and takes the smaller of the two counts. Each row ties to one coverage gap when one is available, carries the builder ref and the builder source digest, and sets write_surface: none, requires_human_review: true, and source_mutation_authorized: false.
\n
Witnessing the builder in check-only mode
\n
_campaign_builder_witness runs the real builder as a check-only subprocess when present, or falls back to the copied source-module manifest in the public clone, storing stdout and stderr hashes rather than bodies. The witness itself is _campaign_builder_witness. It looks for the builder named by BUILDER_REF at the repository root. When that file is present it runs it as a subprocess with --check --report --max-targets, with a 45 second timeout, and accepts the witness only when the process exits zero, returns kind standard_skill_pairing_campaign_summary, is not a no-op, reports at least one candidate target, emits a source_digest, and leaves wrote_packet unset. When the builder file is absent, which is the case in the public clone, it falls back to _public_source_module_builder_witness, which reads the copied source-module manifest and passes only when the target file's sha256 matches, its line count matches, and no required anchor is missing. Either way the witness stores stdout and stderr hashes, not bodies.
\n
Function
Role
run, run_bounded_autonomy_bundle
Route the fixture and exported bundle through run_crown_jewel_organ with the shared evaluators
evaluate
Load the three inputs, run the policy, witness, candidate, and digest checks, return findings
_candidate_packet_subprocess
Convert the witnessed target count into stamped draft candidate rows
_campaign_builder_witness
Watch the real builder in check-only mode, or fall back to the source-module manifest
evaluate_negative_case
Mutate a copy of the input and re-run evaluate to confirm the refusal fires
\n
Module constants pin the boundary: EXPECTED_NEGATIVE_CASES maps mutations to codes, AUTHORITY_CEILING sets every autonomy flag false, and SPEC declares the required inputs, manifest ref, and anchors. The module constants pin the boundary. EXPECTED_NEGATIVE_CASES maps each named mutation to the error code it must produce. AUTHORITY_CEILING sets self_repair_authorized, unsupervised_source_mutation_authorized, source_write_packet_authorized, provider_calls_authorized, and release_authorized all to false. SPEC declares the required inputs, the source manifest ref, and the required anchors.
\nDiagram of the mechanism (8 steps).Diagram source & refs
The refusals are checked by mutating the input rather than trusting a declared label, as evaluate_negative_case copies the inputs, edits one file, and re-runs evaluate. The refusals are checked by mutating the input, not by trusting a declared label. evaluate_negative_case copies the required inputs into a temporary directory, edits one file, and re-runs evaluate.
\n
The source-write policy case
\n
The source_write_campaign_packet case appends write_source to the policy's allowed_actions and must produce BOUNDED_AUTONOMY_SOURCE_WRITE_FORBIDDEN. The source_write_campaign_packet case appends write_source to the policy's allowed_actions. It must produce BOUNDED_AUTONOMY_SOURCE_WRITE_FORBIDDEN.
\n
The repeated failed-digest case
\n
The repeated_failed_campaign_digest case duplicates a digest in the failed-digest ledger and must produce BOUNDED_AUTONOMY_REPEATED_FAILED_DIGEST, since an already-failed plan cannot be re-proposed under a fresh wrapper. The repeated_failed_campaign_digest case rewrites the failed-digest ledger to hold the same digest twice. It must produce BOUNDED_AUTONOMY_REPEATED_FAILED_DIGEST. A plan that has already failed cannot be re-proposed under a fresh wrapper.
\n
Unknown negative cases fail with one stable code
\n
Any negative-case id outside the declared set is rejected as unsupported with the stable bounded-autonomy error code. A negative case with any other id returns BOUNDED_AUTONOMY_NEGATIVE_CASE_UNSUPPORTED.
\n
Prior Art Grounding
\n
Preflight gating of proposed actions
\n
The component follows established preflight-gate patterns from NIST AI RMF, Open Policy Agent, and GitHub Actions, applying that shape to draft self-proposals over synthetic public coverage gaps. The component follows established patterns for gating proposed actions before they run. The NIST AI Risk Management Framework frames AI work as governing, mapping, measuring, and managing risk rather than assuming autonomy is authorized. Open Policy Agent is a policy-engine pattern for deciding whether an action may proceed. GitHub Actions workflow syntax is a common automation surface where jobs and permissions are declared before execution. This component keeps the preflight-gate shape and applies it to draft self-proposals over synthetic public coverage gaps.
\n
Validation Result record Path
\n
What a pass means and how to route drift
\n
A pass means both fixture and bundle drafted a packet, both negative cases fired their codes, and the record stayed metadata-only, with digest drift routed through microcosm_exact_copy_refresh. A pass means the fixture and bundle both produced a draft packet, both named negative cases fired their expected error codes, and the result record stayed metadata-only. If the source-module manifest reports digest drift, route that through microcosm_exact_copy_refresh; this page is not the authority for copied source bodies. If the full projection check fails because another session holds the shared lattice outputs, treat that as unrelated contention and use the corpus check as the local gate.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim is that under public synthetic inputs the component emits a draft self-proposal from a read-only witness and refuses source-write or repeated-digest inputs, both checked by mutation. The strongest claim the evidence supports is this: under public synthetic inputs, the component emits a draft self-proposal derived from a read-only builder witness and refuses a source-write policy or a repeated failed digest, with both refusals checked by mutating the input. That is the proof boundary and the scope limit for this module.
\n
What the module does not claim
\n",
"components": [
{
"anchor": "component-bounded_autonomy_campaign_packet",
"component_id": "bounded_autonomy_campaign_packet",
"public_label": "Bounded Autonomy Campaign Packet"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 7
},
"link_count": 19,
"links": [
{
"id": "bounded_autonomy_campaign_packet",
"kind": "organ",
"kind_label": "component",
"label": "Bounded Autonomy Campaign Packet",
"relation": "explains",
"url": "component-bounded-autonomy-campaign-packet.html"
},
{
"id": "mechanism.bounded_autonomy_campaign_packet.validates_public_bounded_autonomy_campaign_packet",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public bounded autonomy campaign packet",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-bounded-autonomy-campaign-packet-validates-public-bounded-autonomy-campaign-packet"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.provider_context_recipe_budget",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Provider Context Recipe Budget",
"relation": "depends on",
"url": "paper-module-provider-context-recipe-budget.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.bounded_autonomy_campaign_packet",
"reader_packet": {
"abstract": "Bounded autonomy campaign packets propose guarded agent work without authorizing source-file changes or unsupervised repair.",
"authority_ceiling": "Self-proposal campaign packet fixture and exported-bundle result record evidence only; no self-repair authority, unsupervised source-file changes, live scheduler authority, external model access, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.bounded_autonomy_campaign_packet.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-bounded-autonomy-campaign-packet-diagram-1",
"html_url": "docs/paper-module-bounded-autonomy-campaign-packet.html#paper-module-bounded-autonomy-campaign-packet-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Inputs[\"Public synthetic inputs coverage_gaps, campaign_policy, failed_campaign_digests\"] PolicyGate{\"policy allows write_source?\"} Witness[\"Builder witness check-only builder run or source-module manifest\"] WitnessGate{\"reports candidates and wrote_packet unset?\"} Draft[\"Draft candidate rows write_surface none, requires_human_review, source_mutation false\"] DigestGate{\"failed digest repeated?\"} Refuse[\"Refuse source-write, blocked witness, or repeated digest\"] Record[\"metadata-only result record refs, digests, stdout/stderr hashes\"] Inputs --> PolicyGate PolicyGate -- \"yes\" --> Refuse PolicyGate -- \"no\" --> Witness Witness --> WitnessGate WitnessGate -- \"no\" --> Refuse WitnessGate -- \"yes\" --> Draft Draft --> DigestGate DigestGate -- \"yes\" --> Refuse DigestGate -- \"no\" --> Record Refuse --> Record",
"source_format": "mermaid",
"source_hash": "sha256:7a8abed446817e7e686956efcbac2bdb83efbf0279114036b6b58e97551d6245",
"source_line_end": 156,
"source_line_start": 135
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.provider_context_recipe_budget",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Provider Context Recipe Budget",
"relation": "depends on",
"url": "paper-module-provider-context-recipe-budget.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "bounded_autonomy_campaign_packet",
"kind": "organ",
"kind_label": "component",
"label": "Bounded Autonomy Campaign Packet",
"relation": "explains",
"url": "component-bounded-autonomy-campaign-packet.html"
},
{
"id": "mechanism.bounded_autonomy_campaign_packet.validates_public_bounded_autonomy_campaign_packet",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public bounded autonomy campaign packet",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-bounded-autonomy-campaign-packet-validates-public-bounded-autonomy-campaign-packet"
}
],
"relation": "explains"
},
{
"count": 8,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.bounded_autonomy_campaign_packet",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 255,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/bounded_autonomy_campaign_packet.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "NIST AI Risk Management Framework",
"section_id": "preflight-gating-of-proposed-actions",
"url": "https://www.nist.gov/itl/ai-risk-management-framework"
},
{
"kind": "external_reference",
"label": "Open Policy Agent",
"section_id": "preflight-gating-of-proposed-actions",
"url": "https://www.openpolicyagent.org/docs/latest"
},
{
"kind": "external_reference",
"label": "workflow syntax",
"section_id": "preflight-gating-of-proposed-actions",
"url": "https://docs.github.com/en/actions/reference/workflows-and-actions/workflow-syntax"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component lets an agent draft a plan to close its own coverage gaps while stamping every row as a proposal that carries no authority to act.** boundedautonomycampaign_packet is a component that lets an agent draft a plan to close its own coverage gaps without giving that plan any authority to act. It reads synthetic public inputs, watches a read-only campaign builder, and emits a draft candidate packet where every row is stamped as a proposal, not a change.",
"level": 4,
"order": 1,
"section_id": "coverage-gaps-become-proposals-without-action-authority",
"source_line_end": 11,
"source_line_start": 3,
"title": "Coverage gaps become proposals without action authority"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The result record carries only refs, digests, line counts, and stdout and stderr hashes, never a write surface, a true source-file changes flag, or the builder's output bodies.** The result record it writes carries refs, digests, line counts, and stdout and stderr hashes. It never carries a write surface, a source-file changes flag set to true, or the builder's output bodies.",
"level": 4,
"order": 2,
"section_id": "the-packet-carries-hashes-and-refs-never-writes",
"source_line_end": 19,
"source_line_start": 12,
"title": "The packet carries hashes and refs, never writes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**What it demonstrates is narrow: an agent can propose work drawn from a real surface capable of doing it, observed in a mode where it did not.** What it demonstrates is narrow and deliberate: an agent can propose work drawn from a real surface that could do that work, observed in a mode where it did not.",
"level": 4,
"order": 3,
"section_id": "the-demonstration-separates-proposing-from-doing",
"source_line_end": 26,
"source_line_start": 20,
"title": "The demonstration separates proposing from doing"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 28,
"source_line_start": 27,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**An agent can usefully sketch a plan to fix its under-coverage, but the risk is that sketching quietly becomes doing as a proposal grows a write surface.** An agent can usefully notice where it is under-covered and sketch a plan to fix it. The risk is that \"sketch a plan\" quietly becomes \"do the work\". A proposal grows a write surface, and a system meant to suggest starts editing its own source unsupervised.",
"level": 4,
"order": 5,
"section_id": "the-risk-of-sketching-sliding-into-doing",
"source_line_end": 37,
"source_line_start": 29,
"title": "The risk of sketching sliding into doing"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component keeps proposing and acting apart, answering whether an agent can emit a draft campaign proposal from real coverage gaps without that proposal carrying any authority to act.** This component keeps those two steps apart. It answers one question: can an agent emit a draft campaign proposal from real coverage gaps without that proposal carrying any authority to act on them?",
"level": 4,
"order": 6,
"section_id": "keeping-proposing-and-acting-apart",
"source_line_end": 45,
"source_line_start": 38,
"title": "Keeping proposing and acting apart"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The candidate count comes not from an invented list but from a check-only builder witness, and each candidate is stamped with no write surface, no source-file changes, and required review.** The design choice that matters is where the candidate count comes from. The component does not invent a plausible list of work. It watches a real campaign builder run in a check-only mode and accepts its witness only when the builder reports candidate targets and leaves wrote_packet unset. Each drafted candidate is then stamped write_surface: none, sourcemutationauthorized: false, and requireshumanreview: true, so proposing can never be read as authorising.",
"level": 4,
"order": 7,
"section_id": "candidate-count-comes-from-a-witnessed-builder",
"source_line_end": 57,
"source_line_start": 46,
"title": "Candidate count comes from a witnessed builder"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 59,
"source_line_start": 58,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run and runboundedautonomybundle both route through runcrownjewelorgan with the same evaluate and evaluatenegative_case evaluators, so the identical checks guard both command surfaces.** run and runboundedautonomy_bundle both hand the fixture and the exported bundle to runcrownjewel_organ, passing evaluate as the evaluator and evaluatenegativecase as the negative-case evaluator. The bundle entry point adds the bundle input mode. The same checks guard both command surfaces.",
"level": 4,
"order": 9,
"section_id": "two-entry-points-share-the-same-checks",
"source_line_end": 69,
"source_line_start": 60,
"title": "Two entry points share the same checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**evaluate loads three inputs and runs four ordered checks, refusing a source-write policy, a failed witness, an empty list, or an authorising candidate, and passing only when no finding remains.** evaluate loads three declared inputs: coveragegaps.json, campaignpolicy.json, and failedcampaigndigests.json. It then runs four checks in order and collects a finding for each failure. First, if write_source appears in the policy's allowedactions, it records BOUNDEDAUTONOMYSOURCEWRITE_FORBIDDEN before any candidate is drafted. Second, it builds the candidate packet and refuses it if the builder witness did not pass or the packet is not marked drafted. Third, it refuses an empty candidate list. Fo\u2026",
"level": 4,
"order": 10,
"section_id": "evaluate-runs-four-ordered-refusal-checks",
"source_line_end": 86,
"source_line_start": 70,
"title": "`evaluate` runs four ordered refusal checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "candidatepacket_subprocess caps the witnessed count at the policy's maxcandidatecount and turns it into draft rows, each tied to a coverage gap and stamped with no write surface and required human review.** candidatepacket_subprocess turns the witnessed target count into draft rows. It reads maxcandidatecount from the policy (default 2), asks the witness for that many targets, and takes the smaller of the two counts. Each row ties to one coverage gap when one is available, carries the builder ref and the builder source digest, and sets writesurface: none, requireshuman_review: true, and sourcemutationauthorized: false.",
"level": 4,
"order": 11,
"section_id": "turning-the-witnessed-count-into-stamped-rows",
"source_line_end": 98,
"source_line_start": 87,
"title": "Turning the witnessed count into stamped rows"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "campaignbuilder_witness runs the real builder as a check-only subprocess when present, or falls back to the copied source-module manifest in the public clone, storing stdout and stderr hashes rather than bodies.** The witness itself is campaignbuilderwitness. It looks for the builder named by BUILDERREF at the repository root. When that file is present it runs it as a subprocess with --check --report --max-targets, with a 45 second timeout, and accepts the witness only when the process exits zero, returns kind standardskillpairingcampaignsummary, is not a no-op, reports at least one candidate target, emits a sourcedigest, and leaves wrotepacket unset. When the builder file is absent, which\u2026",
"level": 4,
"order": 12,
"section_id": "witnessing-the-builder-in-check-only-mode",
"source_line_end": 123,
"source_line_start": 99,
"title": "Witnessing the builder in check-only mode"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "Module constants pin the boundary: EXPECTEDNEGATIVE_CASES maps mutations to codes, scope_limit sets every autonomy flag false, and SPEC declares the required inputs, manifest ref, and anchors.** The module constants pin the boundary. EXPECTEDNEGATIVECASES maps each named mutation to the error code it must produce. AUTHORITYCEILING sets selfrepair_authorized, unsupervisedsourcemutationauthorized, sourcewritepacketauthorized, providercallsauthorized, and release_authorized all to false. SPEC declares the required inputs, the source manifest ref, and the required anchors.",
"level": 4,
"order": 13,
"section_id": "module-constants-pin-the-boundary",
"source_line_end": 157,
"source_line_start": 124,
"title": "Module constants pin the boundary"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 159,
"source_line_start": 158,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The refusals are checked by mutating the input rather than trusting a declared label, as evaluatenegativecase copies the inputs, edits one file, and re-runs evaluate.** The refusals are checked by mutating the input, not by trusting a declared label. evaluatenegativecase copies the required inputs into a temporary directory, edits one file, and re-runs evaluate.",
"level": 4,
"order": 15,
"section_id": "refusals-checked-by-mutating-the-input",
"source_line_end": 167,
"source_line_start": 160,
"title": "Refusals checked by mutating the input"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The sourcewritecampaignpacket case appends write_source to the policy's allowedactions and must produce BOUNDEDAUTONOMYSOURCEWRITEFORBIDDEN. The sourcewritecampaignpacket case appends writesource to the policy's allowedactions. It must produce BOUNDEDAUTONOMYSOURCEWRITE_FORBIDDEN.",
"level": 4,
"order": 16,
"section_id": "the-source-write-policy-case",
"source_line_end": 174,
"source_line_start": 168,
"title": "The source-write policy case"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The repeatedfailedcampaigndigest case duplicates a digest in the failed-digest ledger and must produce BOUNDEDAUTONOMYREPEATEDFAILEDDIGEST, since an already-failed plan cannot be re-proposed under a fresh wrapper.** The repeatedfailedcampaign_digest case rewrites the failed-digest ledger to hold the same digest twice. It must produce BOUNDEDAUTONOMYREPEATEDFAILEDDIGEST. A plan that has already failed cannot be re-proposed under a fresh wrapper.",
"level": 4,
"order": 17,
"section_id": "the-repeated-failed-digest-case",
"source_line_end": 184,
"source_line_start": 175,
"title": "The repeated failed-digest case"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Any negative-case id outside the declared set is rejected as unsupported with the stable bounded-autonomy error code. A negative case with any other id returns BOUNDEDAUTONOMYNEGATIVECASEUNSUPPORTED.",
"level": 4,
"order": 18,
"section_id": "unknown-negative-cases-fail-with-one-stable-code",
"source_line_end": 189,
"source_line_start": 185,
"title": "Unknown negative cases fail with one stable code"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "prior-art-grounding",
"source_line_end": 191,
"source_line_start": 190,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component follows established preflight-gate patterns from NIST AI RMF, Open Policy Agent, and GitHub Actions, applying that shape to draft self-proposals over synthetic public coverage gaps.** The component follows established patterns for gating proposed actions before they run. The NIST AI Risk Management Framework frames AI work as governing, mapping, measuring, and managing risk rather than assuming autonomy is authorized. Open Policy Agent is a policy-engine pattern for deciding whether an action may proceed. GitHub Actions workflow syntax is a common automation surface where jobs and permissions are declared before execution. This component keeps the preflight-gate shape and ap\u2026",
"level": 4,
"order": 20,
"section_id": "preflight-gating-of-proposed-actions",
"source_line_end": 206,
"source_line_start": 192,
"title": "Preflight gating of proposed actions"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "tracked result record outputs:",
"level": 2,
"order": 21,
"section_id": "validation-receipt-path",
"source_line_end": 219,
"source_line_start": 207,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means both fixture and bundle drafted a packet, both negative cases fired their codes, and the record stayed metadata-only, with digest drift routed through microcosmexactcopyrefresh. A pass means the fixture and bundle both produced a draft packet, both named negative cases fired their expected error codes, and the result record stayed metadata-only. If the source-module manifest reports digest drift, route that through microcosmexactcopy_refresh; this page is not the authority for copied source bodies. If the full projection check fails because another session holds the shared lattice outputs, treat that as unrelated contention and use the corpus check as the local gate.",
"level": 4,
"order": 22,
"section_id": "what-a-pass-means-and-how-to-route-drift",
"source_line_end": 231,
"source_line_start": 220,
"title": "What a pass means and how to route drift"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 23,
"section_id": "claim-ceiling",
"source_line_end": 233,
"source_line_start": 232,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest supported claim is that under public synthetic inputs the component emits a draft self-proposal from a read-only witness and refuses source-write or repeated-digest inputs, both checked by mutation.** The strongest claim the evidence supports is this: under public synthetic inputs, the component emits a draft self-proposal derived from a read-only builder witness and refuses a source-write policy or a repeated failed digest, with both refusals checked by mutating the input. That is the proof boundary and the scope limit for this module.",
"level": 4,
"order": 24,
"section_id": "the-strongest-supported-claim",
"source_line_end": 244,
"source_line_start": 234,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It claims no autonomous repair, unsupervised source-file changes, external model access, launch or public sharing decision, production safety, or whole-system correctness, claim autonomous repair, unsupervised source-file changes, external model access, launch-scope decision, public sharing decision, production campaign safety, whole-system equivalence, or whole-system correctness. If those claims change, the authority is",
"level": 4,
"order": 25,
"section_id": "what-the-module-does-not-claim",
"source_line_end": 254,
"source_line_start": 245,
"title": "What the module does not claim"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/bounded_autonomy_campaign_packet.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Bounded autonomy campaign packets propose guarded agent work without authorizing source-file changes or unsupervised repair."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "bounded_autonomy_campaign_packet.py",
"path": "src/microcosm_core/organs/bounded_autonomy_campaign_packet.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for bounded campaign packet fixture evaluation, failed-campaign digest refusal, negative cases, and scope limits.",
"symbols": [
"run",
"run_bounded_autonomy_bundle",
"evaluate",
"_candidate_packet_subprocess",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/bounded_autonomy_campaign_packet.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/bounded_autonomy_campaign_packet.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/bounded_autonomy_campaign_packet.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Bounded Autonomy Campaign Packet validates public campaign packet fixtures: proposed gaps, policy gates, repeated-failure digests, negative cases, source-open body imports, and scope limits. It keeps campaign proposals separate from self-repair, source writes, live scheduling, external model access, launch-scope decision, public sharing, and whole-system correctness.",
"summary_line": "Bounded autonomy campaign packets propose guarded agent work without authorizing source-file changes or unsupervised repair.",
"title": "Bounded Autonomy Campaign Packet"
},
{
"anchor": "paper-module-computer-use-action-trace-replay",
"authority_ceiling": "Synthetic public computer-use action-trace fixtures, exported bundle metadata, copied source-module digests, and metadata-only result records only; no live account action, account secret entry, external network mutation, purchase/send authority, destructive host action, hidden screen-state claim, benchmark-score claim, launch-scope decision, or whole-system correctness.",
"body_html": "
Component answers offline
\n
The component answers, offline over a synthetic trace, whether a blocked action was later narrated as success and whether a claimed success has any state evidence behind it. A computer-use agent leaves behind a stream of screenshots, clicks, keystrokes, and "it worked" assertions. Reading that stream by hand does not tell you the two things that matter: whether an action that was blocked or sent for review got narrated later as a success, and whether a claimed success has any state evidence behind it. This component answers those two questions on a synthetic trace, offline, before any of the language reaches a reader.
\n
Takes recorded local-agent
\n
It takes a recorded local-agent run with its observations, actions, pre-action verdicts, and state changes, checking the trail is self-consistent against authority limits written down up front. It takes a recorded run of a local agent, the route it picked, the observations it logged, the actions it took, the authority verdict stamped before each action, and the state change each action produced, and checks that this recorded trail is well-formed and self-consistent. The record states up front where the agent's authority was supposed to stop, so the limits are written down and checkable rather than taken on faith.
\n
Result is metadata-only
\n
The result is a metadata-only record of counts, refs, hashes, and redaction posture over a recorded trail; it never watches a live agent. The result is a metadata-only record of counts, references, hashes, and redaction posture. It checks a recorded trail. It does not watch a live agent and it does not establish one stayed in bounds.
\n
Purpose
\n
The complete-looking trace that hides a lie
\n
It guards against a complete-looking trace that hides a lie: a denied action reported as done, a success asserted with no oracle, a screenshot treated as proof. The failure this guards against is a trace that looks complete while hiding a lie. An action can be denied by policy and then reported as done. A success can be asserted with no diff, no oracle, and no before-and-after state. A screenshot can be treated as proof when the pixels prove nothing about what actually changed.
\n
A typed join of verdict and transition
\n
Every action must line up with a prior observation, a pre-action verdict, and a transition whose outcome matches that verdict, through a typed join rather than a screenshot replay. A loose reading of the log cannot catch these. You need every recorded action to line up, row by row, with a prior visible observation, a pre-action authority verdict, and a state-transition record whose outcome agrees with that verdict. If the verdict said allow, the transition has to show the action ran and an oracle confirmed the new state. If the verdict said block or review, the transition has to show the action did not run and the status reads blocked or review_required. The mechanism is a typed join across those tables, not a screenshot replay.
\n
The screenshot is the one distrusted input
\n
Inverting the usual trust, observations enter only as a screenshot_digest and visible_state_hash, with raw pixels and hidden state barred, so weight rests on verdict-transition agreement. The inversion is the point. Most action-trace tooling trusts the screenshot. This component treats the screenshot as the one thing it will not trust: observations enter only as a screenshot_digest and a visible_state_hash, and raw pixel bodies, hidden-state claims, and live-browser state are all required to be absent. The evidence that carries weight is the agreement between verdict and transition.
\n
How it works
\n
run_computer_use_action_trace_bundle and the check sequence
\n
run_computer_use_action_trace_bundle resolves the input, decides fixture versus exported bundle, loads it through strict JSON, then runs a fixed sequence of checks before writing anything. The entry point is run_computer_use_action_trace_bundle. It resolves the input directory, decides whether the input is the full fixture with negative cases or the public exported bundle, loads the bundle through a strict JSON path, then runs a fixed sequence of checks in this order before it writes anything.
\nDiagram of the mechanism (12 steps).Diagram source & refs
Each stage consumes the rows the earlier stage exported, so a break early in the chain propagates downstream. Each stage consumes the rows the earlier stage exported, so a break early in the chain propagates. The relevant functions:
\n
Function
Consumes
Checks and rejects
_validate_computer_use_observations
screen observations, episode rows
Each observation is a redacted visible-state row with affordance refs and a screenshot digest. Rejects raw_screenshot_body_exported, hidden_state_claim, or live_browser_state set true. Requires every episode to have at least one observation.
_validate_computer_use_actions
action trace, episodes, observations
Each action cites an observation and an affordance, uses an allowed action kind, stays redacted, and does not precede the observation it reacts to. Requires the trace to cover click, type, navigate, edit_text_record, and wait.
_validate_computer_use_authority_verdicts
verdicts, action rows
Each action has one pre-action verdict of allow, block, or review that explicitly sets live_account_authorized, account secret entry authorized, external_network_mutation_authorized, destructive_without_review_authorized, and purchase_or_send_authorized to false. Flags any action with no verdict.
_validate_computer_use_state_transitions
transitions, actions, verdicts
Matches each transition to its action's verdict. An allow needs execution_attempted true and oracle_status pass. A block or review needs execution_attempted false and oracle_status in blocked or review_required. Rejects any nondeterministic_success_claim.
_validate_computer_use_recovery_receipts
recovery rows, actions, verdicts
Every blocked or reviewed action needs a redacted recovery row with recovery_status recovered and state_restored true. A blocked action is never upgraded into an executed mutation.
_validate_computer_use_cold_replay
cold replay rows, actions
Replay rows must cover every action id and reproduce the action trace, the authority verdicts, and the state transitions, with a pass_label of accepted or blocked_recovered.
\n
Public trace spans and the pass rule
\n
build_public_computer_use_trace emits bounded public spans, and status is pass only when findings are empty, no negative case is missing, the scan passes, and source-module digests verify. After the checks, build_public_computer_use_trace emits bounded public trace spans over refs, counts, and hashes. run_computer_use_action_trace_bundle collects the findings from every stage. It sets status to pass only when the positive findings are empty, no expected negative case is missing, the non-public-state scan passes, and the copied public source-module digests verify. Otherwise the status is blocked. The written record carries counts, refs, hashes, and the redaction posture, and omits raw screen bodies, account secrets, hidden screen state, model-output data, private source bodies, absolute local paths, and benchmark-score claims.
\n
What the tracked fixture record reports
\n
The tracked fixture record reports fixed counts: four episodes, eight actions and verdicts, eight transitions, one recovery, four cold-replay rows, one blocked action, and eight trace spans. The tracked fixture record reports four episodes, six observations, eight actions, eight authority verdicts, eight state-transition records, one recovery record, four cold-replay rows, one blocked action, and eight public trace spans.
\n
Negative cases
\n
Eight declared perturbations on the fixture path
\n
_validate_computer_use_negative_cases feeds eight declared perturbations through the same posture and requires each to be rejected with its own code, failing the run if any is missing. The fixture path runs _validate_computer_use_negative_cases, which feeds eight declared perturbations through the same posture and confirms each one is rejected with a specific error code. The set is fixed in COMPUTER_USE_EXPECTED_NEGATIVE_CASES; a missing observation there fails the run.
\n
The specific actions each row must reject
\n
Rows are rejected for live-account or account secret actions, external mutation, unapproved purchase or send, destructive file actions, hidden-state claims, missing observations, and benchmark-score claims. Live account action is rejected when a row carries live_account_context or live_account_action. account secret entry is rejected when a row attempts it or carries a account secret value. External network mutation is rejected outright. A purchase or send is rejected when it is requested without a human_approval_ref. A destructive file action is rejected without a human_review_ref. A hidden screen-state claim is rejected when a row claims hidden state or omits its visible_state_hash. An action without a prior observation or affordance ref is rejected. A benchmark-score claim is rejected: a synthetic replay record cannot claim benchmark performance.
\n
Prior Art Grounding
\n
WebArena, Mind2Web, OSWorld, WebDriver accounting
\n
The design borrows inspectable action-trace accounting from WebArena, Mind2Web, OSWorld, and WebDriver, requiring observations, verdicts, transitions, recovery, and cold replay to line up before an episode is credited. This component borrows the action-trace accounting pattern from web and desktop agent benchmarks that make action trajectories inspectable. WebArena and Mind2Web anchor realistic web-task evaluation, and OSWorld extends the concern to agents acting in real computer environments. Browser automation standards such as WebDriver are prior art for representing actions against visible browser state through a controlled protocol. The borrowed idea is the accounting: observations, affordances, actions, pre-action verdicts, transitions, recovery, and cold replay must line up before an episode is credited.
A pass means the exported bundle joined cleanly, every action carrying an observation, verdict, and matching transition, with the cold replay reproducing them and the record staying metadata-only. A pass means the exported bundle joined cleanly: every action had an observation, a pre-action verdict, and a matching transition, the cold replay reproduced them, and the written record stayed metadata-only. The command writes the validation record under the named output directory. The public exported bundle has contains_negative_cases: false, so the eight-case floor is checked on the fixture path, and tests/test_agent_route_observability_runtime.py remains the negative-case authority.
\n
Scope boundary
\n
Scope limit
\n
The synthetic replay strongest claim
\n
The strongest claim is a synthetic action-trace replay over public fixtures, joining observations, verdicts, transitions, recovery, cold replay, digest checks, and the eight negative cases into metadata-only records. The strongest claim the evidence supports is a synthetic computer-use action-trace replay over public fixtures: visible observations, affordances, action rows, pre-action authority verdicts, state-transition records, recovery records, cold-replay rows, public trace spans, source-module digest checks, the eight expected negative cases, and metadata-only result records.
\n
Capabilities the ceiling sets false
\n
It controls no live browser, takes no account actions, enters no account secrets, mutates nothing external, and claims no benchmark, with COMPUTER_USE_AUTHORITY_CEILING setting every such capability false. That is the whole proof boundary. This component does not control a live browser or desktop, take account actions, enter account secrets, mutate external systems, execute purchases or sends, perform destructive host actions, export raw screenshots, or read hidden screen state. It does not claim benchmark performance, use external model services, change source files, include launch operations, or prove whole-system correctness. The scope limit is stated in the record itself as COMPUTER_USE_AUTHORITY_CEILING, where every one of those capabilities is set to false. The generated diagram and Atlas entry are navigation surfaces, not additional proof.
\n",
"components": [
{
"anchor": "component-agent_route_observability_runtime",
"component_id": "agent_route_observability_runtime",
"public_label": "Agent Route Observability Runtime"
}
],
"doctrine_feed": {
"counts": {
"axiom": 1,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 2,
"principle": 2
},
"link_count": 8,
"links": [
{
"id": "agent_route_observability_runtime",
"kind": "organ",
"kind_label": "component",
"label": "Agent Route Observability Runtime",
"relation": "explains",
"url": "component-agent-route-observability-runtime.html"
},
{
"id": "mechanism.agent_route_observability_runtime.validates_public_route_feedback",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public route feedback",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-route-observability-runtime-validates-public-route-feedback"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.agent_sandbox_policy_escape_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Sandbox Policy-Escape Replay",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[35:paper_module.agent_sandbox_policy_escape_replay].depends_on",
"url": "paper-module-agent-sandbox-policy-escape-replay.html"
},
{
"id": "paper_module.batch7_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Source Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[69:paper_module.batch7_macro_engines_capsule].depends_on",
"url": "paper-module-batch7-macro-engines-capsule.html"
}
],
"paper_module_id": "paper_module.computer_use_action_trace_replay",
"reader_packet": {
"abstract": "Validator-backed public replay for synthetic computer-use action traces under the route-observability runtime.",
"authority_ceiling": "Synthetic public computer-use action-trace fixtures, exported bundle metadata, copied source-module digests, and metadata-only result records only; no live account action, account secret entry, external network mutation, purchase/send authority, destructive host action, hidden screen-state claim, benchmark-score claim, launch-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.computer_use_action_trace_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-computer-use-action-trace-replay-diagram-1",
"html_url": "docs/paper-module-computer-use-action-trace-replay.html#paper-module-computer-use-action-trace-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD load[\"load bundle, strict JSON\"] load --> proj[\"_validate_computer_use_projection_protocol\"] proj --> pol[\"_validate_computer_use_interaction_policy\"] pol --> ep[\"_validate_computer_use_episodes\"] ep --> obs[\"_validate_computer_use_observations\"] obs --> act[\"_validate_computer_use_actions\"] act --> ver[\"_validate_computer_use_authority_verdicts\"] ver --> tr[\"_validate_computer_use_state_transitions\"] tr --> rec[\"_validate_computer_use_recovery_receipts\"] rec --> cold[\"_validate_computer_use_cold_replay\"] cold --> neg[\"_validate_computer_use_negative_cases fixture path only\"] neg --> out[\"status pass, else blocked\"]",
"source_format": "mermaid",
"source_hash": "sha256:a119d7f4ed9e1982f774216d796f28cf8050e2984d48dd1130fc42bfc7ee23c6",
"source_line_end": 49,
"source_line_start": 35
}
],
"grouped_relation_feeds": [
{
"count": 1,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
}
],
"relation": "abides by"
},
{
"count": 2,
"links": [
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "agent_route_observability_runtime",
"kind": "organ",
"kind_label": "component",
"label": "Agent Route Observability Runtime",
"relation": "explains",
"url": "component-agent-route-observability-runtime.html"
},
{
"id": "mechanism.agent_route_observability_runtime.validates_public_route_feedback",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public route feedback",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-agent-route-observability-runtime-validates-public-route-feedback"
}
],
"relation": "explains"
},
{
"count": 3,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.computer_use_action_trace_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 113,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/computer_use_action_trace_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "WebArena",
"section_id": "webarena-mind2web-osworld-webdriver-accounting",
"url": "https://arxiv.org/abs/2307.13854"
},
{
"kind": "external_reference",
"label": "Mind2Web",
"section_id": "webarena-mind2web-osworld-webdriver-accounting",
"url": "https://arxiv.org/abs/2306.06070"
},
{
"kind": "external_reference",
"label": "OSWorld",
"section_id": "webarena-mind2web-osworld-webdriver-accounting",
"url": "https://arxiv.org/abs/2404.07972"
},
{
"kind": "external_reference",
"label": "WebDriver",
"section_id": "webarena-mind2web-osworld-webdriver-accounting",
"url": "https://www.w3.org/TR/webdriver2/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component answers, offline over a synthetic trace, whether a blocked action was later narrated as success and whether a claimed success has any state evidence behind it. A computer-use agent leaves behind a stream of screenshots, clicks, keystrokes, and \"it worked\" assertions. Reading that stream by hand does not tell you the two things that matter: whether an action that was blocked or sent for review got narrated later as a success, and whether a claimed success has any state evidence behind it. This component answers those two questions on a synthetic trace, offline, before any of the language reaches a reader.",
"level": 4,
"order": 1,
"section_id": "component-answers-offline",
"source_line_end": 6,
"source_line_start": 3,
"title": "Component answers offline"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It takes a recorded local-agent run with its observations, actions, pre-action verdicts, and state changes, checking the trail is self-consistent against authority limits written down up front. It takes a recorded run of a local agent, the route it picked, the observations it logged, the actions it took, the authority verdict stamped before each action, and the state change each action produced, and checks that this recorded trail is well-formed and self-consistent. The record states up front where the agent's authority was supposed to stop, so the limits are written down and checkable rather than taken on faith.",
"level": 4,
"order": 2,
"section_id": "takes-recorded-local-agent",
"source_line_end": 10,
"source_line_start": 7,
"title": "Takes recorded local-agent"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The result is a metadata-only record of counts, refs, hashes, and redaction posture over a recorded trail; it never watches a live agent. The result is a metadata-only record of counts, references, hashes, and redaction posture. It checks a recorded trail. It does not watch a live agent and it does not establish one stayed in bounds.",
"level": 4,
"order": 3,
"section_id": "result-is-body-free",
"source_line_end": 14,
"source_line_start": 11,
"title": "Result is metadata-only"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It guards against a complete-looking trace that hides a lie: a denied action reported as done, a success asserted with no oracle, a screenshot treated as proof. The failure this guards against is a trace that looks complete while hiding a lie. An action can be denied by policy and then reported as done. A success can be asserted with no diff, no oracle, and no before-and-after state. A screenshot can be treated as proof when the pixels prove nothing about what actually changed.",
"level": 4,
"order": 5,
"section_id": "the-complete-looking-trace-that-hides-a-lie",
"source_line_end": 20,
"source_line_start": 17,
"title": "The complete-looking trace that hides a lie"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Every action must line up with a prior observation, a pre-action verdict, and a transition whose outcome matches that verdict, through a typed join rather than a screenshot replay. A loose reading of the log cannot catch these. You need every recorded action to line up, row by row, with a prior visible observation, a pre-action authority verdict, and a state-transition record whose outcome agrees with that verdict. If the verdict said allow, the transition has to show the action ran and an oracle confirmed the new state. If the verdict said block or review, the transition has to show the action did not run and the status reads blocked or review_required. The mechanism is a typed join across\u2026",
"level": 4,
"order": 6,
"section_id": "a-typed-join-of-verdict-and-transition",
"source_line_end": 24,
"source_line_start": 21,
"title": "A typed join of verdict and transition"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Inverting the usual trust, observations enter only as a screenshotdigest and visiblestatehash, with raw pixels and hidden state barred, so weight rests on verdict-transition agreement. The inversion is the point. Most action-trace tooling trusts the screenshot. This component treats the screenshot as the one thing it will not trust: observations enter only as a screenshotdigest and a visiblestatehash, and raw pixel bodies, hidden-state claims, and live-browser state are all required to be absent. The evidence that carries weight is the agreement between verdict and transition.",
"level": 4,
"order": 7,
"section_id": "the-screenshot-is-the-one-distrusted-input",
"source_line_end": 28,
"source_line_start": 25,
"title": "The screenshot is the one distrusted input"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 30,
"source_line_start": 29,
"title": "How it works"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "runcomputeruseactiontracebundle resolves the input, decides fixture versus exported bundle, loads it through strict JSON, then runs a fixed sequence of checks before writing anything. The entry point is runcomputeruseactiontracebundle. It resolves the input directory, decides whether the input is the full fixture with negative cases or the public exported bundle, loads the bundle through a strict JSON path, then runs a fixed sequence of checks in this order before it writes anything.",
"level": 4,
"order": 9,
"section_id": "run-computer-use-action-trace-bundle-and-the-check-sequence",
"source_line_end": 50,
"source_line_start": 31,
"title": "`run_computer_use_action_trace_bundle` and the check sequence"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "Each stage consumes the rows the earlier stage exported, so a break early in the chain propagates downstream. Each stage consumes the rows the earlier stage exported, so a break early in the chain propagates. The relevant functions:",
"level": 4,
"order": 10,
"section_id": "stages-chain-so-early-breaks-propagate",
"source_line_end": 63,
"source_line_start": 51,
"title": "Stages chain, so early breaks propagate"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildpubliccomputerusetrace emits bounded public spans, and status is pass only when findings are empty, no negative case is missing, the scan passes, and source-module digests verify. After the checks, buildpubliccomputerusetrace emits bounded public trace spans over refs, counts, and hashes. runcomputeruseactiontrace_bundle collects the findings from every stage. It sets status to pass only when the positive findings are empty, no expected negative case is missing, the non-public-state scan passes, and the copied public source-module digests verify. Otherwise the status is blocked. The written record carries counts, refs, hashes, and the redaction posture, and omits raw screen bodies, acc\u2026",
"level": 4,
"order": 11,
"section_id": "public-trace-spans-and-the-pass-rule",
"source_line_end": 67,
"source_line_start": 64,
"title": "Public trace spans and the pass rule"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The tracked fixture record reports fixed counts: four episodes, eight actions and verdicts, eight transitions, one recovery, four cold-replay rows, one blocked action, and eight trace spans. The tracked fixture record reports four episodes, six observations, eight actions, eight authority verdicts, eight state-transition records, one recovery record, four cold-replay rows, one blocked action, and eight public trace spans.",
"level": 4,
"order": 12,
"section_id": "what-the-tracked-fixture-record-reports",
"source_line_end": 71,
"source_line_start": 68,
"title": "What the tracked fixture record reports"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 73,
"source_line_start": 72,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatecomputerusenegativecases feeds eight declared perturbations through the same posture and requires each to be rejected with its own code, failing the run if any is missing. The fixture path runs validatecomputerusenegativecases, which feeds eight declared perturbations through the same posture and confirms each one is rejected with a specific error code. The set is fixed in COMPUTERUSEEXPECTEDNEGATIVE_CASES; a missing observation there fails the run.",
"level": 4,
"order": 14,
"section_id": "eight-declared-perturbations-on-the-fixture-path",
"source_line_end": 77,
"source_line_start": 74,
"title": "Eight declared perturbations on the fixture path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Rows are rejected for live-account or account secret actions, external mutation, unapproved purchase or send, destructive file actions, hidden-state claims, missing observations, and benchmark-score claims. Live account action is rejected when a row carries liveaccountcontext or liveaccountaction. account secret entry is rejected when a row attempts it or carries a credentialvalue. External network mutation is rejected outright. A purchase or send is rejected when it is requested without a humanapprovalref. A destructive file action is rejected without a humanreviewref. A hidden screen-state claim is rejected when a row claims hidden state or omits its visiblestate_hash. An action without a\u2026",
"level": 4,
"order": 15,
"section_id": "the-specific-actions-each-row-must-reject",
"source_line_end": 81,
"source_line_start": 78,
"title": "The specific actions each row must reject"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 83,
"source_line_start": 82,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The design borrows inspectable action-trace accounting from WebArena, Mind2Web, OSWorld, and WebDriver, requiring observations, verdicts, transitions, recovery, and cold replay to line up before an episode is credited. This component borrows the action-trace accounting pattern from web and desktop agent benchmarks that make action trajectories inspectable. WebArena and Mind2Web anchor realistic web-task evaluation, and OSWorld extends the concern to agents acting in real computer environments. Browser automation standards such as WebDriver are prior art for representing actions against visible browser state through a controlled protocol. The borrowed idea is the accounting: observations, af\u2026",
"level": 4,
"order": 17,
"section_id": "webarena-mind2web-osworld-webdriver-accounting",
"source_line_end": 87,
"source_line_start": 84,
"title": "WebArena, Mind2Web, OSWorld, WebDriver accounting"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from microcosm-substrate/:",
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 98,
"source_line_start": 88,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means the exported bundle joined cleanly, every action carrying an observation, verdict, and matching transition, with the cold replay reproducing them and the record staying metadata-only. A pass means the exported bundle joined cleanly: every action had an observation, a pre-action verdict, and a matching transition, the cold replay reproduced them, and the written record stayed metadata-only. The command writes the validation record under the named output directory. The public exported bundle has containsnegativecases: false, so the eight-case floor is checked on the fixture path, and tests/testagentrouteobservabilityruntime.py remains the negative-case authority.",
"level": 4,
"order": 19,
"section_id": "what-a-clean-bundle-join-means",
"source_line_end": 102,
"source_line_start": 99,
"title": "What a clean bundle join means"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 104,
"source_line_start": 103,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim is a synthetic action-trace replay over public fixtures, joining observations, verdicts, transitions, recovery, cold replay, digest checks, and the eight negative cases into metadata-only records. The strongest claim the evidence supports is a synthetic computer-use action-trace replay over public fixtures: visible observations, affordances, action rows, pre-action authority verdicts, state-transition records, recovery records, cold-replay rows, public trace spans, source-module digest checks, the eight expected negative cases, and metadata-only result records.",
"level": 4,
"order": 21,
"section_id": "the-synthetic-replay-strongest-claim",
"source_line_end": 108,
"source_line_start": 105,
"title": "The synthetic replay strongest claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It controls no live browser, takes no account actions, enters no account secrets, mutates nothing external, and claims no benchmark, with COMPUTERUSEAUTHORITYCEILING setting every such capability false. That is the whole proof boundary. This component does not control a live browser or desktop, take account actions, enter account secrets, mutate external systems, execute purchases or sends, perform destructive host actions, export raw screenshots, or read hidden screen state. It does not claim benchmark performance, use external model services, change source files, include launch operations, or prove whole-system correctness. The scope limit is stated in the record itself as COMPUTERUSEAUTH\u2026",
"level": 4,
"order": 22,
"section_id": "capabilities-the-ceiling-sets-false",
"source_line_end": 112,
"source_line_start": 109,
"title": "Capabilities the ceiling sets false"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/computer_use_action_trace_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Validator-backed public replay for synthetic computer-use action traces under the route-observability runtime."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "agent_route_observability_runtime.py",
"path": "src/microcosm_core/organs/agent_route_observability_runtime.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for the synthetic computer-use action-trace bundle, visible observation/action binding, scope limit checks, negative cases, and public trace spans.",
"symbols": [
"run_computer_use_action_trace_bundle",
"validate_computer_use_source_manifest",
"_validate_computer_use_projection_protocol",
"_validate_computer_use_negative_cases"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/agent_route_observability_runtime.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/computer_use_action_trace_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/computer_use_action_trace_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Computer-Use Action Trace Replay validates a synthetic computer-use episode through visible observations, affordances, action rows, authority verdicts, state-transition and recovery result records, cold replay rows, public trace spans, source-module manifests, negative cases, and metadata-only result records. It is a reader-facing contract under Agent Route Observability Runtime, not live browser or desktop control.",
"summary_line": "Validator-backed public replay for synthetic computer-use action traces under the route-observability runtime.",
"title": "Computer-Use Action Trace Replay"
},
{
"anchor": "paper-module-concurrency-mission-control",
"authority_ceiling": "Verified concurrency mission-control fixture and source-module import evidence only; no live scheduler, external model access, hosted orchestration, production concurrency-safety proof, source authority, whole-system equivalence, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
Component turns quiet
\n
concurrency_mission_control turns that moment into a checkable verdict.
\n
Copies real coordination
\n
It copies the real coordination builder and bridge fixtures, runs them over synthetic lanes and a work log seed-speed snapshot, and emits metadata-only records of refs, digests, and verdicts. The component imports the real coordination specimen builder and its public provider and work log bridge fixtures as exact copies, runs the copied builder over synthetic multi-agent lanes, and reads a public snapshot of the work log's seed-speed topology. It emits metadata-only result records: refs, digests, counts, and verdicts, with no session bodies, model-output data, or proof text.
\n
Is no live
\n
It is no live scheduler, private runtime, or production safety evidence; it demonstrates fail-closed transaction gating over declared copied bodies and fixtures. It is not a live scheduler, not the private mission-control runtime, not external model access, and outside production claims concurrency-safety evidence. It demonstrates fail-closed transaction gating over declared copied bodies and fixtures.
\n
Purpose
\n
Is acting on a dirty path safe
\n
Given a dirty path and its live claim topology, each case resolves to one named classification and one allowed action, so a lane proceeds, hands off, or waits. The question the component answers is narrow: given a dirty path and the live claim topology around it, is acting on that path safe, and if not, what must happen first. A loose answer like "probably fine" is exactly the failure mode. Each case resolves to a named classification and one allowed action, so a lane can decide to proceed, hand off, or wait.
\n
Real builder and two cautious classifiers
\n
The evidence runs the real specimen builder over synthetic lanes, reading claims, heartbeats, and collisions, with two classifiers that default to the cautious verdict when evidence is thin. The evidence does not come from a re-implemented scheduler. The component runs the same specimen builder the private system uses, over public synthetic lanes, and reads who holds which claim, whether their heartbeat is current, and where path claims collide. Two classifier functions carry the judgment, and both default to the cautious verdict when the evidence is thin. That default is the behaviour the page is about.
\n
How it works
\n
Two entry points share one SPEC
\n
run and run_concurrency_mission_control_bundle both hand a shared SPEC to run_crown_jewel_organ with _evaluate and evaluate_negative_case, the bundle validating the same contract from a shipped copy.run and run_concurrency_mission_control_bundle are the two entry points. Both hand a shared SPEC to run_crown_jewel_organ with _evaluate as the evaluator and evaluate_negative_case as the negative-case checker. The bundle entry sets input_mode to the exported bundle so a reader can validate the same contract from a shipped copy.
\n
_evaluate runs the builder and five gates
\n
_evaluate loads the exercise manifest, runs the copied builder, then five gates, passing only when all six engines pass and every expected engine id is present._evaluate loads concurrency_mission_control_exercise_manifest.json, runs the copied builder, then runs five gates. It passes only when all six engines pass and every expected engine id is present.
\n
Function
Role
_run_original_builder
Copies the public seed root into a temporary workspace and calls build_concurrency_mission_control_specimen from the copied source. Reads the emitted mission_board.json and receipt.json, summarizes case, accept, block, repair, and authority-collapse counts. If the seed root is missing it returns blocked with CONCURRENCY_MISSION_CONTROL_SEED_ROOT_MISSING instead of running.
_failure_matrix_gate
Checks the board against the manifest's expected counts: exact case and accept counts, minimum block, repair, and source-bundle counts, and every required failure class present (owner-path conflict, duplicate run, dependency gap, stale lease, missing result record, supervised-scope gap, missing parent finalizer, misanchored claim).
_bridge_authority_membrane
Checks that required bridge statuses match, that the summed authority-collapse count is zero, that bridge case-count shapes match, and that the manifest's forbidden-claim set covers seven required bars including private_mission_control_runtime and production_concurrency_safety.
_work_ledger_seed_speed_gate
Checks the seed-speed snapshot: schema, non-negative count shape, minimum active sessions and claims, a collision gate (claim_collisions equals expected, per-session collision count zero, no collision rows), heartbeat, session cards, bound source refs, and forbidden private inputs.
classify_generated_surface_claim_lens
Takes a case's drift paths and claim rows and returns one classification with one allowed action.
classify_concurrency_closure_state_lens
Decides whether a unit of work is genuinely closed or only looks closed.
\n
The generated-surface lens and its verdicts
\n
The generated-surface lens intersects drift paths with the generated entry surfaces, returning owned_live, owned_stale, unowned_generated_drift, unrelated_dirty_state, or clean with a matching allowed action. The generated-surface lens intersects a case's drift paths with GENERATED_ENTRY_SURFACES (ORGANS.md, ARCHITECTURE.md, AGENT_ROUTES.md, atlas/agent_task_routes.json). If a live owner holds one of those paths it returns owned_live with the action do_not_patch_from_sibling_lane. Drift on other paths is unrelated_dirty_state, and no drift is clean.
\n
Separating closed from looks-closed
\n
The closure-state lens folds classification, validation, commitability, and residuals to separate genuinely closed work from cases that only look done, defaulting to open_unclassified when evidence is insufficient. The closure-state lens folds that generated classification together with validation state, commitability, and any open residual. It separates closed_and_committed (product closed, validation complete, commit landed) from the cases that only look done: closed_validation_deferred when validation is parked under host pressure with a stated reason, closed_uncommitted_authority when event authority exists but shared append logs are unsafe to stage, false_residual_stale when a residual is left open against a passing generator check, and owned_live_handoff or owned_stale_reentry when a live or stale owner still holds the path. When the closure evidence is insufficient it returns open_unclassified, which is the default, so absent evidence never reads as success.
\nDiagram of the mechanism (7 steps).Diagram source & refs
Source refs
owner live / stale / absent
classify_generated_surface_claim_lens
closed vs looks-closed
classify_concurrency_closure_state_lens
flowchart TD Builder["_run_original_builder copied specimen builder in temp seed root"] Matrix["_failure_matrix_gate counts and failure classes"] Membrane["_bridge_authority_membrane bridges green, authority-collapse zero, forbidden claims blocked"] Seed["_work_ledger_seed_speed_gate heartbeat, counts, collision-free claims"] Surface["classify_generated_surface_claim_lens owner live / stale / absent"] Closure["classify_concurrency_closure_state_lens closed vs looks-closed"] Record["metadata-only result record refs, digests, counts, verdicts"] Builder --> Matrix Builder --> Membrane Matrix --> Record Membrane --> Record Seed --> Record Surface --> Record Closure --> Record
\n
Negative cases
\n
Five cases that must each block
\n
evaluate_negative_case drives five stable cases that must each block with a named code, and any non-block counts as a wrong-kind pass the checker treats as a failure.evaluate_negative_case drives five stable cases, each expected to block with a named code. missing_seed_root runs the builder against an empty temporary root and expects CONCURRENCY_MISSION_CONTROL_SEED_ROOT_MISSING. provider_bridge_missing sets the provider repair bridge status to missing and expects the authority membrane to block with CONCURRENCY_MISSION_CONTROL_PROVIDER_BRIDGE_BLOCKED. authority_collapse_claim sets an authority-collapse count of one and expects CONCURRENCY_MISSION_CONTROL_AUTHORITY_COLLAPSE. private_runtime_claim drops private_mission_control_runtime from the manifest's forbidden claims and expects CONCURRENCY_MISSION_CONTROL_PRIVATE_RUNTIME_OVERCLAIM. work_ledger_seed_speed_collision injects a claim collision into the seed-speed snapshot and expects CONCURRENCY_MISSION_CONTROL_WORK_LEDGER_COLLISION_UNRESOLVED. A case that does not produce its code counts as a pass of the wrong kind, so the checker treats a non-block as a failure.
\n
Prior Art Grounding
\n
Airflow, Kubernetes leases, two-phase commit
\n
The design borrows coordination shapes from Airflow DAGs, Kubernetes leases, two-phase commit, and GitHub Actions concurrency, keeping only the fail-closed synthetic gating shape, not their runtimes. The component borrows established coordination shapes without claiming their runtimes. Apache Airflow DAGs separate task dependency and scheduling from task internals. Kubernetes lease-based leader election is the lease-holder and renewal pattern behind live-versus-stale owner state. IBM Research on two-phase commit is the transaction-consistency pattern under participant failure. GitHub Actions workflow syntax is the declared workflow-concurrency control. This component keeps only the fail-closed synthetic gating shape.
\n
Validation Result record Path
\n
From microcosm-substrate, validate with throwaway result outputs first:
\n
What a pass proves, and drift repair
\n
A pass proves fail-closed coordination evidence over the declared copied bodies and synthetic fixtures, with source-copy digest drift repaired through microcosm_exact_copy_refresh, not here. A pass means the fixture and bundle runs proved fail-closed coordination evidence over the declared copied bodies and synthetic fixtures. Source-copy digest drift is repaired through microcosm_exact_copy_refresh, not here.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported fixture-contract claim
\n
The strongest claim is that the copied builder, bridge, matrix, membrane, and seed-speed fixtures satisfy the declared contract with five named blocks and no non-public body export. The strongest claim the evidence supports: the exact copied specimen builder, provider and work log bridge fixtures, failure-matrix fixture, bridge authority membrane, and work log seed-speed topology fixture together satisfy the declared concurrency mission-control fixture contract, with five negative cases blocking on their named codes and no non-public body export. The proof boundary is those copied bodies, fixtures, source manifests, and metadata-only result records.
\n
What the scope limit excludes
\n
The ceiling proves no private runtime, live scheduling, external model access, or production safety, and the classifiers lower their verdict rather than claim success when evidence is missing. The scope limit excludes everything past that. This component does not establish private mission-control runtime truth, live scheduling, provider or model dispatch, hosted orchestration, production concurrency safety, source-file changes, launch or publishing-scope decision, or whole-system correctness. When evidence, preconditions, provenance, or a refusal reason is missing, the classifiers lower their verdict rather than claim success.
\n",
"components": [
{
"anchor": "component-concurrency_mission_control",
"component_id": "concurrency_mission_control",
"public_label": "Concurrency Mission Control"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 5
},
"link_count": 15,
"links": [
{
"id": "concurrency_mission_control",
"kind": "organ",
"kind_label": "component",
"label": "Concurrency Mission Control",
"relation": "explains",
"url": "component-concurrency-mission-control.html"
},
{
"id": "mechanism.concurrency_mission_control.validates_public_concurrency_mission_control",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public concurrency mission control",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-concurrency-mission-control-validates-public-concurrency-mission-control"
},
{
"id": "P-10",
"kind": "principle",
"kind_label": "principle",
"label": "Do not land effects without compensation",
"relation": "governed by",
"url": "doctrine.html#dcard-p-10"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.work_landing_and_continuity_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "work landing and continuity controls as transaction-scoped coordination evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-work-landing-and-continuity-control-bundle"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.bridge_phase_continuity_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Bridge Phase Continuity Runtime",
"relation": "depends on",
"url": "paper-module-bridge-phase-continuity-runtime.html"
},
{
"id": "paper_module.work_landing_control_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Work Landing Control Spine",
"relation": "depends on",
"url": "paper-module-work-landing-control-spine.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch10_live_source_drift_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 10 Live Source Drift Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[68:paper_module.batch10_live_source_drift_capsule].depends_on",
"url": "paper-module-batch10-live-source-drift-capsule.html"
}
],
"paper_module_id": "paper_module.concurrency_mission_control",
"reader_packet": {
"abstract": "Concurrency Mission Control validates metadata-only coordination result records without becoming a live scheduler or production concurrency proof.",
"authority_ceiling": "Verified concurrency mission-control fixture and source-module import evidence only; no live scheduler, external model access, hosted orchestration, production concurrency-safety proof, source authority, whole-system equivalence, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.concurrency_mission_control.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-concurrency-mission-control-diagram-1",
"html_url": "docs/paper-module-concurrency-mission-control.html#paper-module-concurrency-mission-control-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Builder[\"_run_original_builder copied specimen builder in temp seed root\"] Matrix[\"_failure_matrix_gate counts and failure classes\"] Membrane[\"_bridge_authority_membrane bridges green, authority-collapse zero, forbidden claims blocked\"] Seed[\"_work_ledger_seed_speed_gate heartbeat, counts, collision-free claims\"] Surface[\"classify_generated_surface_claim_lens owner live / stale / absent\"] Closure[\"classify_concurrency_closure_state_lens closed vs looks-closed\"] Record[\"metadata-only result record refs, digests, counts, verdicts\"] Builder --> Matrix Builder --> Membrane Matrix --> Record Membrane --> Record Seed --> Record Surface --> Record Closure --> Record",
"source_format": "mermaid",
"source_hash": "sha256:d26f1cbdc37827a4ef791fd6da252bf01e9fe8059bde94b116189bbe69e664dc",
"source_line_end": 68,
"source_line_start": 52
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.bridge_phase_continuity_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Bridge Phase Continuity Runtime",
"relation": "depends on",
"url": "paper-module-bridge-phase-continuity-runtime.html"
},
{
"id": "paper_module.work_landing_control_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Work Landing Control Spine",
"relation": "depends on",
"url": "paper-module-work-landing-control-spine.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "concurrency_mission_control",
"kind": "organ",
"kind_label": "component",
"label": "Concurrency Mission Control",
"relation": "explains",
"url": "component-concurrency-mission-control.html"
},
{
"id": "mechanism.concurrency_mission_control.validates_public_concurrency_mission_control",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public concurrency mission control",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-concurrency-mission-control-validates-public-concurrency-mission-control"
}
],
"relation": "explains"
},
{
"count": 6,
"links": [
{
"id": "P-10",
"kind": "principle",
"kind_label": "principle",
"label": "Do not land effects without compensation",
"relation": "governed by",
"url": "doctrine.html#dcard-p-10"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "concept.work_landing_and_continuity_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "work landing and continuity controls as transaction-scoped coordination evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-work-landing-and-continuity-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.concurrency_mission_control",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 108,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/concurrency_mission_control.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "DAGs",
"section_id": "airflow-kubernetes-leases-two-phase-commit",
"url": "https://airflow.apache.org/docs/apache-airflow/stable/core-concepts/dags.html"
},
{
"kind": "external_reference",
"label": "lease-based leader election",
"section_id": "airflow-kubernetes-leases-two-phase-commit",
"url": "https://kubernetes.io/docs/concepts/cluster-administration/coordinated-leader-election/"
},
{
"kind": "external_reference",
"label": "two-phase commit",
"section_id": "airflow-kubernetes-leases-two-phase-commit",
"url": "https://research.ibm.com/publications/two-phase-commit-optimizations-and-tradeoffs-in-the-commercial-environment"
},
{
"kind": "external_reference",
"label": "workflow syntax",
"section_id": "airflow-kubernetes-leases-two-phase-commit",
"url": "https://docs.github.com/en/actions/reference/workflows-and-actions/workflow-syntax"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "concurrencymissioncontrol turns that moment into a checkable verdict.",
"level": 4,
"order": 1,
"section_id": "component-turns-quiet",
"source_line_end": 6,
"source_line_start": 3,
"title": "Component turns quiet"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It copies the real coordination builder and bridge fixtures, runs them over synthetic lanes and a work log seed-speed snapshot, and emits metadata-only records of refs, digests, and verdicts. The component imports the real coordination specimen builder and its public provider and work log bridge fixtures as exact copies, runs the copied builder over synthetic multi-agent lanes, and reads a public snapshot of the work log's seed-speed topology. It emits metadata-only result records: refs, digests, counts, and verdicts, with no session bodies, model-output data, or proof text.",
"level": 4,
"order": 2,
"section_id": "copies-real-coordination",
"source_line_end": 10,
"source_line_start": 7,
"title": "Copies real coordination"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It is no live scheduler, private runtime, or production safety evidence; it demonstrates fail-closed transaction gating over declared copied bodies and fixtures. It is not a live scheduler, not the private mission-control runtime, not external model access, and outside production claims concurrency-safety evidence. It demonstrates fail-closed transaction gating over declared copied bodies and fixtures.",
"level": 4,
"order": 3,
"section_id": "is-no-live",
"source_line_end": 14,
"source_line_start": 11,
"title": "Is no live"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Given a dirty path and its live claim topology, each case resolves to one named classification and one allowed action, so a lane proceeds, hands off, or waits. The question the component answers is narrow: given a dirty path and the live claim topology around it, is acting on that path safe, and if not, what must happen first. A loose answer like \"probably fine\" is exactly the failure mode. Each case resolves to a named classification and one allowed action, so a lane can decide to proceed, hand off, or wait.",
"level": 4,
"order": 5,
"section_id": "is-acting-on-a-dirty-path-safe",
"source_line_end": 20,
"source_line_start": 17,
"title": "Is acting on a dirty path safe"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The evidence runs the real specimen builder over synthetic lanes, reading claims, heartbeats, and collisions, with two classifiers that default to the cautious verdict when evidence is thin. The evidence does not come from a re-implemented scheduler. The component runs the same specimen builder the private system uses, over public synthetic lanes, and reads who holds which claim, whether their heartbeat is current, and where path claims collide. Two classifier functions carry the judgment, and both default to the cautious verdict when the evidence is thin. That default is the behaviour the page is about.",
"level": 4,
"order": 6,
"section_id": "real-builder-and-two-cautious-classifiers",
"source_line_end": 24,
"source_line_start": 21,
"title": "Real builder and two cautious classifiers"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run and runconcurrencymissioncontrolbundle both hand a shared SPEC to runcrownjewelorgan with evaluate and evaluatenegativecase, the bundle validating the same contract from a shipped copy. run and runconcurrencymissioncontrolbundle are the two entry points. Both hand a shared SPEC to runcrownjewelorgan with evaluate as the evaluator and evaluatenegativecase as the negative-case checker. The bundle entry sets input_mode to the exported bundle so a reader can validate the same contract from a shipped copy.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-share-one-spec",
"source_line_end": 30,
"source_line_start": 27,
"title": "Two entry points share one `SPEC`"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "evaluate loads the exercise manifest, runs the copied builder, then five gates, passing only when all six engines pass and every expected engine id is present. evaluate loads concurrencymissioncontrolexercise_manifest.json, runs the copied builder, then runs five gates. It passes only when all six engines pass and every expected engine id is present.",
"level": 4,
"order": 9,
"section_id": "evaluate-runs-the-builder-and-five-gates",
"source_line_end": 43,
"source_line_start": 31,
"title": "`_evaluate` runs the builder and five gates"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The generated-surface lens intersects drift paths with the generated entry surfaces, returning ownedlive, ownedstale, unownedgenerateddrift, unrelateddirtystate, or clean with a matching allowed action. The generated-surface lens intersects a case's drift paths with GENERATEDENTRYSURFACES (Components.md, ARCHITECTURE.md, AGENTROUTES.md, atlas/agenttaskroutes.json). If a live owner holds one of those paths it returns ownedlive with the action donotpatchfromsiblinglane. Drift on other paths is unrelateddirtystate, and no drift is clean.",
"level": 4,
"order": 10,
"section_id": "the-generated-surface-lens-and-its-verdicts",
"source_line_end": 47,
"source_line_start": 44,
"title": "The generated-surface lens and its verdicts"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "The closure-state lens folds classification, validation, commitability, and residuals to separate genuinely closed work from cases that only look done, defaulting to openunclassified when evidence is insufficient. The closure-state lens folds that generated classification together with validation state, commitability, and any open residual. It separates closedandcommitted (product closed, validation complete, commit landed) from the cases that only look done: closedvalidationdeferred when validation is parked under host pressure with a stated reason, closeduncommittedauthority when event authority exists but shared append logs are unsafe to stage, falseresidualstale when a residual is left\u2026",
"level": 4,
"order": 11,
"section_id": "separating-closed-from-looks-closed",
"source_line_end": 69,
"source_line_start": 48,
"title": "Separating closed from looks-closed"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 71,
"source_line_start": 70,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluatenegativecase drives five stable cases that must each block with a named code, and any non-block counts as a wrong-kind pass the checker treats as a failure. evaluatenegativecase drives five stable cases, each expected to block with a named code. missingseedroot runs the builder against an empty temporary root and expects CONCURRENCYMISSIONCONTROLSEEDROOTMISSING. providerbridgemissing sets the provider repair bridge status to missing and expects the authority membrane to block with CONCURRENCYMISSIONCONTROLPROVIDERBRIDGEBLOCKED. authoritycollapseclaim sets an authority-collapse count of one and expects CONCURRENCYMISSIONCONTROLAUTHORITYCOLLAPSE. privateruntimeclaim drops privatemissi\u2026",
"level": 4,
"order": 13,
"section_id": "five-cases-that-must-each-block",
"source_line_end": 75,
"source_line_start": 72,
"title": "Five cases that must each block"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "prior-art-grounding",
"source_line_end": 77,
"source_line_start": 76,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The design borrows coordination shapes from Airflow DAGs, Kubernetes leases, two-phase commit, and GitHub Actions concurrency, keeping only the fail-closed synthetic gating shape, not their runtimes. The component borrows established coordination shapes without claiming their runtimes. Apache Airflow DAGs separate task dependency and scheduling from task internals. Kubernetes lease-based leader election is the lease-holder and renewal pattern behind live-versus-stale owner state. IBM Research on two-phase commit is the transaction-consistency pattern under participant failure. GitHub Actions workflow syntax is the declared workflow-concurrency control. This component keeps only the fail-clo\u2026",
"level": 4,
"order": 15,
"section_id": "airflow-kubernetes-leases-two-phase-commit",
"source_line_end": 81,
"source_line_start": 78,
"title": "Airflow, Kubernetes leases, two-phase commit"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate, validate with throwaway result outputs first:",
"level": 2,
"order": 16,
"section_id": "validation-receipt-path",
"source_line_end": 93,
"source_line_start": 82,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass proves fail-closed coordination evidence over the declared copied bodies and synthetic fixtures, with source-copy digest drift repaired through microcosmexactcopyrefresh, not here. A pass means the fixture and bundle runs proved fail-closed coordination evidence over the declared copied bodies and synthetic fixtures. Source-copy digest drift is repaired through microcosmexactcopyrefresh, not here.",
"level": 4,
"order": 17,
"section_id": "what-a-pass-proves-and-drift-repair",
"source_line_end": 97,
"source_line_start": 94,
"title": "What a pass proves, and drift repair"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "claim-ceiling",
"source_line_end": 99,
"source_line_start": 98,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim is that the copied builder, bridge, matrix, membrane, and seed-speed fixtures satisfy the declared contract with five named blocks and no non-public body export. The strongest claim the evidence supports: the exact copied specimen builder, provider and work log bridge fixtures, failure-matrix fixture, bridge authority membrane, and work log seed-speed topology fixture together satisfy the declared concurrency mission-control fixture contract, with five negative cases blocking on their named codes and no non-public body export. The proof boundary is those copied bodies, fixtures, source manifests, and metadata-only result records.",
"level": 4,
"order": 19,
"section_id": "the-strongest-supported-fixture-contract-claim",
"source_line_end": 103,
"source_line_start": 100,
"title": "The strongest supported fixture-contract claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The ceiling proves no private runtime, live scheduling, external model access, or production safety, and the classifiers lower their verdict rather than claim success when evidence is missing. The scope limit excludes everything past that. This component does not establish private mission-control runtime truth, live scheduling, provider or model dispatch, hosted orchestration, production concurrency safety, source-file changes, launch or public sharing decision, or whole-system correctness. When evidence, preconditions, provenance, or a refusal reason is missing, the classifiers lower their verdict rather than claim success.",
"level": 4,
"order": 20,
"section_id": "what-the-authority-ceiling-excludes",
"source_line_end": 107,
"source_line_start": 104,
"title": "What the scope limit excludes"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/concurrency_mission_control.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Concurrency Mission Control validates metadata-only coordination result records without becoming a live scheduler or production concurrency proof."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "concurrency_mission_control.py",
"path": "src/microcosm_core/organs/concurrency_mission_control.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for concurrency mission-control fixture validation, work log topology classification, failure-class gates, negative cases, and scope limits.",
"symbols": [
"run",
"run_concurrency_mission_control_bundle",
"_evaluate",
"classify_generated_surface_claim_lens",
"classify_concurrency_closure_state_lens",
"result_card",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/concurrency_mission_control.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/concurrency_mission_control.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/concurrency_mission_control.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Concurrency Mission Control validates the public concurrency mission-control membrane: copied source-builder digests, bridge artifacts, failure classes, work log seed-speed topology, heartbeat and claim-collision anchors, negative cases, source-open body imports, and metadata-only result records. It separates coordination evidence from hosted orchestration, external model access, live scheduling, whole-system equivalence, source authority, launch, public sharing, and production concurrency guarantees.",
"summary_line": "Concurrency Mission Control validates metadata-only coordination result records without becoming a live scheduler or production concurrency proof.",
"title": "Concurrency Mission Control"
},
{
"anchor": "paper-module-doctrine-fact-claim-audit",
"authority_ceiling": "Declared fact-assertion, code-locus, and DAG fixture truth gate only; no comprehension engine, no minimum-read-graph proof, no doctrine saturation claim, no source-file changes, no launch-scope decision, no publishing-scope decision, and no whole-system correctness.",
"body_html": "
Documentation about live
\n
Documentation about a live system rots: counts and code citations go stale as the code moves, and a reader cannot tell a current number from one true only once. Documentation about a live system rots. A page says there are forty-seven of something, or cites a function in a named file, and both claims go stale as the code moves underneath them. A reader cannot tell a current count from a number that was true once and never rechecked.
\n
Doctrine fact claim
\n
doctrine_fact_claim_audit asks which factual claims on a page can be re-derived from source now and which have drifted, emitting a metadata-only record of counts, verdicts, and scope limits.doctrine_fact_claim_audit answers one question about a page of factual claims: which of them can be re-derived from source right now, and which have become untracked drift. It reads a small set of public fixtures, recomputes a bounded relation over them, and writes a result record that carries counts, verdicts, and scope limits but no copied source bodies.
\n
Does not judge
\n
It does not judge whether prose sounds true; it checks four concrete things: fact count, code-locus existence with anchor, route-graph edges pointing at known facts, and volatile numbers bound to assertions. The component does not read prose and judge whether it sounds true. It checks four concrete things: that a declared fact count matches the fact table, that every cited code location exists on disk with its quoted anchor present, that every route-graph edge points at a known fact, and that volatile numbers in current-state prose are bound to a fact assertion or surfaced as a blocking finding.
\n
Purpose
\n
A claim is a cache with no invalidation
\n
A loose documentation claim is a cached value with no invalidation strategy: a bare number tells you nothing about how to recheck it, so stale claims persist unnoticed as code changes. A loose documentation claim is a cached value with no invalidation strategy. The number 47 sitting in a sentence tells you nothing about how to recheck it, or whether the file it references still exists. When the code changes, the sentence keeps asserting the old value and nobody notices.
\n
Admissible only when bound to something recomputable
\n
A claim is admissible only when bound to something recomputable: a bare number must tie to a fact assertion or get flagged, and cited paths are resolved on disk with anchors confirmed. This component makes a claim admissible only when it is bound to something recomputable. A bare number in a current-state section is not enough. It has to be tied to a fact assertion that records how the value is derived, or it gets flagged. Cited file paths are resolved on disk and their anchor text is confirmed present, so a plausible-but-dead reference becomes a typed finding instead of inert prose.
\n
How it works
\n
A fixture evaluator on the crown-jewel runner
\n
The component is a public fixture evaluator whose SPEC declares four required inputs and source anchors, while the shared runner handles validation, record writing, and negative cases, and this module supplies the evaluator and mutator. The component is a public fixture evaluator built on the shared crown-jewel runner. SPEC is a CrownJewelSpec that declares four required inputs, fact_assertions.json, fact_dag.json, numeric_claims.json, and projection_protocol.json, plus the source manifest reference and the anchors each copied source body must contain. The runner handles manifest validation, result-record writing, negative-case execution, and scope-limit attachment. This module supplies the domain evaluator and the negative-case mutator.
\n
evaluate checks count, loci, anchors, edges
\n
evaluate compares the declared fact count to the table, resolves each fact's code locus and anchor on disk, and rejects graph edges whose endpoints are not known facts, each with a typed error.evaluate is the core pass. It loads the fact table and the route graph, then runs the numeric-claim check, then walks the fact rows. It compares expected_fact_count against the number of fact rows and records DOCTRINE_FACT_COUNT_MISMATCH on disagreement. For each fact it requires at least one code locus, resolves the relative path through _resolve_code_locus against the copied source bundle, reads the file, and confirms the declared anchor string is in the body. A missing file or a missing anchor becomes DOCTRINE_CODE_LOCUS_MISSING or DOCTRINE_CODE_LOCUS_ANCHOR_MISSING. It then builds the set of known fact ids and rejects any graph edge whose from or to endpoint is not in that set with DOCTRINE_FACT_DAG_DEAD_REF.
\n
_evaluate_numeric_claims binds volatile numbers
\n
_evaluate_numeric_claims imports the copied derived_fact_hologram.py and runs find_unbound_numeric_claims per case, treating a detector case as positive and blocking any unbound current-state number._evaluate_numeric_claims handles the volatile-number check. It loads numeric_claims.json and imports the copied derived_fact_hologram.py body through _load_derived_fact_module, which resolves the module out of the exported bundle and reports DOCTRINE_NUMERIC_CLAIM_SOURCE_MISSING if the copied body is absent. It then calls that body's find_unbound_numeric_claims against each case. A detector case declares expect_unbound: true and must surface the stated section and number, so it is positive evidence. Any other case that leaves current-state numeric prose without a matching fact-assertion section is blocked with DOCTRINE_UNBOUND_NUMERIC_CLAIM and DOCTRINE_VOLATILE_NUMERIC_UNBOUND.
\n
Pass requires an empty findings list
\n
evaluate returns pass only when the findings list is empty, reporting the fact, code-locus, edge, and numeric-claim tallies in a metadata-only record that never embeds copied source.evaluate returns a status of pass only when the findings list is empty, along with the fact count, verified code-locus count, edge count, and the numeric-claim tallies. The status record keeps body_in_receipt: false and never embeds copied source text.
\n
Negative cases mutate a fixture in memory
\n
evaluate_negative_case applies one real defect to the positive fixture in memory and reruns evaluate, so the proof of each error code stays attached to the evaluator, not to fixture filenames. The negative floor is semantic, not label-trusting. evaluate_negative_case copies the positive fixture in memory, applies one real defect for the named case (wrong_fact_count bumps the expected count, missing_code_locus empties a fact's loci, dead_code_locus corrupts an anchor, dead_dag_ref repoints an edge, unbound_numeric_claim strips a case's asserted sections), writes the mutated payloads to a temporary directory, and reruns evaluate there. The expected error codes live in EXPECTED_NEGATIVE_CASES. Because the mutation happens after load, the proof stays attached to the evaluator rather than to fixture filenames.
\n
Function
Role
run, run_doctrine_fact_bundle
CLI entrypoints; forward to the shared crown-jewel runner in fixture or exported-bundle mode
evaluate
Loads the fixtures, checks fact count, code loci, anchors, and graph edges, returns the status record
_evaluate_numeric_claims
Imports the copied source body and runs the volatile-number binding check per case
_load_derived_fact_module
Resolves and imports derived_fact_hologram.py from the copied bundle
evaluate_negative_case
Mutates the positive fixture in memory and reruns evaluate to prove each stable error code
\nDiagram of the mechanism (10 steps).Diagram source & refs
Source refs
facts + expected_fact_count
fact_assertions.json
edges
fact_dag.json
cases
numeric_claims.json
flowchart TD Facts["fact_assertions.json facts + expected_fact_count"] --> Eval["evaluate"] Dag["fact_dag.json edges"] --> Eval Numerics["numeric_claims.json cases"] --> Eval Eval --> Count{"count matches table length?"} Eval --> Loci{"path on disk + anchor in body?"} Eval --> DagRef{"edge endpoints are known facts?"} Eval --> Bound{"current-state numbers bound to a fact?"} Count -->|no| Block["typed blocking finding"] Loci -->|no| Block DagRef -->|no| Block Bound -->|no| Block Count -->|yes| Pass["metadata-only result record"] Loci -->|yes| Pass DagRef -->|yes| Pass Bound -->|yes| Pass
\n
Negative cases
\n
Five failure cases, five stable error codes
\n
Five declared failure cases each derive a stable error code, while the positive run reports three facts, three code loci, two edges, and two numeric-claim cases including one detector. The fixture ships five declared failure cases and the component derives a stable error code for each. wrong_fact_count gives DOCTRINE_FACT_COUNT_MISMATCH. missing_code_locus gives DOCTRINE_CODE_LOCUS_MISSING. dead_code_locus gives DOCTRINE_CODE_LOCUS_ANCHOR_MISSING. dead_dag_ref gives DOCTRINE_FACT_DAG_DEAD_REF. unbound_numeric_claim gives DOCTRINE_UNBOUND_NUMERIC_CLAIM together with DOCTRINE_VOLATILE_NUMERIC_UNBOUND. The accepted positive run reports three facts, three verified code loci, two graph edges, and two numeric-claim cases of which one is the detector for the number 123 in a current-state section.
\n
Prior Art Grounding
\n
Provenance and fact-check metadata prior art
\n
The design borrows from provenance modeling and structured fact-check metadata, keeping only the parts that apply to fixture fact counts, code-locus existence, anchors, graph references, and numeric binding. This component borrows shapes from provenance modeling and structured fact-check metadata. W3C PROV models entities, activities, and agents so a reader can assess how derived information was produced. Schema.org ClaimReview is a web pattern for recording a reviewed claim and its checking context. JSON Schema covers rejecting malformed or incomplete claim records, and SLSA provenance covers tracing an artifact back to its source. The component keeps only the parts that apply to public fixture fact counts, code-locus existence, anchor presence, graph references, and synthetic numeric binding cases.
\n
Validation Result record Path
\n
From microcosm-substrate, run against external output paths so the reader check does not churn tracked result records. From microcosm-substrate, run against external output paths so the reader check does not churn tracked result records:
\n
What a pass confirms
\n
A pass means the fact count, code loci, anchors, graph edges, numeric bindings, manifest digests, and five negative cases all agreed with the copied public basis. A pass means the fixture fact count, code loci, anchors, graph edges, numeric-claim bindings, manifest digests, and the five negative cases all agreed with the copied public basis. If the copied source bodies drift, refresh the exported bundle through its owning import lane before treating a bundle failure as a reader-page defect.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The strongest claim is that, over these fixtures, the component verifies fact count, code-locus existence with anchors, graph edges, and bound volatile numbers, with negative cases proving each rejection derives from the evaluator. The strongest claim the evidence supports: given these public fixtures, the component checks that a declared fact count matches its table, that every cited code location exists on disk with its anchor present, that every route-graph edge points at a known fact, and that volatile current-state numbers are either bound to a fact assertion or surfaced as a blocking finding. The five negative cases prove each rejection path derives a stable error code from the evaluator, not from a fixture label. Result records stay metadata-only.
\n
What it refuses to claim
\n
It is not a comprehension engine and judges no prose as true: it exports no private doctrine, mutates no source, promotes no axiom, and excludes launch, bounded to the public fixtures and copied bodies. What it refuses to claim: it is not a comprehension engine and does not judge whether prose reads as true. It does not establish a minimum read graph for the wider doctrine plane, does not export private doctrine, and excludes launch or hosted-public posture. The proof boundary is the synthetic public fixtures, the copied source bodies, their manifest digests and anchors, the metadata-only result records, and the named negative cases. The scope limit excludes source-file changes, candidate-axiom promotion, and any claim that these counts prove whole-system correctness.
\n",
"components": [
{
"anchor": "component-doctrine_fact_claim_audit",
"component_id": "doctrine_fact_claim_audit",
"public_label": "Doctrine Fact Claim Audit"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 14,
"links": [
{
"id": "doctrine_fact_claim_audit",
"kind": "organ",
"kind_label": "component",
"label": "Doctrine Fact Claim Audit",
"relation": "explains",
"url": "component-doctrine-fact-claim-audit.html"
},
{
"id": "mechanism.doctrine_fact_claim_audit.validates_public_doctrine_fact_claim_audit",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public doctrine fact claim audit",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-doctrine-fact-claim-audit-validates-public-doctrine-fact-claim-audit"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-7",
"kind": "principle",
"kind_label": "principle",
"label": "Track known unknowns without claiming the unknown is mapped",
"relation": "governed by",
"url": "doctrine.html#dcard-p-7"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.executable_doctrine_grammar",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Executable Doctrine Grammar",
"relation": "depends on",
"url": "paper-module-executable-doctrine-grammar.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.engine_room_derived_fact_provider_engine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Derived Fact Provider Engine",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[85:paper_module.engine_room_derived_fact_provider_engine].depends_on",
"url": "paper-module-engine-room-derived-fact-provider-engine.html"
},
{
"id": "paper_module.microcosm_axiom_substrate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Microcosm Axiom System",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[92:paper_module.microcosm_axiom_substrate].depends_on",
"url": "paper-module-microcosm-axiom-substrate.html"
},
{
"id": "paper_module.self_ignorance_coverage_ledger",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Self-Ignorance Coverage Ledger",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[49:paper_module.self_ignorance_coverage_ledger].depends_on",
"url": "paper-module-self-ignorance-coverage-ledger.html"
}
],
"paper_module_id": "paper_module.doctrine_fact_claim_audit",
"reader_packet": {
"abstract": "Doctrine Fact Claim Audit rejects wrong fact counts and dead anchors without claiming comprehension or route completeness.",
"authority_ceiling": "Declared fact-assertion, code-locus, and DAG fixture truth gate only; no comprehension engine, no minimum-read-graph proof, no doctrine saturation claim, no source-file changes, no launch-scope decision, no publishing-scope decision, and no whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.doctrine_fact_claim_audit.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-doctrine-fact-claim-audit-diagram-1",
"html_url": "docs/paper-module-doctrine-fact-claim-audit.html#paper-module-doctrine-fact-claim-audit-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Facts[\"fact_assertions.json facts + expected_fact_count\"] --> Eval[\"evaluate\"] Dag[\"fact_dag.json edges\"] --> Eval Numerics[\"numeric_claims.json cases\"] --> Eval Eval --> Count{\"count matches table length?\"} Eval --> Loci{\"path on disk + anchor in body?\"} Eval --> DagRef{\"edge endpoints are known facts?\"} Eval --> Bound{\"current-state numbers bound to a fact?\"} Count -->|no| Block[\"typed blocking finding\"] Loci -->|no| Block DagRef -->|no| Block Bound -->|no| Block Count -->|yes| Pass[\"metadata-only result record\"] Loci -->|yes| Pass DagRef -->|yes| Pass Bound -->|yes| Pass",
"source_format": "mermaid",
"source_hash": "sha256:c0f821a73c80cbd873a322d7f8522749a67188caec5dc60a28a29361160b14be",
"source_line_end": 82,
"source_line_start": 65
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.executable_doctrine_grammar",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Executable Doctrine Grammar",
"relation": "depends on",
"url": "paper-module-executable-doctrine-grammar.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "doctrine_fact_claim_audit",
"kind": "organ",
"kind_label": "component",
"label": "Doctrine Fact Claim Audit",
"relation": "explains",
"url": "component-doctrine-fact-claim-audit.html"
},
{
"id": "mechanism.doctrine_fact_claim_audit.validates_public_doctrine_fact_claim_audit",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public doctrine fact claim audit",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-doctrine-fact-claim-audit-validates-public-doctrine-fact-claim-audit"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-7",
"kind": "principle",
"kind_label": "principle",
"label": "Track known unknowns without claiming the unknown is mapped",
"relation": "governed by",
"url": "doctrine.html#dcard-p-7"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.doctrine_fact_claim_audit",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 130,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/doctrine_fact_claim_audit.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "PROV",
"section_id": "provenance-and-fact-check-metadata-prior-art",
"url": "https://www.w3.org/TR/prov-overview/"
},
{
"kind": "external_reference",
"label": "ClaimReview",
"section_id": "provenance-and-fact-check-metadata-prior-art",
"url": "https://schema.org/ClaimReview"
},
{
"kind": "external_reference",
"label": "JSON Schema",
"section_id": "provenance-and-fact-check-metadata-prior-art",
"url": "https://json-schema.org/"
},
{
"kind": "external_reference",
"label": "SLSA provenance",
"section_id": "provenance-and-fact-check-metadata-prior-art",
"url": "https://slsa.dev/spec/v1.2/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "Documentation about a live system rots: counts and code citations go stale as the code moves, and a reader cannot tell a current number from one true only once. Documentation about a live system rots. A page says there are forty-seven of something, or cites a function in a named file, and both claims go stale as the code moves underneath them. A reader cannot tell a current count from a number that was true once and never rechecked.",
"level": 4,
"order": 1,
"section_id": "documentation-about-live",
"source_line_end": 7,
"source_line_start": 3,
"title": "Documentation about live"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "doctrinefactclaimaudit asks which factual claims on a page can be re-derived from source now and which have drifted, emitting a metadata-only record of counts, verdicts, and scope limits.** doctrinefactclaim_audit answers one question about a page of factual claims: which of them can be re-derived from source right now, and which have become untracked drift. It reads a small set of public fixtures, recomputes a bounded relation over them, and writes a result record that carries counts, verdicts, and scope limits but no copied source bodies.",
"level": 4,
"order": 2,
"section_id": "doctrine-fact-claim",
"source_line_end": 12,
"source_line_start": 8,
"title": "Doctrine fact claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It does not judge whether prose sounds true; it checks four concrete things: fact count, code-locus existence with anchor, route-graph edges pointing at known facts, and volatile numbers bound to assertions. The component does not read prose and judge whether it sounds true. It checks four concrete things: that a declared fact count matches the fact table, that every cited code location exists on disk with its quoted anchor present, that every route-graph edge points at a known fact, and that volatile numbers in current-state prose are bound to a fact assertion or surfaced as a blocking finding.",
"level": 4,
"order": 3,
"section_id": "does-not-judge",
"source_line_end": 17,
"source_line_start": 13,
"title": "Does not judge"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 19,
"source_line_start": 18,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A loose documentation claim is a cached value with no invalidation strategy: a bare number tells you nothing about how to recheck it, so stale claims persist unnoticed as code changes. A loose documentation claim is a cached value with no invalidation strategy. The number 47 sitting in a sentence tells you nothing about how to recheck it, or whether the file it references still exists. When the code changes, the sentence keeps asserting the old value and nobody notices.",
"level": 4,
"order": 5,
"section_id": "a-claim-is-a-cache-with-no-invalidation",
"source_line_end": 24,
"source_line_start": 20,
"title": "A claim is a cache with no invalidation"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A claim is admissible only when bound to something recomputable: a bare number must tie to a fact assertion or get flagged, and cited paths are resolved on disk with anchors confirmed. This component makes a claim admissible only when it is bound to something recomputable. A bare number in a current-state section is not enough. It has to be tied to a fact assertion that records how the value is derived, or it gets flagged. Cited file paths are resolved on disk and their anchor text is confirmed present, so a plausible-but-dead reference becomes a typed finding instead of inert prose.",
"level": 4,
"order": 6,
"section_id": "admissible-only-when-bound-to-something-recomputable",
"source_line_end": 29,
"source_line_start": 25,
"title": "Admissible only when bound to something recomputable"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 31,
"source_line_start": 30,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component is a public fixture evaluator whose SPEC declares four required inputs and source anchors, while the shared runner handles validation, record writing, and negative cases, and this module supplies the evaluator and mutator. The component is a public fixture evaluator built on the shared crown-jewel runner. SPEC is a CrownJewelSpec that declares four required inputs, factassertions.json, factdag.json, numericclaims.json, and projectionprotocol.json, plus the source manifest reference and the anchors each copied source body must contain. The runner handles manifest validation, result-record writing, negative-case execution, and scope-limit attachment. This module supplies the dom\u2026",
"level": 4,
"order": 8,
"section_id": "a-fixture-evaluator-on-the-crown-jewel-runner",
"source_line_end": 36,
"source_line_start": 32,
"title": "A fixture evaluator on the crown-jewel runner"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluate compares the declared fact count to the table, resolves each fact's code locus and anchor on disk, and rejects graph edges whose endpoints are not known facts, each with a typed error. evaluate is the core pass. It loads the fact table and the route graph, then runs the numeric-claim check, then walks the fact rows. It compares expectedfactcount against the number of fact rows and records DOCTRINEFACTCOUNTMISMATCH on disagreement. For each fact it requires at least one code locus, resolves the relative path through resolvecodelocus against the copied source bundle, reads the file, and confirms the declared anchor string is in the body. A missing file or a missing anchor becomes DOC\u2026",
"level": 4,
"order": 9,
"section_id": "evaluate-checks-count-loci-anchors-edges",
"source_line_end": 41,
"source_line_start": 37,
"title": "`evaluate` checks count, loci, anchors, edges"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluatenumericclaims imports the copied derivedfacthologram.py and runs findunboundnumericclaims per case, treating a detector case as positive and blocking any unbound current-state number.** evaluatenumericclaims handles the volatile-number check. It loads numericclaims.json and imports the copied derivedfacthologram.py body through loadderivedfactmodule, which resolves the module out of the exported bundle and reports DOCTRINENUMERICCLAIMSOURCEMISSING if the copied body is absent. It then calls that body's findunboundnumericclaims against each case. A detector case declares expectunbound: true and must surface the stated section and number, so it is positive evidence. Any other case tha\u2026",
"level": 4,
"order": 10,
"section_id": "evaluate-numeric-claims-binds-volatile-numbers",
"source_line_end": 46,
"source_line_start": 42,
"title": "`_evaluate_numeric_claims` binds volatile numbers"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluate returns pass only when the findings list is empty, reporting the fact, code-locus, edge, and numeric-claim tallies in a metadata-only record that never embeds copied source. evaluate returns a status of pass only when the findings list is empty, along with the fact count, verified code-locus count, edge count, and the numeric-claim tallies. The status record keeps bodyinreceipt: false and never embeds copied source text.",
"level": 4,
"order": 11,
"section_id": "pass-requires-an-empty-findings-list",
"source_line_end": 51,
"source_line_start": 47,
"title": "Pass requires an empty findings list"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "evaluatenegativecase applies one real defect to the positive fixture in memory and reruns evaluate, so the proof of each error code stays attached to the evaluator, not to fixture filenames. The negative floor is semantic, not label-trusting. evaluatenegativecase copies the positive fixture in memory, applies one real defect for the named case (wrongfactcount bumps the expected count, missingcodelocus empties a fact's loci, deadcodelocus corrupts an anchor, deaddagref repoints an edge, unboundnumericclaim strips a case's asserted sections), writes the mutated payloads to a temporary directory, and reruns evaluate there. The expected error codes live in EXPECTEDNEGATIVECASES. Because the mut\u2026",
"level": 4,
"order": 12,
"section_id": "negative-cases-mutate-a-fixture-in-memory",
"source_line_end": 83,
"source_line_start": 52,
"title": "Negative cases mutate a fixture in memory"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 85,
"source_line_start": 84,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Five declared failure cases each derive a stable error code, while the positive run reports three facts, three code loci, two edges, and two numeric-claim cases including one detector. The fixture ships five declared failure cases and the component derives a stable error code for each. wrongfactcount gives DOCTRINEFACTCOUNTMISMATCH. missingcodelocus gives DOCTRINECODELOCUSMISSING. deadcodelocus gives DOCTRINECODELOCUSANCHORMISSING. deaddagref gives DOCTRINEFACTDAGDEADREF. unboundnumericclaim gives DOCTRINEUNBOUNDNUMERICCLAIM together with DOCTRINEVOLATILENUMERICUNBOUND. The accepted positive run reports three facts, three verified code loci, two graph edges, and two numeric-claim cases of w\u2026",
"level": 4,
"order": 14,
"section_id": "five-failure-cases-five-stable-error-codes",
"source_line_end": 90,
"source_line_start": 86,
"title": "Five failure cases, five stable error codes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 92,
"source_line_start": 91,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The design borrows from provenance modeling and structured fact-check metadata, keeping only the parts that apply to fixture fact counts, code-locus existence, anchors, graph references, and numeric binding. This component borrows shapes from provenance modeling and structured fact-check metadata. W3C PROV models entities, activities, and agents so a reader can assess how derived information was produced. Schema.org ClaimReview is a web pattern for recording a reviewed claim and its checking context. JSON Schema covers rejecting malformed or incomplete claim records, and SLSA provenance covers tracing an artifact back to its source. The component keeps only the parts that apply to public fi\u2026",
"level": 4,
"order": 16,
"section_id": "provenance-and-fact-check-metadata-prior-art",
"source_line_end": 97,
"source_line_start": 93,
"title": "Provenance and fact-check metadata prior art"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 99,
"source_line_start": 98,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "**From microcosm-substrate, run against external output paths so the reader check does not churn tracked result records.** From microcosm-substrate, run against external output paths so the reader check does not churn tracked result records:",
"level": 4,
"order": 18,
"section_id": "from-microcosm-substrate-run",
"source_line_end": 112,
"source_line_start": 100,
"title": "From microcosm-substrate run"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means the fact count, code loci, anchors, graph edges, numeric bindings, manifest digests, and five negative cases all agreed with the copied public basis. A pass means the fixture fact count, code loci, anchors, graph edges, numeric-claim bindings, manifest digests, and the five negative cases all agreed with the copied public basis. If the copied source bodies drift, refresh the exported bundle through its owning import lane before treating a bundle failure as a reader-page defect.",
"level": 4,
"order": 19,
"section_id": "what-a-pass-confirms",
"source_line_end": 117,
"source_line_start": 113,
"title": "What a pass confirms"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 119,
"source_line_start": 118,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim is that, over these fixtures, the component verifies fact count, code-locus existence with anchors, graph edges, and bound volatile numbers, with negative cases proving each rejection derives from the evaluator. The strongest claim the evidence supports: given these public fixtures, the component checks that a declared fact count matches its table, that every cited code location exists on disk with its anchor present, that every route-graph edge points at a known fact, and that volatile current-state numbers are either bound to a fact assertion or surfaced as a blocking finding. The five negative cases prove each rejection path derives a stable error code from the evalua\u2026",
"level": 4,
"order": 21,
"section_id": "the-strongest-supported-claim",
"source_line_end": 124,
"source_line_start": 120,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It is not a comprehension engine and judges no prose as true: it exports no private doctrine, mutates no source, promotes no axiom, and excludes launch, bounded to the public fixtures and copied bodies. What it refuses to claim: it is not a comprehension engine and does not judge whether prose reads as true. It does not establish a minimum read graph for the wider doctrine plane, does not export private doctrine, and excludes launch or hosted-public posture. The proof boundary is the synthetic public fixtures, the copied source bodies, their manifest digests and anchors, the metadata-only result records, and the named negative cases. The scope limit excludes source-file changes, candidate-a\u2026",
"level": 4,
"order": 22,
"section_id": "what-it-refuses-to-claim",
"source_line_end": 129,
"source_line_start": 125,
"title": "What it refuses to claim"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/doctrine_fact_claim_audit.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Doctrine Fact Claim Audit rejects wrong fact counts and dead anchors without claiming comprehension or route completeness."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "doctrine_fact_claim_audit.py",
"path": "src/microcosm_core/organs/doctrine_fact_claim_audit.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for doctrine fact assertion checks, numeric claim evaluation, code-locus anchor validation, negative cases, and scope limits.",
"symbols": [
"run",
"run_doctrine_fact_bundle",
"evaluate",
"_evaluate_numeric_claims",
"_load_derived_fact_module",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/doctrine_fact_claim_audit.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/doctrine_fact_claim_audit.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/doctrine_fact_claim_audit.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Doctrine Fact Claim Audit validates public doctrine fact assertions against declared sections, numeric claim gates, code-locus anchors, route DAG fixtures, negative cases, source-open body imports, and scope limits. It lowers claim strength to fixture truth: fact assertion, code-loci, and DAG evidence are checked, but the component does not become a comprehension engine, minimum-read graph, doctrine saturation proof, source-file changes lane, or launch-scope decision.",
"summary_line": "Doctrine Fact Claim Audit rejects wrong fact counts and dead anchors without claiming comprehension or route completeness.",
"title": "Doctrine Fact Claim Audit"
},
{
"anchor": "paper-module-self-ignorance-coverage-ledger",
"authority_ceiling": "Known Kind Atlas coverage-debt projection only; no literal unknown-unknown omniscience, no absence proof, no total repository search proof, no source-file changes, no launch-scope decision, no publishing-scope decision, and no whole-system correctness.",
"body_html": "
Known unknowns in the current surface
\n
It answers one question: across declared Kind Atlas families, how many option-surface rows has the generated System Atlas not yet materialised? A navigation system that lists what it knows is easy to build. One that can state precisely what it has not yet covered is harder, and it is the more honest signal to a cold reader. This component answers a single question: for a declared set of Kind Atlas families, how many rows does the option surface expose that the generated System Atlas has not yet materialised?
\n
Computed unknown-coverage debt vector
\n
The answer is a computed debt vector: for each kind it subtracts materialised entities from the live row count and reports the difference as known coverage debt. The answer is a small debt vector, computed rather than asserted. For each selected kind the component recomputes the live Kind Atlas row count, counts the entities the System Atlas graph has actually materialised, and reports the difference as known coverage debt. It emits a metadata-only result record with the counts, the source references, and an explicit scope limit.
\n
Refusing false completeness
\n
Its unusual part is refusal: it rejects any claim that unknown-unknowns are exhaustive, treating known debt as a bounded count of visible gaps, never a completeness proof. The unusual part is what it refuses. It will not accept input that claims its unknown-unknowns are exhaustive, and it will not treat the absence of a row as proof that nothing is missing. Known debt is a bounded count of gaps it can see. It is never a completeness proof.
\n
Purpose
\n
The narrow middle between vague and overconfident
\n
Systems overstate coverage either vaguely or overconfidently, so this component computes the finite gap between what the option surface exposes and what the graph materialised. Navigation systems overstate themselves in two opposite ways. A vague "coverage is incomplete" tells a cold reader nothing operational. A confident "nothing else is missing" is worse, because it converts absence of evidence into evidence of absence. This component occupies the narrow middle. For a declared finite domain of Kind Atlas families it computes the gap between what the option surface exposes and what the System Atlas graph has materialised, and it reports that gap as a number with its basis attached.
\n
The negative output the checker must reject
\n
Beside the bounded debt vector, the checker must reject exhaustive-coverage claims, hand-authored counts, substituted ids, stale baked ids, a tampered builder marker, or an understated debt floor. The positive output is a bounded debt vector. The negative output matters as much: the checker must reject input that claims exhaustive unknown-unknown coverage, hand-authored materialisation counts, substituted entity ids, stale baked ids used as authority, a tampered graph builder marker, or a coverage debt that understates the declared floor.
\n
How it works
\n
Entrypoints and the evaluate core
\n
run and run_self_ignorance_bundle are the entrypoints, evaluate is the core over four bundle files, and evaluate_negative_case replays the semantic negatives. The runtime lives in src/microcosm_core/organs/self_ignorance_coverage_ledger.py. The command entrypoint is run; the exported-bundle entrypoint is run_self_ignorance_bundle; the core evaluator is evaluate; the semantic negative-case replayer is evaluate_negative_case. evaluate consumes four bundle files and runs the following order.
\n
Binding the coverage scope and check result record
\n
_projection_protocol_receipt checks the protocol against the expected scope and System Atlas check command, raising a mismatch when either the scope string or the result record status is wrong. First it loads kind_atlas_rows.json, materialized_entities.json, and projection_protocol.json. _projection_protocol_receipt checks the protocol against two constants: EXPECTED_COVERAGE_SCOPE and EXPECTED_SYSTEM_ATLAS_CHECK_COMMAND. A scope that is not the expected live-versus-generated string raises SELF_IGNORANCE_PROJECTION_PROTOCOL_SCOPE_MISMATCH. A check result record whose command is wrong, or whose status is outside ALLOWED_SYSTEM_ATLAS_CHECK_STATUSES (pass or a declared blocked refresh), raises SELF_IGNORANCE_SYSTEM_ATLAS_CHECK_RECEIPT_INVALID.
\n
Recomputing live row counts through build_kind_atlas
\n
_rows_with_live_kind_atlas_counts overwrites fixture counts with freshly built build_kind_atlas row counts, or falls back to declared counts at a lower realness rank when the builder is absent. Next _rows_with_live_kind_atlas_counts recomputes the live row counts. When the wider repository is present it imports build_kind_atlas from system.lib.kind_atlas and overwrites each fixture-declared count with the freshly built row_count. When the builder is not importable it records unavailable and falls back to the declared counts, which lowers the realness rank.
\n
Materialised ids derived from the graph
\n
_system_atlas_graph_materialization rejects an empty graph or wrong builder marker, then buckets entity ids by kind so the materialised set comes from the graph, not a declared number. Then _system_atlas_graph_materialization loads system_atlas_graph.json. It requires non-empty entity rows and rejects an empty graph with SELF_IGNORANCE_REAL_ATLAS_GRAPH_EMPTY. It requires the graph to carry the build_system_atlas.py builder marker; a different marker raises SELF_IGNORANCE_ATLAS_GRAPH_BUILDER_MISMATCH. _materialized_ids_by_kind_from_graph then buckets the entity ids by kind using MATERIALIZED_PREFIX_BY_KIND (concept_, mechanism_, pm_, std_, and so on), so the materialised set is derived from the graph rather than from a declared number.
\n
Graph ids must resolve to real source files
\n
_source_validate_expected_entity_ids requires each graph id to resolve to a real source file so the set cannot be inflated, and cross-checks the bundled slice against the live graph._source_validate_expected_entity_ids checks that graph-derived ids for concepts, mechanisms, and standards resolve to real source files through _entity_source_exists. An id with no file behind it raises SELF_IGNORANCE_EXPECTED_ENTITY_ID_SOURCE_MISSING, so the materialised set cannot be inflated with names that do not exist. _live_system_atlas_graph_materialization cross-checks the bundled graph slice against the live System Atlas graph when it is available; a divergence raises SELF_IGNORANCE_LIVE_SYSTEM_ATLAS_GRAPH_MISMATCH.
\n
Computing debt and rejecting tampered counts
\n
evaluate takes debt as live rows minus materialised, and raises errors when a declared count disagrees with the graph, understates the floor, or claims exhaustive absence. With those bound, evaluate computes the debt. For each kind it takes known_coverage_debt_count = max(live_kind_atlas_row_count - materialised_count, 0). _expected_entity_ids_source_backed decides whether declared ids may stand in for graph ids. A declared count that disagrees with the graph-derived count raises SELF_IGNORANCE_MATERIALIZATION_COUNT_NOT_GRAPH_DERIVED. A declared debt below the fixture floor raises SELF_IGNORANCE_COVERAGE_DEBT_MISMATCH. If absence_policy.claims_unknown_unknowns_exhaustive is true, it raises SELF_IGNORANCE_FORBIDDEN_ABSENCE_INFERENCE.
\n
Result record and the realness rank ladder
\n
The record reports status, per-kind debt, evidence, and scope limit, earning R4 only when every recompute and cross-check ran, else R3 or R2, with no source bodies inside. The result record reports status, the per-kind debt rows, the graph and live-graph evidence, the realness rank, and the scope limit from AUTHORITY_CEILING. The rank is R4 only when the debt was recomputed from live counts, the ids are source-backed, the Kind Atlas recompute ran, the protocol result record passed, and the live graph cross-check ran or the wider repository is absent. Otherwise it is R3 when a graph was supplied and R2 when it was not. No copied source bodies enter the record.
\n
Function
Role
evaluate
Runs the full check and emits the debt vector and evidence
_projection_protocol_receipt
Binds the coverage scope and the build_system_atlas.py --check result record
_rows_with_live_kind_atlas_counts
Recomputes live row counts through build_kind_atlas
_system_atlas_graph_materialization
Loads the graph, checks the builder marker, derives materialised ids
_source_validate_expected_entity_ids
Confirms graph ids resolve to real concept, mechanism, and standard files
_live_system_atlas_graph_materialization
Cross-checks the bundled slice against the live graph
evaluate_negative_case
Replays a named negative case from clean input copies
\nDiagram of the mechanism (7 steps).Diagram source & refs
Source refs
live row counts
kind_atlas_rows.json
materialised ids by kind
system_atlas_graph.json
scope + check result record
projection_protocol.json
flowchart TD KA["kind_atlas_rows.json live row counts"] Graph["system_atlas_graph.json materialised ids by kind"] Proto["projection_protocol.json scope + check result record"] Src["concept / mechanism / standard source files"] Eval["evaluate()"] Debt["known debt vector per kind"] Rec["metadata-only result record counts, refs, scope limit"] KA --> Eval Graph --> Eval Proto --> Eval Src --> Eval Eval --> Debt Debt --> Rec
\n
The current bundle's debt snapshot
\n
The current bundle shows 503 live rows against 307 materialised for a known debt of 196, all snapshot facts to rerun after any Kind Atlas, System Atlas, or source-set move. For the current bundle the count vector is concepts 41 rows against 30 materialised (debt 11), mechanisms 36 against 28 (debt 8), paper modules 225 against 220 (debt 5), and standards 201 against 29 (debt 172). Total live rows 503, materialised 307, known debt 196. These are snapshot facts read from the bundle, not fixed constants. Rerun the check after Kind Atlas, System Atlas, or the source set moves.
\n
Negative cases
\n
Two negatives replayed from clean copies
\n
evaluate_negative_case replays two named cases from fresh input, so an absence-inference claim or an understated debt is caught by the semantic evaluator even if it mislabels its own error code. The fixture declares two named negative cases in EXPECTED_NEGATIVE_CASES, and evaluate_negative_case replays each one against a fresh copy of the input rather than trusting a declared error label. forbidden_absence_inference sets claims_unknown_unknowns_exhaustive to true and must produce SELF_IGNORANCE_FORBIDDEN_ABSENCE_INFERENCE. coverage_debt_mismatch drops one materialised entity while raising the declared debt floor and must produce SELF_IGNORANCE_COVERAGE_DEBT_MISMATCH. Because the replay recomputes from clean copies, a case that lies about its own error code is still caught by the semantic evaluator.
\n
Perturbation proves the count tracks the graph
\n
Adding one source-backed standard entity moves known debt from 196 to 195, showing the count is coupled to the graph-derived set rather than a fixed number in prose. A perturbation check adds one real, source-backed standard entity to the graph and confirms the known debt moves from 196 to 195 while the result still passes. That shows the count is coupled to the graph-derived materialised set, not to a fixed number in prose.
\n
Prior Art Grounding
\n
Bounded coverage measurement as the analogue
\n
The analogue is software coverage measurement like coverage.py and bounded checks like OpenSSF Scorecard, which give a useful signal over a declared set without claiming a complete proof. The nearest ordinary analogue is software coverage measurement. A coverage tool reports what was exercised or missed over a declared source set, not every possible missing behaviour. coverage.py is a useful reference for that bounded, observed-coverage pattern. The health-signal side is adjacent to bounded repository checks such as OpenSSF Scorecard, where a fixed set of checks yields a useful risk signal without becoming a complete proof. This component applies the same discipline to navigation coverage debt and keeps the scope limit in the same record as the count.
A pass means the bundle, manifest, two negative cases, and paper-module corpus are still coherent, and it proves no freshness beyond the checked snapshot. A pass means the declared bundle, the source-module manifest, the two negative cases, and the paper-module corpus are still coherent. It does not establish freshness beyond the checked snapshot.
\n
Scope boundary
\n
Scope limit
\n
The narrow strongest claim
\n
Given the fixture, a marked graph, a live recompute, source-backed ids, and a valid result record, the component makes the coverage debt visible, recomputable, and checkable while refusing the tamper cases. The strongest claim the evidence supports is narrow. Given the public fixture, a graph carrying the build_system_atlas.py marker, a live Kind Atlas recompute, source-backed entity ids, and a valid protocol result record, the component makes the declared Kind Atlas coverage debt visible, recomputable, and checkable, and it refuses the named tamper cases.
\n
What the scope limit excludes
\n
The boundary is the fixture, graph-derived rows, source checks, and result records, and the ceiling excludes unknown-unknown omniscience, absence proof, source-file changes, and launch-scope decision. The proof boundary is the public coverage fixture, the graph-derived materialised rows, the source-file checks, the two semantic negative cases, and the listed validation result records. The scope limit excludes everything past that. It is not unknown-unknown omniscience, not absence proof, not a total repository search, not source-file changes, not launch or publishing-scope decision, not private-source equivalence, and not whole-system correctness. AUTHORITY_CEILING records the same limit in the result record, with literal_unknown_unknown_omniscience_authorized, absence_proof_authorized, source_mutation_authorized, and release_authorized all set to false.
\n",
"components": [
{
"anchor": "component-self_ignorance_coverage_ledger",
"component_id": "self_ignorance_coverage_ledger",
"public_label": "Self Ignorance Coverage Ledger"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 14,
"links": [
{
"id": "self_ignorance_coverage_ledger",
"kind": "organ",
"kind_label": "component",
"label": "Self Ignorance Coverage Ledger",
"relation": "explains",
"url": "component-self-ignorance-coverage-ledger.html"
},
{
"id": "mechanism.self_ignorance_coverage_ledger.validates_public_self_ignorance_coverage_ledger",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public self ignorance coverage ledger",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-self-ignorance-coverage-ledger-validates-public-self-ignorance-coverage-ledger"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-7",
"kind": "principle",
"kind_label": "principle",
"label": "Track known unknowns without claiming the unknown is mapped",
"relation": "governed by",
"url": "doctrine.html#dcard-p-7"
},
{
"id": "P-11",
"kind": "principle",
"kind_label": "principle",
"label": "Bind volatile facts to refresh routes",
"relation": "governed by",
"url": "doctrine.html#dcard-p-11"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
},
{
"id": "paper_module.doctrine_fact_claim_audit",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Doctrine Fact Claim Audit",
"relation": "depends on",
"url": "paper-module-doctrine-fact-claim-audit.html"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.self_ignorance_coverage_ledger",
"reader_packet": {
"abstract": "Self-Ignorance Coverage Ledger counts known coverage debt while refusing unknown-unknown or absence-proof claims.",
"authority_ceiling": "Known Kind Atlas coverage-debt projection only; no literal unknown-unknown omniscience, no absence proof, no total repository search proof, no source-file changes, no launch-scope decision, no publishing-scope decision, and no whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.self_ignorance_coverage_ledger.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-self-ignorance-coverage-ledger-diagram-1",
"html_url": "docs/paper-module-self-ignorance-coverage-ledger.html#paper-module-self-ignorance-coverage-ledger-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD KA[\"kind_atlas_rows.json live row counts\"] Graph[\"system_atlas_graph.json materialised ids by kind\"] Proto[\"projection_protocol.json scope + check result record\"] Src[\"concept / mechanism / standard source files\"] Eval[\"evaluate()\"] Debt[\"known debt vector per kind\"] Rec[\"metadata-only result record counts, refs, scope limit\"] KA --> Eval Graph --> Eval Proto --> Eval Src --> Eval Eval --> Debt Debt --> Rec",
"source_format": "mermaid",
"source_hash": "sha256:6f0d21beab5417d50eb15295614cd67fe101f395b8a5597c4fceff39dc9eb467",
"source_line_end": 80,
"source_line_start": 65
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.doctrine_fact_claim_audit",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Doctrine Fact Claim Audit",
"relation": "depends on",
"url": "paper-module-doctrine-fact-claim-audit.html"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "self_ignorance_coverage_ledger",
"kind": "organ",
"kind_label": "component",
"label": "Self Ignorance Coverage Ledger",
"relation": "explains",
"url": "component-self-ignorance-coverage-ledger.html"
},
{
"id": "mechanism.self_ignorance_coverage_ledger.validates_public_self_ignorance_coverage_ledger",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public self ignorance coverage ledger",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-self-ignorance-coverage-ledger-validates-public-self-ignorance-coverage-ledger"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-7",
"kind": "principle",
"kind_label": "principle",
"label": "Track known unknowns without claiming the unknown is mapped",
"relation": "governed by",
"url": "doctrine.html#dcard-p-7"
},
{
"id": "P-11",
"kind": "principle",
"kind_label": "principle",
"label": "Bind volatile facts to refresh routes",
"relation": "governed by",
"url": "doctrine.html#dcard-p-11"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.self_ignorance_coverage_ledger",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 132,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/self_ignorance_coverage_ledger.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "It answers one question: across declared Kind Atlas families, how many option-surface rows has the generated System Atlas not yet materialised? A navigation system that lists what it knows is easy to build. One that can state precisely what it has not yet covered is harder, and it is the more honest signal to a cold reader. This component answers a single question: for a declared set of Kind Atlas families, how many rows does the option surface expose that the generated System Atlas has not yet materialised?",
"level": 4,
"order": 1,
"section_id": "known-unknowns-in-the-current-surface",
"source_line_end": 6,
"source_line_start": 3,
"title": "Known unknowns in the current surface"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The answer is a computed debt vector: for each kind it subtracts materialised entities from the live row count and reports the difference as known coverage debt. The answer is a small debt vector, computed rather than asserted. For each selected kind the component recomputes the live Kind Atlas row count, counts the entities the System Atlas graph has actually materialised, and reports the difference as known coverage debt. It emits a metadata-only result record with the counts, the source references, and an explicit scope limit.",
"level": 4,
"order": 2,
"section_id": "computed-unknown-coverage-debt-vector",
"source_line_end": 10,
"source_line_start": 7,
"title": "Computed unknown-coverage debt vector"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Its unusual part is refusal: it rejects any claim that unknown-unknowns are exhaustive, treating known debt as a bounded count of visible gaps, never a completeness proof. The unusual part is what it refuses. It will not accept input that claims its unknown-unknowns are exhaustive, and it will not treat the absence of a row as proof that nothing is missing. Known debt is a bounded count of gaps it can see. It is never a completeness proof.",
"level": 4,
"order": 3,
"section_id": "refusing-false-completeness",
"source_line_end": 14,
"source_line_start": 11,
"title": "Refusing false completeness"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Systems overstate coverage either vaguely or overconfidently, so this component computes the finite gap between what the option surface exposes and what the graph materialised. Navigation systems overstate themselves in two opposite ways. A vague \"coverage is incomplete\" tells a cold reader nothing operational. A confident \"nothing else is missing\" is worse, because it converts absence of evidence into evidence of absence. This component occupies the narrow middle. For a declared finite domain of Kind Atlas families it computes the gap between what the option surface exposes and what the System Atlas graph has materialised, and it reports that gap as a number with its basis attached.",
"level": 4,
"order": 5,
"section_id": "the-narrow-middle-between-vague-and-overconfident",
"source_line_end": 20,
"source_line_start": 17,
"title": "The narrow middle between vague and overconfident"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Beside the bounded debt vector, the checker must reject exhaustive-coverage claims, hand-authored counts, substituted ids, stale baked ids, a tampered builder marker, or an understated debt floor. The positive output is a bounded debt vector. The negative output matters as much: the checker must reject input that claims exhaustive unknown-unknown coverage, hand-authored materialisation counts, substituted entity ids, stale baked ids used as authority, a tampered graph builder marker, or a coverage debt that understates the declared floor.",
"level": 4,
"order": 6,
"section_id": "the-negative-output-the-checker-must-reject",
"source_line_end": 24,
"source_line_start": 21,
"title": "The negative output the checker must reject"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run and runselfignorancebundle are the entrypoints, evaluate is the core over four bundle files, and evaluatenegativecase replays the semantic negatives. The runtime lives in src/microcosmcore/components/selfignorancecoverageledger.py. The command entrypoint is run; the exported-bundle entrypoint is runselfignorancebundle; the core evaluator is evaluate; the semantic negative-case replayer is evaluatenegativecase. evaluate consumes four bundle files and runs the following order.",
"level": 4,
"order": 8,
"section_id": "entrypoints-and-the-evaluate-core",
"source_line_end": 30,
"source_line_start": 27,
"title": "Entrypoints and the `evaluate` core"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "projectionprotocolreceipt checks the protocol against the expected scope and System Atlas check command, raising a mismatch when either the scope string or the result record status is wrong. First it loads kindatlasrows.json, materializedentities.json, and projectionprotocol.json. projectionprotocolreceipt checks the protocol against two constants: EXPECTEDCOVERAGESCOPE and EXPECTEDSYSTEMATLASCHECKCOMMAND. A scope that is not the expected live-versus-generated string raises SELFIGNORANCEPROJECTIONPROTOCOLSCOPEMISMATCH. A check result record whose command is wrong, or whose status is outside ALLOWEDSYSTEMATLASCHECKSTATUSES (pass or a declared blocked refresh), raises SELFIGNORANCESYSTEMATLAS\u2026",
"level": 4,
"order": 9,
"section_id": "binding-the-coverage-scope-and-check-receipt",
"source_line_end": 34,
"source_line_start": 31,
"title": "Binding the coverage scope and check result record"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "rowswithlivekindatlascounts overwrites fixture counts with freshly built buildkindatlas row counts, or falls back to declared counts at a lower realness rank when the builder is absent. Next rowswithlivekindatlascounts recomputes the live row counts. When the wider repository is present it imports buildkindatlas from system.lib.kindatlas and overwrites each fixture-declared count with the freshly built row_count. When the builder is not importable it records unavailable and falls back to the declared counts, which lowers the realness rank.",
"level": 4,
"order": 10,
"section_id": "recomputing-live-row-counts-through-build-kind-atlas",
"source_line_end": 38,
"source_line_start": 35,
"title": "Recomputing live row counts through `build_kind_atlas`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "systematlasgraphmaterialization rejects an empty graph or wrong builder marker, then buckets entity ids by kind so the materialised set comes from the graph, not a declared number. Then systematlasgraphmaterialization loads systematlasgraph.json. It requires non-empty entity rows and rejects an empty graph with SELFIGNORANCEREALATLASGRAPHEMPTY. It requires the graph to carry the buildsystematlas.py builder marker; a different marker raises SELFIGNORANCEATLASGRAPHBUILDERMISMATCH. materializedidsbykindfromgraph then buckets the entity ids by kind using MATERIALIZEDPREFIXBYKIND (concept, mechanism, pm, std, and so on), so the materialised set is derived from the graph rather than from a declar\u2026",
"level": 4,
"order": 11,
"section_id": "materialised-ids-derived-from-the-graph",
"source_line_end": 42,
"source_line_start": 39,
"title": "Materialised ids derived from the graph"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourcevalidateexpectedentityids requires each graph id to resolve to a real source file so the set cannot be inflated, and cross-checks the bundled slice against the live graph. sourcevalidateexpectedentityids checks that graph-derived ids for concepts, mechanisms, and standards resolve to real source files through entitysourceexists. An id with no file behind it raises SELFIGNORANCEEXPECTEDENTITYIDSOURCEMISSING, so the materialised set cannot be inflated with names that do not exist. livesystematlasgraphmaterialization cross-checks the bundled graph slice against the live System Atlas graph when it is available; a divergence raises SELFIGNORANCELIVESYSTEMATLASGRAPH_MISMATCH.",
"level": 4,
"order": 12,
"section_id": "graph-ids-must-resolve-to-real-source-files",
"source_line_end": 46,
"source_line_start": 43,
"title": "Graph ids must resolve to real source files"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluate takes debt as live rows minus materialised, and raises errors when a declared count disagrees with the graph, understates the floor, or claims exhaustive absence. With those bound, evaluate computes the debt. For each kind it takes knowncoveragedebtcount = max(livekindatlasrowcount - materialisedcount, 0). expectedentityidssourcebacked decides whether declared ids may stand in for graph ids. A declared count that disagrees with the graph-derived count raises SELFIGNORANCEMATERIALIZATIONCOUNTNOTGRAPHDERIVED. A declared debt below the fixture floor raises SELFIGNORANCECOVERAGEDEBTMISMATCH. If absencepolicy.claimsunknownunknownsexhaustive is true, it raises SELFIGNORANCEFORBIDDENABSEN\u2026",
"level": 4,
"order": 13,
"section_id": "computing-debt-and-rejecting-tampered-counts",
"source_line_end": 50,
"source_line_start": 47,
"title": "Computing debt and rejecting tampered counts"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "The record reports status, per-kind debt, evidence, and scope limit, earning R4 only when every recompute and cross-check ran, else R3 or R2, with no source bodies inside. The result record reports status, the per-kind debt rows, the graph and live-graph evidence, the realness rank, and the scope limit from scope_limit. The rank is R4 only when the debt was recomputed from live counts, the ids are source-backed, the Kind Atlas recompute ran, the protocol result record passed, and the live graph cross-check ran or the wider repository is absent. Otherwise it is R3 when a graph was supplied and R2 when it was not. No copied source bodies enter the record.",
"level": 4,
"order": 14,
"section_id": "result-record-and-the-realness-rank-ladder",
"source_line_end": 81,
"source_line_start": 51,
"title": "Result record and the realness rank ladder"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The current bundle shows 503 live rows against 307 materialised for a known debt of 196, all snapshot facts to rerun after any Kind Atlas, System Atlas, or source-set move. For the current bundle the count vector is concepts 41 rows against 30 materialised (debt 11), mechanisms 36 against 28 (debt 8), paper modules 225 against 220 (debt 5), and standards 201 against 29 (debt 172). Total live rows 503, materialised 307, known debt 196. These are snapshot facts read from the bundle, not fixed constants. Rerun the check after Kind Atlas, System Atlas, or the source set moves.",
"level": 4,
"order": 15,
"section_id": "the-current-bundle-s-debt-snapshot",
"source_line_end": 85,
"source_line_start": 82,
"title": "The current bundle's debt snapshot"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "negative-cases",
"source_line_end": 87,
"source_line_start": 86,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluatenegativecase replays two named cases from fresh input, so an absence-inference claim or an understated debt is caught by the semantic evaluator even if it mislabels its own error code. The fixture declares two named negative cases in EXPECTEDNEGATIVECASES, and evaluatenegativecase replays each one against a fresh copy of the input rather than trusting a declared error label. forbiddenabsenceinference sets claimsunknownunknownsexhaustive to true and must produce SELFIGNORANCEFORBIDDENABSENCEINFERENCE. coveragedebtmismatch drops one materialised entity while raising the declared debt floor and must produce SELFIGNORANCECOVERAGEDEBT_MISMATCH. Because the replay recomputes from clean co\u2026",
"level": 4,
"order": 17,
"section_id": "two-negatives-replayed-from-clean-copies",
"source_line_end": 91,
"source_line_start": 88,
"title": "Two negatives replayed from clean copies"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Adding one source-backed standard entity moves known debt from 196 to 195, showing the count is coupled to the graph-derived set rather than a fixed number in prose. A perturbation check adds one real, source-backed standard entity to the graph and confirms the known debt moves from 196 to 195 while the result still passes. That shows the count is coupled to the graph-derived materialised set, not to a fixed number in prose.",
"level": 4,
"order": 18,
"section_id": "perturbation-proves-the-count-tracks-the-graph",
"source_line_end": 95,
"source_line_start": 92,
"title": "Perturbation proves the count tracks the graph"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "prior-art-grounding",
"source_line_end": 97,
"source_line_start": 96,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The analogue is software coverage measurement like coverage.py and bounded checks like OpenSSF Scorecard, which give a useful signal over a declared set without claiming a complete proof. The nearest ordinary analogue is software coverage measurement. A coverage tool reports what was exercised or missed over a declared source set, not every possible missing behaviour. coverage.py is a useful reference for that bounded, observed-coverage pattern. The health-signal side is adjacent to bounded repository checks such as OpenSSF Scorecard, where a fixed set of checks yields a useful risk signal without becoming a complete proof. This component applies the same discipline to navigation coverage d\u2026",
"level": 4,
"order": 20,
"section_id": "bounded-coverage-measurement-as-the-analogue",
"source_line_end": 101,
"source_line_start": 98,
"title": "Bounded coverage measurement as the analogue"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate, run: For a manual bundle run into a throwaway directory:",
"level": 2,
"order": 21,
"section_id": "validation-receipt-path",
"source_line_end": 117,
"source_line_start": 102,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means the bundle, manifest, two negative cases, and paper-module corpus are still coherent, and it proves no freshness beyond the checked snapshot. A pass means the declared bundle, the source-module manifest, the two negative cases, and the paper-module corpus are still coherent. It does not establish freshness beyond the checked snapshot.",
"level": 4,
"order": 22,
"section_id": "what-a-pass-means-for-the-snapshot",
"source_line_end": 121,
"source_line_start": 118,
"title": "What a pass means for the snapshot"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 23,
"section_id": "claim-ceiling",
"source_line_end": 123,
"source_line_start": 122,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Given the fixture, a marked graph, a live recompute, source-backed ids, and a valid result record, the component makes the coverage debt visible, recomputable, and checkable while refusing the tamper cases. The strongest claim the evidence supports is narrow. Given the public fixture, a graph carrying the buildsystematlas.py marker, a live Kind Atlas recompute, source-backed entity ids, and a valid protocol result record, the component makes the declared Kind Atlas coverage debt visible, recomputable, and checkable, and it refuses the named tamper cases.",
"level": 4,
"order": 24,
"section_id": "the-narrow-strongest-claim",
"source_line_end": 127,
"source_line_start": 124,
"title": "The narrow strongest claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The boundary is the fixture, graph-derived rows, source checks, and result records, and the ceiling excludes unknown-unknown omniscience, absence proof, source-file changes, and launch-scope decision. The proof boundary is the public coverage fixture, the graph-derived materialised rows, the source-file checks, the two semantic negative cases, and the listed validation result records. The scope limit excludes everything past that. It is not unknown-unknown omniscience, not absence proof, not a total repository search, not source-file changes, not launch or public sharing decision, not private-source equivalence, and not whole-system correctness. AUTHORITYCEILING records the same limit in th\u2026",
"level": 4,
"order": 25,
"section_id": "what-the-authority-ceiling-excludes",
"source_line_end": 131,
"source_line_start": 128,
"title": "What the scope limit excludes"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/self_ignorance_coverage_ledger.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Self-Ignorance Coverage Ledger counts known coverage debt while refusing unknown-unknown or absence-proof claims."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "self_ignorance_coverage_ledger.py",
"path": "src/microcosm_core/organs/self_ignorance_coverage_ledger.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for known coverage-debt fixture validation, missing-category checks, negative cases, and scope limits.",
"symbols": [
"run",
"run_self_ignorance_bundle",
"evaluate",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/self_ignorance_coverage_ledger.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/self_ignorance_coverage_ledger.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/self_ignorance_coverage_ledger.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Self-Ignorance Coverage Ledger validates a public known-debt coverage fixture: declared Kind Atlas gaps, missing coverage categories, negative cases, source-open body imports, and scope limits. It records what the system knows it has not covered without claiming omniscience, absence proof, total search, source-file changes, public sharing, launch-scope decision, or whole-system correctness.",
"summary_line": "Self-Ignorance Coverage Ledger counts known coverage debt while refusing unknown-unknown or absence-proof claims.",
"title": "Self-Ignorance Coverage Ledger"
},
{
"anchor": "paper-module-tool-server-pressure-inventory",
"authority_ceiling": "Declared public helper-process pressure inventory fixture and source-module digest evidence only; no live process reads, process signalling, host mutation, launch-scope decision, external model access, non-public data equivalence, publishing-scope decision, or whole-system correctness.",
"body_html": "
Long sessions leave helper processes that slow the host, but a reaper that kills stale helpers is dangerous when an abandoned process looks like one a live session still uses. Long agent sessions leave helper processes behind: MCP servers, dev servers, keepalives. They pile up and the host slows down. The obvious fix is a reaper that walks the process table and kills stale helpers. A reaper is also dangerous, because the hard case is telling a genuinely abandoned process apart from one a live session is still using. Kill the wrong one and you break work in flight.
\n
It is the read-only classifier over synthetic input: it classifies each helper by kind and owner chain, marks the few safe to close, and requests launch for over-budget owners.tool_server_pressure_inventory is the read-only classifier from that problem, ported to run over synthetic input. It takes a fixture that supplies a ps-shaped process table, a helper-kind policy, and an owner-status taxonomy. It classifies each helper by kind and by owner chain, marks the few that are safe to close, and for owners that are over their helper budget it emits a launch request rather than a kill. Every row carries a command_hash, not a command line.
\n
It reads no live process table, imports no process-control module, and sends no signal, writing metadata-only records that show a reaper's safety reasoning without the harmful part. The component reads no live process table, imports no process-control module, and sends no signal. It writes metadata-only result records so a reader can see the safety reasoning of a reaper without the part that could harm a host.
\n
Purpose
\n
The owner chain is the real decision
\n
A loose reaper closes from a shallow signal like age or name, but the real decision is the owner chain: a launchd-parented helper is detached while a session-traced one is not. A loose reaper decides "close this" from a shallow signal: the process is old, or its name matches a helper pattern. That is enough to close a live session's MCP server by mistake. The interesting decision is the owner chain. A helper whose parent is launchd (ppid == 1) has been detached from any session and is a candidate. A helper that still traces back through a live agent session is not.
\n
The narrow safe-close predicate
\n
A row becomes a close candidate only when it is detached, its kind is allowlisted, and its idle age passes the 300-second threshold; everything else routes to owner-check or keep. So the safe-close predicate is deliberately narrow. A row becomes a close candidate only when it is detached, its kind is on an allowlist, and its idle age is past a threshold (DEFAULT_MIN_AGE_SECONDS, 300 seconds). Everything else routes to requires_owner_check or keep.
\n
Active owners get a launch request, not a kill
\n
When an owner is over budget it emits a launch request asking the session to launch its own lease, and an audit enforces that an active-owner descendant is never a close candidate. The second half of the design is that active owners never get a kill. When an owner is over its helper budget the component emits a launch request: a row that asks the owning session to launch or reuse its own lease. The inventory is not a kill list. The invariant, enforced by an audit pass over the component's own output, is that an active-owner descendant can never be a close candidate.
\n
How it works
\n
One pass, two entry points
\n
The pipeline runs in one pass; run validates a fixture while run_pressure_bundle validates an exported source-body bundle, both delegating to _build_result. The pipeline runs in one pass over the fixture. run loads the positive and negative inputs and writes the result records. run_pressure_bundle is the second entrypoint: it validates an exported source-body bundle instead of a fixture. Both call _build_result, which orchestrates loading, classification, redaction, and the negative-case sweep.
\n
The classifier's decision law
\n
build_tool_server_pressure_inventory parses rows, matches each command to a helper kind, walks the parent chain for owner status, and routes to keep, safe-close, or owner-check. The classifier is build_tool_server_pressure_inventory. It calls _parse_process_rows to turn the synthetic ps_text into rows of pid, ppid, age, and command, then for each row calls _process_kind to match the command against the policy's match_substrings. Rows whose command matches no helper kind are dropped. For each kept row _owner_status_for_process walks the parent chain up to eight hops, stopping at a keep runtime, a detached parent, an active owner hint, or a cycle. _inventory_owner_and_decision then applies the decision law: keep kinds route to keep, detached and allowlisted and old-enough rows route to candidate_safe_close, and everything else routes to requires_owner_check. Over-budget active owners are summarized by _active_owner_pressure_groups, and _owner_release_request_for_group builds a helper_owner_release_request_v1 whose permitted_action is ask_owner_to_release and whose result is requested.
\n
Two guards over the classifier
\n
_redaction_findings rejects command previews, signal keys, and absolute paths, and _audit_inventory_claim re-reads the output to fail any close candidate that breaks the safety invariant. Two guards sit on top of the classifier. _redaction_findings walks every emitted dict and rejects a command preview key, a truthy process-signal key, or an absolute host path. _audit_inventory_claim re-reads the component's own inventory and fails it if any close candidate is an active owner, is not a ppid == 1 allowlisted detached orphan, or is younger than the age threshold, and it fails any launch request that reads as a kill. _positive_findings runs these over the live fixture and _negative_findings runs them over the eight crafted failure inputs.
\n
The bundle lane's manifest check
\n
_source_module_manifest_result verifies each exported module's relation, recomputes the target SHA-256, confirms required anchors, and reruns the redaction guard, keeping result records metadata-only._source_module_manifest_result covers the bundle lane. It reads source_module_manifest.json, checks each declared module's source_to_target_relation against the accepted set, recomputes the target body's SHA-256 against the declared target_sha256, confirms the required anchors are present in the body, and reruns the redaction guard on the body text. Result records carry refs, hashes, counts, and verdicts only; body_in_receipt stays false. _write_receipts then emits the result, board, validation, and sign-off records.
\n
Function
Role
run
Validate a fixture and write the four result records
run_pressure_bundle
Validate an exported source-body bundle
_build_result
Load inputs, classify, run guards, assemble the result
build_tool_server_pressure_inventory
Classify each helper by kind, owner chain, and decision
_parse_process_rows
Parse synthetic ps_text into pid, ppid, age, command
_owner_status_for_process
Walk the parent chain up to eight hops for owner status
_inventory_owner_and_decision
Apply the keep / safe-close / owner-check decision law
_active_owner_pressure_groups
Group over-budget active owners into launch requests
_redaction_findings
Reject command previews, process signals, absolute paths
_audit_inventory_claim
Re-check the component's own output against the invariant
_source_module_manifest_result
Verify the exported source body by digest and anchors
_write_receipts
Write the metadata-only result, board, and validation records
\nDiagram of the mechanism (9 steps).Diagram source & refs
flowchart TD Fixture["Synthetic process table pressure policy, owner classes"] Parse["_parse_process_rows ps rows to pid, ppid, age, cmd"] Kind["_process_kind match command to helper kind"] Owner["_owner_status_for_process walk parent chain up to 8 hops"] Decision["_inventory_owner_and_decision keep / safe-close / owner-check"] Groups["_active_owner_pressure_groups over-budget owners to launch request"] Audit["_audit_inventory_claim re-check own output"] Redact["_redaction_findings reject previews, paths, signals"] Result records["_write_receipts metadata-only result records"] Fixture --> Parse Parse --> Kind Kind --> Owner Owner --> Decision Decision --> Groups Decision --> Audit Groups --> Audit Audit --> Redact Redact --> Result records
\n
Negative cases
\n
EXPECTED_NEGATIVE_CASES lists eight inputs the fixture must reject, each keyed to the error code it must raise.EXPECTED_NEGATIVE_CASES lists eight inputs the fixture must reject, each keyed to the error code it must raise.
\n
Eight rejection codes for kill-unsafe inputs
\n
The eight inputs each raise a specific code, covering active-owner or cycle kill candidates, unknown-owner kills, premature closes, sent signals, command or path leaks, and launch overclaims.active_owner_kill_candidate and owner_chain_cycle_safe_close both raise TSPI_ACTIVE_OWNER_KILL_CANDIDATE: an active-owner descendant, including one reached through a parent-chain cycle, is marked as a close candidate. unknown_owner_kill raises TSPI_UNKNOWN_OWNER_KILL_FORBIDDEN when a close candidate is not a ppid == 1 allowlisted detached orphan. premature_safe_close raises TSPI_PREMATURE_SAFE_CLOSE when a detached orphan younger than the age threshold is marked closeable. process_signal_sent raises TSPI_PROCESS_SIGNAL_FORBIDDEN. command_preview_leak raises TSPI_COMMAND_PREVIEW_FORBIDDEN when a row carries a live command line instead of a hash. absolute_path_leak raises TSPI_ABSOLUTE_PATH_FORBIDDEN. owner_release_overclaim raises TSPI_OWNER_RELEASE_OVERCLAIM when a launch request reads as a kill, terminate, or force-close.
\n
A run passes only when every expected case is observed, the positive fixture raises nothing, and the secret scan is clean. A run passes only when every expected case is observed, the positive fixture raises nothing, and the secret scan is clean.
\n
Prior Art Grounding
\n
Process-inventory and owner-reference prior art
\n
It draws on psutil process iteration, Kubernetes owner-reference garbage collection, and MCP tool servers, but stays weaker: it classifies synthetic rows and reads no live host state. The classifier draws on established process-inventory and ownership patterns. psutil.process_iter() is a common way to iterate process metadata without ad hoc ps parsing. Kubernetes garbage collection uses owner references to separate objects that may be collected from objects a live controller still owns, which is the same detached-versus-owned distinction applied here. The Model Context Protocol gives the local "server exposes callable tools" shape that names these helpers. This component keeps the result weaker than any of them: it classifies synthetic rows for pressure and close eligibility, and it reads no live host state.
\n
psutil process iteration: https://psutil.readthedocs.io/en/latest/#psutil.process_iter
Model Context Protocol tool servers: https://modelcontextprotocol.io/docs/concepts/tools
\n
Validation Result record Path
\n
From microcosm-substrate, with result records written under /tmp:
\n
What a pass certifies
\n
A pass means rows stayed digest-only, the eight negative cases raised their codes, and the bundle body matched its declared digest and anchors. A pass means synthetic classification and source-manifest shape held: rows stayed digest-only, the eight negative cases raised their codes, and the bundle body matched its declared digest and anchors.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
Over synthetic fixtures it classifies kind, owner status, and close eligibility correctly, keeps active owners out of the close set, requests launch for over-budget owners, and rejects the eight failures. The strongest claim this component supports is that over synthetic process-table fixtures it classifies helper kind, owner status, and close eligibility correctly, keeps active-owner descendants out of the close set, turns over-budget owners into launch requests, and rejects the eight named boundary failures. The bundle lane adds that an exported source body matched its declared digest and anchors while staying metadata-only.
\n
The whole proof boundary and ceiling
\n
That is the whole boundary: the component reads no live table, sends no signal, mutates no host, and AUTHORITY_CEILING holds every action flag at false as projection and validation only. That is the whole proof boundary. The component does not read a live process table, send a process signal, mutate host state, or authorize cleanup. It does not use external model services, export private account or session state, or claim whole-system correctness. AUTHORITY_CEILING holds process_signal_authority, live_process_table_read_authorized, host_mutation_authorized, release_authorized, and whole_system_correctness_claim at false. The scope limit is projection and validation only.
\n",
"components": [
{
"anchor": "component-tool_server_pressure_inventory",
"component_id": "tool_server_pressure_inventory",
"public_label": "Tool Server Pressure Inventory"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 14,
"links": [
{
"id": "tool_server_pressure_inventory",
"kind": "organ",
"kind_label": "component",
"label": "Tool Server Pressure Inventory",
"relation": "explains",
"url": "component-tool-server-pressure-inventory.html"
},
{
"id": "mechanism.tool_server_pressure_inventory.validates_public_tool_server_pressure_inventory",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public tool server pressure inventory",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-tool-server-pressure-inventory-validates-public-tool-server-pressure-inventory"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-4",
"kind": "principle",
"kind_label": "principle",
"label": "Possession is not permission",
"relation": "governed by",
"url": "doctrine.html#dcard-p-4"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "AX-3",
"kind": "axiom",
"kind_label": "axiom",
"label": "Authority by derivation, not possession",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-3"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.provider_context_recipe_budget",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Provider Context Recipe Budget",
"relation": "depends on",
"url": "paper-module-provider-context-recipe-budget.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.tool_server_pressure_inventory",
"reader_packet": {
"abstract": "Tool Server Pressure Inventory validates public pressure-inventory fixtures without reading or mutating live host process state.",
"authority_ceiling": "Declared public helper-process pressure inventory fixture and source-module digest evidence only; no live process reads, process signalling, host mutation, launch-scope decision, external model access, non-public data equivalence, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.tool_server_pressure_inventory.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-tool-server-pressure-inventory-diagram-1",
"html_url": "docs/paper-module-tool-server-pressure-inventory.html#paper-module-tool-server-pressure-inventory-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Fixture[\"Synthetic process table pressure policy, owner classes\"] Parse[\"_parse_process_rows ps rows to pid, ppid, age, cmd\"] Kind[\"_process_kind match command to helper kind\"] Owner[\"_owner_status_for_process walk parent chain up to 8 hops\"] Decision[\"_inventory_owner_and_decision keep / safe-close / owner-check\"] Groups[\"_active_owner_pressure_groups over-budget owners to launch request\"] Audit[\"_audit_inventory_claim re-check own output\"] Redact[\"_redaction_findings reject previews, paths, signals\"] Result records[\"_write_receipts metadata-only result records\"] Fixture --> Parse Parse --> Kind Kind --> Owner Owner --> Decision Decision --> Groups Decision --> Audit Groups --> Audit Audit --> Redact Redact --> Result records",
"source_format": "mermaid",
"source_hash": "sha256:dad89b1312d4e0d2f18340e645cca35c39b0490686e09d0189d97bd7f6fb3b64",
"source_line_end": 83,
"source_line_start": 62
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-3",
"kind": "axiom",
"kind_label": "axiom",
"label": "Authority by derivation, not possession",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-3"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.provider_context_recipe_budget",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Provider Context Recipe Budget",
"relation": "depends on",
"url": "paper-module-provider-context-recipe-budget.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "tool_server_pressure_inventory",
"kind": "organ",
"kind_label": "component",
"label": "Tool Server Pressure Inventory",
"relation": "explains",
"url": "component-tool-server-pressure-inventory.html"
},
{
"id": "mechanism.tool_server_pressure_inventory.validates_public_tool_server_pressure_inventory",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public tool server pressure inventory",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-tool-server-pressure-inventory-validates-public-tool-server-pressure-inventory"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-4",
"kind": "principle",
"kind_label": "principle",
"label": "Possession is not permission",
"relation": "governed by",
"url": "doctrine.html#dcard-p-4"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.tool_server_pressure_inventory",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 136,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/tool_server_pressure_inventory.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "Long sessions leave helper processes that slow the host, but a reaper that kills stale helpers is dangerous when an abandoned process looks like one a live session still uses. Long agent sessions leave helper processes behind: MCP servers, dev servers, keepalives. They pile up and the host slows down. The obvious fix is a reaper that walks the process table and kills stale helpers. A reaper is also dangerous, because the hard case is telling a genuinely abandoned process apart from one a live session is still using. Kill the wrong one and you break work in flight.",
"level": 4,
"order": 1,
"section_id": "long-sessions-leave-helper-processes-that-slow-the-host",
"source_line_end": 6,
"source_line_start": 3,
"title": "Long sessions leave helper processes that slow the host"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It is the read-only classifier over synthetic input: it classifies each helper by kind and owner chain, marks the few safe to close, and requests launch for over-budget owners. toolserverpressureinventory is the read-only classifier from that problem, ported to run over synthetic input. It takes a fixture that supplies a ps-shaped process table, a helper-kind policy, and an owner-status taxonomy. It classifies each helper by kind and by owner chain, marks the few that are safe to close, and for owners that are over their helper budget it emits a launch request rather than a kill. Every row carries a commandhash, not a command line.",
"level": 4,
"order": 2,
"section_id": "it-is-the-read-only-classifier-over-synthetic-input",
"source_line_end": 10,
"source_line_start": 7,
"title": "It is the read-only classifier over synthetic input"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It reads no live process table, imports no process-control module, and sends no signal, writing metadata-only records that show a reaper's safety reasoning without the harmful part. The component reads no live process table, imports no process-control module, and sends no signal. It writes metadata-only result records so a reader can see the safety reasoning of a reaper without the part that could harm a host.",
"level": 4,
"order": 3,
"section_id": "it-reads-no-live-process-table",
"source_line_end": 14,
"source_line_start": 11,
"title": "It reads no live process table"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A loose reaper closes from a shallow signal like age or name, but the real decision is the owner chain: a launchd-parented helper is detached while a session-traced one is not. A loose reaper decides \"close this\" from a shallow signal: the process is old, or its name matches a helper pattern. That is enough to close a live session's MCP server by mistake. The interesting decision is the owner chain. A helper whose parent is launchd (ppid == 1) has been detached from any session and is a candidate. A helper that still traces back through a live agent session is not.",
"level": 4,
"order": 5,
"section_id": "the-owner-chain-is-the-real-decision",
"source_line_end": 20,
"source_line_start": 17,
"title": "The owner chain is the real decision"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A row becomes a close candidate only when it is detached, its kind is allowlisted, and its idle age passes the 300-second threshold; everything else routes to owner-check or keep. So the safe-close predicate is deliberately narrow. A row becomes a close candidate only when it is detached, its kind is on an allowlist, and its idle age is past a threshold (DEFAULTMINAGESECONDS, 300 seconds). Everything else routes to requiresowner_check or keep.",
"level": 4,
"order": 6,
"section_id": "the-narrow-safe-close-predicate",
"source_line_end": 24,
"source_line_start": 21,
"title": "The narrow safe-close predicate"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "When an owner is over budget it emits a launch request asking the session to launch its own lease, and an audit enforces that an active-owner descendant is never a close candidate. The second half of the design is that active owners never get a kill. When an owner is over its helper budget the component emits a launch request: a row that asks the owning session to launch or reuse its own lease. The inventory is not a kill list. The invariant, enforced by an audit pass over the component's own output, is that an active-owner descendant can never be a close candidate.",
"level": 4,
"order": 7,
"section_id": "active-owners-get-a-release-request-not-a-kill",
"source_line_end": 28,
"source_line_start": 25,
"title": "Active owners get a launch request, not a kill"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 30,
"source_line_start": 29,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The pipeline runs in one pass; run validates a fixture while runpressurebundle validates an exported source-body bundle, both delegating to buildresult. The pipeline runs in one pass over the fixture. run loads the positive and negative inputs and writes the result records. runpressurebundle is the second entrypoint: it validates an exported source-body bundle instead of a fixture. Both call buildresult, which orchestrates loading, classification, redaction, and the negative-case sweep.",
"level": 4,
"order": 9,
"section_id": "one-pass-two-entry-points",
"source_line_end": 34,
"source_line_start": 31,
"title": "One pass, two entry points"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildtoolserverpressureinventory parses rows, matches each command to a helper kind, walks the parent chain for owner status, and routes to keep, safe-close, or owner-check. The classifier is buildtoolserverpressureinventory. It calls parseprocessrows to turn the synthetic pstext into rows of pid, ppid, age, and command, then for each row calls processkind to match the command against the policy's matchsubstrings. Rows whose command matches no helper kind are dropped. For each kept row ownerstatusforprocess walks the parent chain up to eight hops, stopping at a keep runtime, a detached parent, an active owner hint, or a cycle. inventoryowneranddecision then applies the decision law: keep ki\u2026",
"level": 4,
"order": 10,
"section_id": "the-classifier-s-decision-law",
"source_line_end": 38,
"source_line_start": 35,
"title": "The classifier's decision law"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "redactionfindings rejects command previews, signal keys, and absolute paths, and auditinventoryclaim re-reads the output to fail any close candidate that breaks the safety invariant. Two guards sit on top of the classifier. redactionfindings walks every emitted dict and rejects a command preview key, a truthy process-signal key, or an absolute host path. auditinventoryclaim re-reads the component's own inventory and fails it if any close candidate is an active owner, is not a ppid == 1 allowlisted detached orphan, or is younger than the age threshold, and it fails any launch request that reads as a kill. positivefindings runs these over the live fixture and negative_findings runs them over\u2026",
"level": 4,
"order": 11,
"section_id": "two-guards-over-the-classifier",
"source_line_end": 42,
"source_line_start": 39,
"title": "Two guards over the classifier"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "sourcemodulemanifestresult verifies each exported module's relation, recomputes the target SHA-256, confirms required anchors, and reruns the redaction guard, keeping result records metadata-only. sourcemodulemanifestresult covers the bundle lane. It reads sourcemodulemanifest.json, checks each declared module's sourcetotargetrelation against the accepted set, recomputes the target body's SHA-256 against the declared targetsha256, confirms the required anchors are present in the body, and reruns the redaction guard on the body text. Result records carry refs, hashes, counts, and verdicts only; bodyinreceipt stays false. write_receipts then emits the result, board, validation, and sign-off r\u2026",
"level": 4,
"order": 12,
"section_id": "the-bundle-lane-s-manifest-check",
"source_line_end": 84,
"source_line_start": 43,
"title": "The bundle lane's manifest check"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 86,
"source_line_start": 85,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVECASES lists eight inputs the fixture must reject, each keyed to the error code it must raise. EXPECTEDNEGATIVECASES lists eight inputs the fixture must reject, each keyed to the error code it must raise.",
"level": 4,
"order": 14,
"section_id": "expected-negative-cases-lists-eight-inputs-the-fixture-must-reject",
"source_line_end": 91,
"source_line_start": 87,
"title": "`EXPECTED_NEGATIVE_CASES` lists eight inputs the fixture must reject"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The eight inputs each raise a specific code, covering active-owner or cycle kill candidates, unknown-owner kills, premature closes, sent signals, command or path leaks, and launch overclaims. activeownerkillcandidate and ownerchaincyclesafeclose both raise TSPIACTIVEOWNERKILLCANDIDATE: an active-owner descendant, including one reached through a parent-chain cycle, is marked as a close candidate. unknownownerkill raises TSPIUNKNOWNOWNERKILLFORBIDDEN when a close candidate is not a ppid == 1 allowlisted detached orphan. prematuresafeclose raises TSPIPREMATURESAFECLOSE when a detached orphan younger than the age threshold is marked closeable. processsignalsent raises TSPIPROCESSSIGNALFORBIDDEN\u2026",
"level": 4,
"order": 15,
"section_id": "eight-rejection-codes-for-kill-unsafe-inputs",
"source_line_end": 95,
"source_line_start": 92,
"title": "Eight rejection codes for kill-unsafe inputs"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A run passes only when every expected case is observed, the positive fixture raises nothing, and the secret scan is clean. A run passes only when every expected case is observed, the positive fixture raises nothing, and the secret scan is clean.",
"level": 4,
"order": 16,
"section_id": "a-run-passes-only-when-every-expected-case-is-observed",
"source_line_end": 100,
"source_line_start": 96,
"title": "A run passes only when every expected case is observed"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 102,
"source_line_start": 101,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "It draws on psutil process iteration, Kubernetes owner-reference garbage collection, and MCP tool servers, but stays weaker: it classifies synthetic rows and reads no live host state. The classifier draws on established process-inventory and ownership patterns. psutil.process_iter() is a common way to iterate process metadata without ad hoc ps parsing. Kubernetes garbage collection uses owner references to separate objects that may be collected from objects a live controller still owns, which is the same detached-versus-owned distinction applied here. The Model Context Protocol gives the local \"server exposes callable tools\" shape that names these helpers. This component keeps the result we\u2026",
"level": 4,
"order": 18,
"section_id": "process-inventory-and-owner-reference-prior-art",
"source_line_end": 110,
"source_line_start": 103,
"title": "Process-inventory and owner-reference prior art"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate, with result records written under /tmp:",
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 121,
"source_line_start": 111,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means rows stayed digest-only, the eight negative cases raised their codes, and the bundle body matched its declared digest and anchors. A pass means synthetic classification and source-manifest shape held: rows stayed digest-only, the eight negative cases raised their codes, and the bundle body matched its declared digest and anchors.",
"level": 4,
"order": 20,
"section_id": "what-a-pass-certifies",
"source_line_end": 125,
"source_line_start": 122,
"title": "What a pass certifies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 127,
"source_line_start": 126,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Over synthetic fixtures it classifies kind, owner status, and close eligibility correctly, keeps active owners out of the close set, requests launch for over-budget owners, and rejects the eight failures. The strongest claim this component supports is that over synthetic process-table fixtures it classifies helper kind, owner status, and close eligibility correctly, keeps active-owner descendants out of the close set, turns over-budget owners into launch requests, and rejects the eight named boundary failures. The bundle lane adds that an exported source body matched its declared digest and anchors while staying metadata-only.",
"level": 4,
"order": 22,
"section_id": "the-strongest-supported-claim",
"source_line_end": 131,
"source_line_start": 128,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "That is the whole boundary: the component reads no live table, sends no signal, mutates no host, and AUTHORITYCEILING holds every action flag at false as projection and validation only. That is the whole proof boundary. The component does not read a live process table, send a process signal, mutate host state, or authorize cleanup. It does not use external model services, export private account or session state, or claim whole-system correctness. AUTHORITYCEILING holds processsignalauthority, liveprocesstablereadauthorized, hostmutationauthorized, releaseauthorized, and wholesystemcorrectnessclaim at false. The scope limit is projection and validation only.",
"level": 4,
"order": 23,
"section_id": "the-whole-proof-boundary-and-ceiling",
"source_line_end": 135,
"source_line_start": 132,
"title": "The whole proof boundary and ceiling"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/tool_server_pressure_inventory.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Tool Server Pressure Inventory validates public pressure-inventory fixtures without reading or mutating live host process state."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "tool_server_pressure_inventory.py",
"path": "src/microcosm_core/organs/tool_server_pressure_inventory.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for tool-server pressure fixture validation, redaction checks, source-module digest checks, negative cases, and scope limits.",
"symbols": [
"run",
"run_pressure_bundle",
"build_tool_server_pressure_inventory",
"build_pressure_hygiene_relief_receipt",
"_redaction_findings",
"_build_result",
"_write_receipts",
"EXPECTED_NEGATIVE_CASES"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/tool_server_pressure_inventory.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/tool_server_pressure_inventory.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/tool_server_pressure_inventory.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Tool Server Pressure Inventory validates the public tool-server pressure membrane: synthetic process rows, active-owner descendants, over-budget owner launch requests, redaction findings, source-module digests, metadata-only result records, negative cases, and scope limits. It treats pressure inventory rows as fixture evidence, not live process control, process signalling, host mutation, provider authority, launch-scope decision, public sharing, or whole-system correctness.",
"summary_line": "Tool Server Pressure Inventory validates public pressure-inventory fixtures without reading or mutating live host process state.",
"title": "Tool Server Pressure Inventory"
},
{
"anchor": "paper-module-mechanistic-interpretability-circuit-attribution-replay",
"authority_ceiling": "Declared public circuit-attribution runtime result record and source-module import evidence only; no live model access, model-transparency product claim, export of private weights, raw activations, proprietary prompts, hidden reasoning, external model access, benchmark claims, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
Interpretability writing is easy to overstate. A named feature can read like understanding. Interpretability writing is easy to overstate. A named feature can read like understanding. A graph picture can read like a discovered circuit. A small local script can read like access to a real model. This component exists to hold one kind of claim to a checkable size.
\n
It answers one question. Before the system lets a circuit-attribution story stand as public evidence, does the story survive a deterministic replay instead of being taken on trust? It answers one question. Before the system lets a circuit-attribution story stand as public evidence, does the story survive a deterministic replay instead of being taken on trust? The runtime lives in src/microcosm_core/organs/mechanistic_interpretability_circuit_attribution_replay.py. It reads a fixture or an exported bundle, recomputes a tiny toy transformer, checks the graph and source-body evidence, and writes result records that carry refs, digests, counts, and verdicts. They carry no model weights, no raw activations, no prompts, and no hidden reasoning.
\n
The current fixture has six sparse features and six replay rows. A pass is bounded evidence about that fixture. The current fixture has six sparse features and six replay rows. A pass is bounded evidence about that fixture. It is not access to any trained model.
\n
Purpose
\n
The problem is that an interpretability-shaped fixture can look convincing while being wrong or stale. A feature label can lack a catalog id. The problem is that an interpretability-shaped fixture can look convincing while being wrong or stale. A feature label can lack a catalog id. A graph can be a screenshot with no traversable edges. A row can claim transparency with no intervention result record behind it. And an attribution claim that was once true of its inputs can stay in place after the inputs change.
\n
A loose check passes all of these. This component makes the claim smaller and forces each claim-bearing row across four gates: a replay-schema gate, a graph-traversal gate, a toy-transformer recomputation gate, and a source-body boundary gate. A loose check passes all of these. This component makes the claim smaller and forces each claim-bearing row across four gates: a replay-schema gate, a graph-traversal gate, a toy-transformer recomputation gate, and a source-body boundary gate. A row passes only when all four hold.
\n
How it works
\n
Two entry points drive the same gates. run validates the first-wave fixture with negative cases included. Two entry points drive the same gates. run validates the first-wave fixture with negative cases included. run_attribution_bundle validates an exported bundle without the negative-case files, and can reuse a cached bundle result when _fresh_bundle_receipt confirms the cache is still coupled to the current inputs. Both call _build_result, which runs the checks in order and then _write_receipts, which writes the result, board, validation, and sign-off records.
\n
_build_result loads the fixture payloads, then for each replay row it calls _replay_policy_findings, _graph_analysis_for_replay, and a feature-catalog resolution check._build_result loads the fixture payloads, then for each replay row it calls _replay_policy_findings, _graph_analysis_for_replay, and a feature-catalog resolution check. After the loop it runs _weight_sequence_analysis once across all rows, _toy_transformer_attribution_runtime once, and scan_paths over the inputs and copied source modules. The overall status is pass only when every gate passes and the secret scan passes.
\n
_replay_policy_findings is the schema gate. It requires all 29 fields in REQUIRED_REPLAY_FIELDS, a toy prompt ref with no inline prompt body, at least one sparse feature id, and a replacement_model_approximation_score inside the range 0 to 1._replay_policy_findings is the schema gate. It requires all 29 fields in REQUIRED_REPLAY_FIELDS, a toy prompt ref with no inline prompt body, at least one sparse feature id, and a replacement_model_approximation_score inside the range 0 to 1. It emits INTERPRETABILITY_INTERVENTION_RECEIPT_REQUIRED when a row claims transparency without a causal_intervention_receipt_ref, and INTERPRETABILITY_FAITHFULNESS_REQUIRES_LIMITS when a row claims faithfulness without a faithfulness_limit_ref. It rejects any row whose private_model_weights_exported, raw_activation_dump_exported, proprietary_prompt_exported, hidden_chain_of_thought_exported, or release_authorized flag is true, and any row whose serialized text contains a private needle from PRIVATE_NEEDLES.
\n
_graph_analysis_for_replay is the traversal gate. It reads graph_nodes and graph_edges, checks that every edge endpoint resolves to a declared node, then does a depth-first walk from the row's sparse_feature_ids looking for a path to a node whose kind is public_error_node._graph_analysis_for_replay is the traversal gate. It reads graph_nodes and graph_edges, checks that every edge endpoint resolves to a declared node, then does a depth-first walk from the row's sparse_feature_ids looking for a path to a node whose kind is public_error_node. A row with no such path gets INTERPRETABILITY_GRAPH_PATH_REQUIRED. _weight_sequence_analysis collects edge weights column by column across rows and calls _constant_delta_sequence; a column of four or more weights with one constant nonzero step is flagged as a decorative arithmetic sequence, not real circuit evidence.
\n
_toy_transformer_attribution_runtime is the recomputation gate. It reads token_ids, embeddings, layer1, layer2, and target_logit_index from the fixture._toy_transformer_attribution_runtime is the recomputation gate. It reads token_ids, embeddings, layer1, layer2, and target_logit_index from the fixture. _toy_transformer_forward averages the token embeddings into a context vector, applies layer1, applies a tanh hidden activation, applies layer2, and reads the target logit. The runtime then computes gradient scores using the analytic tanh derivative 1 - h^2, so the attribution rests on the same forward pass, and ablates each hidden feature in turn to measure its logit delta. It picks the top feature by attribution and by ablation and compares both against the fixture's declared winners. A mismatch emits INTERPRETABILITY_TOY_TRANSFORMER_DECLARED_TOP_FEATURE_MISMATCH. It also records a weight_digest so a cached or exported card can prove which fixture basis it is coupled to. If the spec came from the internal default rather than the fixture, input_coupled_fixture is false and _build_result raises INTERPRETABILITY_TOY_TRANSFORMER_FIXTURE_SPEC_REQUIRED.
\n
On the current fixture the target logit is 0.044176, and both the gradient attribution and the ablation delta select toy_hidden_feature_1, which matches the declaration. On the current fixture the target logit is 0.044176, and both the gradient attribution and the ablation delta select toy_hidden_feature_1, which matches the declaration. The row passes because two independent recomputed paths agree with each other and with the fixture.
\n
_source_module_manifest_result is the boundary gate. For each row in the source-module manifest it checks that the classification is copied_non_secret_macro_body, that the material class is one of the allowed public source classes, that the body was copied without any result record body text, that the target file exists and its digest matches target_sha256 and source_sha256, that sha256_match is true, and that every required anchor appears in the target text._source_module_manifest_result is the boundary gate. For each row in the source-module manifest it checks that the classification is copied_non_secret_macro_body, that the material class is one of the allowed public source classes, that the body was copied without any result record body text, that the target file exists and its digest matches target_sha256 and source_sha256, that sha256_match is true, and that every required anchor appears in the target text. _source_open_body_import_summary folds the verified modules into a summary whose scope limit flags are all false. _write_receipts emits the records with body_in_receipt set to false throughout, and result_card omits the large payload keys listed in CARD_OMITTED_FULL_PAYLOAD_KEYS.
\nDiagram of the mechanism (7 steps).Diagram source & refs
flowchart TD Input["Fixture or exported bundle features, replay rows, toy-transformer spec"] Schema["_replay_policy_findings required fields, refs, forbidden flags"] Graph["_graph_analysis_for_replay feature ids to public error node"] Weights["_weight_sequence_analysis reject arithmetic edge weights"] Toy["_toy_transformer_attribution_runtime forward + gradient + ablation vs declared"] Source["_source_module_manifest_result copied bodies, digest, anchors"] Result["_write_receipts refs, digests, counts, verdicts"] Input --> Schema Input --> Graph Input --> Weights Input --> Toy Input --> Source Schema --> Result Graph --> Result Weights --> Result Toy --> Result Source --> Result
\n
Negative cases
\n
run loads eight named negative fixtures and checks that each produces its expected error code, listed in EXPECTED_NEGATIVE_CASES.run loads eight named negative fixtures and checks that each produces its expected error code, listed in EXPECTED_NEGATIVE_CASES. private_model_weights_export must raise INTERPRETABILITY_PRIVATE_WEIGHTS_FORBIDDEN. raw_activation_dump must raise INTERPRETABILITY_RAW_ACTIVATION_DUMP_FORBIDDEN. proprietary_prompt_export must raise INTERPRETABILITY_PROPRIETARY_PROMPT_FORBIDDEN. hidden_chain_of_thought_export must raise INTERPRETABILITY_HIDDEN_COT_FORBIDDEN. unverifiable_feature_name must raise INTERPRETABILITY_FEATURE_NAME_UNVERIFIABLE. graph_screenshot_without_edges must raise INTERPRETABILITY_MACHINE_READABLE_EDGES_REQUIRED. transparency_claim_without_intervention must raise INTERPRETABILITY_INTERVENTION_RECEIPT_REQUIRED. faithfulness_claim_without_sufficiency_limit must raise INTERPRETABILITY_FAITHFULNESS_REQUIRES_LIMITS. A case that fails to raise its code leaves an entry in expected_missing, and the overall status is then blocked.
\n
Prior Art Grounding
\n
The accounting shape borrows from the mechanistic interpretability tradition, the study of the internal circuits of neural networks (Anthropic, Transformer Circuits): circuits should be graph-structured, features should be identifiable, causal language should be backed by interventions, and faithfulness language should be bounded. The accounting shape borrows from the mechanistic interpretability tradition, the study of the internal circuits of neural networks (Anthropic, Transformer Circuits): circuits should be graph-structured, features should be identifiable, causal language should be backed by interventions, and faithfulness language should be bounded. This component does not reproduce those methods. It puts a replay boundary around an interpretability-shaped fixture: machine-readable edges instead of screenshots, causal-intervention refs instead of bare transparency language, and fixture recomputation instead of trusting a stale row. The result is fixture-bound runtime evidence, not live model access.
The focused test exercises the fixture and bundle: recomputation, stale-row rejection, the graph and weight gates, source-body import checks, and cached-card reuse. The focused test exercises the fixture and bundle: recomputation, stale-row rejection, the graph and weight gates, source-body import checks, and cached-card reuse. A pass is reader-verifiable evidence only. It excludes launch, external model access, source-file changes, or whole-system correctness.
\n
Scope boundary
\n
Scope limit
\n
The bounded attribution-replay claim
\n
The strongest claim the evidence supports is a public, cold-replayable circuit-attribution accounting for the named fixture and exported bundle. The strongest claim the evidence supports is a public, cold-replayable circuit-attribution accounting for the named fixture and exported bundle. Feature ids are tied to machine-readable graph edges with a traversable path to a public error node. Transparency and faithfulness language passes only when it carries a causal-intervention result record ref and an explicit limit ref. The toy-transformer declaration is coupled to recomputed forward, gradient, and ablation results, and a stale declaration is rejected. Copied source bodies are verified by manifest and digest checks while the records stay metadata-only.
\n
Where the replay proof boundary stops
\n
The proof boundary is that fixture and bundle. This component does not access a live model, does not export private weights, raw activations, proprietary prompts, or hidden reasoning, and is not a model-transparency product. The proof boundary is that fixture and bundle. This component does not access a live model, does not export private weights, raw activations, proprietary prompts, or hidden reasoning, and is not a model-transparency product. It carries no benchmark, public sharing, hosting, or launch-scope decision. The scope limit is a public circuit-attribution runtime record and copied source-body evidence, nothing wider.
\n",
"components": [
{
"anchor": "component-mechanistic_interpretability_circuit_attribution_replay",
"component_id": "mechanistic_interpretability_circuit_attribution_replay",
"public_label": "Toy-Transformer Attribution Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 14,
"links": [
{
"id": "mechanistic_interpretability_circuit_attribution_replay",
"kind": "organ",
"kind_label": "component",
"label": "Toy-Transformer Attribution Replay",
"relation": "explains",
"url": "component-mechanistic-interpretability-circuit-attribution-replay.html"
},
{
"id": "mechanism.mechanistic_interpretability_circuit_attribution_replay.validates_public_mechanistic_interpretability_circuit_attribution_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public mechanistic interpretability circuit attribution replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-mechanistic-interpretability-circuit-attribution-replay-validates-public-mechanistic-interpretability-circuit-attribution-replay"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-4",
"kind": "principle",
"kind_label": "principle",
"label": "Possession is not permission",
"relation": "governed by",
"url": "doctrine.html#dcard-p-4"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "AX-3",
"kind": "axiom",
"kind_label": "axiom",
"label": "Authority by derivation, not possession",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-3"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
},
{
"id": "paper_module.research_replication_rubric_artifact_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Research Replication Rubric Artifact Replay",
"relation": "depends on",
"url": "paper-module-research-replication-rubric-artifact-replay.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
},
{
"id": "paper_module.public_reveal_walkthrough",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Public Reveal Walkthrough",
"relation": "depends on",
"url": "paper-module-public-reveal-walkthrough.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.mechanistic_interpretability_circuit_attribution_replay",
"reader_packet": {
"abstract": "Mechanistic interpretability replay validates public circuit-attribution result record contracts without live model access or private activation export.",
"authority_ceiling": "Declared public circuit-attribution runtime result record and source-module import evidence only; no live model access, model-transparency product claim, export of private weights, raw activations, proprietary prompts, hidden reasoning, external model access, benchmark claims, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.mechanistic_interpretability_circuit_attribution_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-mechanistic-interpretability-circuit-attribution-replay-diagram-1",
"html_url": "docs/paper-module-mechanistic-interpretability-circuit-attribution-replay.html#paper-module-mechanistic-interpretability-circuit-attribution-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Input[\"Fixture or exported bundle features, replay rows, toy-transformer spec\"] Schema[\"_replay_policy_findings required fields, refs, forbidden flags\"] Graph[\"_graph_analysis_for_replay feature ids to public error node\"] Weights[\"_weight_sequence_analysis reject arithmetic edge weights\"] Toy[\"_toy_transformer_attribution_runtime forward + gradient + ablation vs declared\"] Source[\"_source_module_manifest_result copied bodies, digest, anchors\"] Result[\"_write_receipts refs, digests, counts, verdicts\"] Input --> Schema Input --> Graph Input --> Weights Input --> Toy Input --> Source Schema --> Result Graph --> Result Weights --> Result Toy --> Result Source --> Result",
"source_format": "mermaid",
"source_hash": "sha256:b0931024e99237145188e40bb3374a6cc073bc8dfd3a2952477878f5813a42e2",
"source_line_end": 87,
"source_line_start": 67
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-3",
"kind": "axiom",
"kind_label": "axiom",
"label": "Authority by derivation, not possession",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-3"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.research_replication_rubric_artifact_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Research Replication Rubric Artifact Replay",
"relation": "depends on",
"url": "paper-module-research-replication-rubric-artifact-replay.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
},
{
"id": "paper_module.public_reveal_walkthrough",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Public Reveal Walkthrough",
"relation": "depends on",
"url": "paper-module-public-reveal-walkthrough.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "mechanistic_interpretability_circuit_attribution_replay",
"kind": "organ",
"kind_label": "component",
"label": "Toy-Transformer Attribution Replay",
"relation": "explains",
"url": "component-mechanistic-interpretability-circuit-attribution-replay.html"
},
{
"id": "mechanism.mechanistic_interpretability_circuit_attribution_replay.validates_public_mechanistic_interpretability_circuit_attribution_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public mechanistic interpretability circuit attribution replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-mechanistic-interpretability-circuit-attribution-replay-validates-public-mechanistic-interpretability-circuit-attribution-replay"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-4",
"kind": "principle",
"kind_label": "principle",
"label": "Possession is not permission",
"relation": "governed by",
"url": "doctrine.html#dcard-p-4"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.mechanistic_interpretability_circuit_attribution_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 129,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/mechanistic_interpretability_circuit_attribution_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Anthropic, Transformer Circuits",
"section_id": "the-accounting-shape-borrows-from-the-mechanistic-interpretability-tradition",
"url": "https://transformer-circuits.pub/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "Interpretability writing is easy to overstate. A named feature can read like understanding. Interpretability writing is easy to overstate. A named feature can read like understanding. A graph picture can read like a discovered circuit. A small local script can read like access to a real model. This component exists to hold one kind of claim to a checkable size.",
"level": 4,
"order": 1,
"section_id": "interpretability-writing-is-easy-to-overstate",
"source_line_end": 7,
"source_line_start": 3,
"title": "Interpretability writing is easy to overstate"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It answers one question. Before the system lets a circuit-attribution story stand as public evidence, does the story survive a deterministic replay instead of being taken on trust? It answers one question. Before the system lets a circuit-attribution story stand as public evidence, does the story survive a deterministic replay instead of being taken on trust? The runtime lives in src/microcosmcore/components/mechanisticinterpretabilitycircuitattribution_replay.py. It reads a fixture or an exported bundle, recomputes a tiny toy transformer, checks the graph and source-body evidence, and writes result records that carry refs, digests, counts, and verdicts. They carry no model weights, no raw\u2026",
"level": 4,
"order": 2,
"section_id": "it-answers-one-question",
"source_line_end": 12,
"source_line_start": 8,
"title": "It answers one question"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The current fixture has six sparse features and six replay rows. A pass is bounded evidence about that fixture. The current fixture has six sparse features and six replay rows. A pass is bounded evidence about that fixture. It is not access to any trained model.",
"level": 4,
"order": 3,
"section_id": "the-current-fixture-has-six-sparse-features-and-six-replay-rows",
"source_line_end": 17,
"source_line_start": 13,
"title": "The current fixture has six sparse features and six replay rows"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 19,
"source_line_start": 18,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The problem is that an interpretability-shaped fixture can look convincing while being wrong or stale. A feature label can lack a catalog id. The problem is that an interpretability-shaped fixture can look convincing while being wrong or stale. A feature label can lack a catalog id. A graph can be a screenshot with no traversable edges. A row can claim transparency with no intervention result record behind it. And an attribution claim that was once true of its inputs can stay in place after the inputs change.",
"level": 4,
"order": 5,
"section_id": "the-problem-is-that-an-interpretability-shaped-fixture-can-look-convincing-while-being-wrong-or-stale",
"source_line_end": 24,
"source_line_start": 20,
"title": "The problem is that an interpretability-shaped fixture can look convincing while being wrong or stale"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A loose check passes all of these. This component makes the claim smaller and forces each claim-bearing row across four gates: a replay-schema gate, a graph-traversal gate, a toy-transformer recomputation gate, and a source-body boundary gate. A loose check passes all of these. This component makes the claim smaller and forces each claim-bearing row across four gates: a replay-schema gate, a graph-traversal gate, a toy-transformer recomputation gate, and a source-body boundary gate. A row passes only when all four hold.",
"level": 4,
"order": 6,
"section_id": "a-loose-check-passes-all-of-these",
"source_line_end": 29,
"source_line_start": 25,
"title": "A loose check passes all of these"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 31,
"source_line_start": 30,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Two entry points drive the same gates. run validates the first-wave fixture with negative cases included. Two entry points drive the same gates. run validates the first-wave fixture with negative cases included. runattributionbundle validates an exported bundle without the negative-case files, and can reuse a cached bundle result when freshbundlereceipt confirms the cache is still coupled to the current inputs. Both call buildresult, which runs the checks in order and then write_receipts, which writes the result, board, validation, and sign-off records.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-drive-the-same-gates",
"source_line_end": 36,
"source_line_start": 32,
"title": "Two entry points drive the same gates"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult loads the fixture payloads, then for each replay row it calls replaypolicyfindings, graphanalysisforreplay, and a feature-catalog resolution check. buildresult loads the fixture payloads, then for each replay row it calls replaypolicyfindings, graphanalysisforreplay, and a feature-catalog resolution check. After the loop it runs weightsequenceanalysis once across all rows, toytransformerattributionruntime once, and scan_paths over the inputs and copied source modules. The overall status is pass only when every gate passes and the secret scan passes.",
"level": 4,
"order": 9,
"section_id": "build-result-loads-the-fixture-payloads",
"source_line_end": 41,
"source_line_start": 37,
"title": "`_build_result` loads the fixture payloads"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "replaypolicyfindings is the schema gate. It requires all 29 fields in REQUIREDREPLAYFIELDS, a toy prompt ref with no inline prompt body, at least one sparse feature id, and a replacementmodelapproximationscore inside the range 0 to 1.** replaypolicyfindings is the schema gate. It requires all 29 fields in REQUIREDREPLAYFIELDS, a toy prompt ref with no inline prompt body, at least one sparse feature id, and a replacementmodelapproximationscore inside the range 0 to 1. It emits INTERPRETABILITYINTERVENTIONRECEIPTREQUIRED when a row claims transparency without a causalinterventionreceiptref, and INTERPRETABILITYFAITHFULNESSREQUIRESLIMITS when a row claims faithfulness without a faithfulnesslim\u2026",
"level": 4,
"order": 10,
"section_id": "replay-policy-findings-is-the-schema-gate",
"source_line_end": 46,
"source_line_start": 42,
"title": "`_replay_policy_findings` is the schema gate"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "graphanalysisforreplay is the traversal gate. It reads graphnodes and graphedges, checks that every edge endpoint resolves to a declared node, then does a depth-first walk from the row's sparsefeatureids looking for a path to a node whose kind is publicerrornode.** graphanalysisforreplay is the traversal gate. It reads graphnodes and graphedges, checks that every edge endpoint resolves to a declared node, then does a depth-first walk from the row's sparsefeatureids looking for a path to a node whose kind is publicerrornode. A row with no such path gets INTERPRETABILITYGRAPHPATHREQUIRED. weightsequenceanalysis collects edge weights column by column across rows and calls constantdelta_sequenc\u2026",
"level": 4,
"order": 11,
"section_id": "graph-analysis-for-replay-is-the-traversal-gate",
"source_line_end": 51,
"source_line_start": 47,
"title": "`_graph_analysis_for_replay` is the traversal gate"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "toytransformerattributionruntime is the recomputation gate. It reads tokenids, embeddings, layer1, layer2, and targetlogitindex from the fixture. toytransformerattributionruntime is the recomputation gate. It reads tokenids, embeddings, layer1, layer2, and targetlogitindex from the fixture. toytransformerforward averages the token embeddings into a context vector, applies layer1, applies a tanh hidden activation, applies layer2, and reads the target logit. The runtime then computes gradient scores using the analytic tanh derivative 1 - h^2, so the attribution rests on the same forward pass, and ablates each hidden feature in turn to measure its logit delta. It picks the top feature by attri\u2026",
"level": 4,
"order": 12,
"section_id": "toy-transformer-attribution-runtime-is-the-recomputation-gate",
"source_line_end": 56,
"source_line_start": 52,
"title": "`_toy_transformer_attribution_runtime` is the recomputation gate"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "On the current fixture the target logit is 0.044176, and both the gradient attribution and the ablation delta select toyhiddenfeature1, which matches the declaration.** On the current fixture the target logit is 0.044176, and both the gradient attribution and the ablation delta select toyhiddenfeature_1, which matches the declaration. The row passes because two independent recomputed paths agree with each other and with the fixture.",
"level": 4,
"order": 13,
"section_id": "on-the-current-fixture-the-target-logit-is-0-044176",
"source_line_end": 61,
"source_line_start": 57,
"title": "On the current fixture the target logit is `0.044176`"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "sourcemodulemanifestresult is the boundary gate. For each row in the source-module manifest it checks that the classification is copiednonsecretmacrobody, that the material class is one of the allowed public source classes, that the body was copied without any result record body text, that the target file exists and its digest matches targetsha256 and sourcesha256, that sha256match is true, and that every required anchor appears in the target text. sourcemodulemanifestresult is the boundary gate. For each row in the source-module manifest it checks that the classification is copiednonsecretmacrobody, that the material class is one of the allowed public source classes, that the body was copi\u2026",
"level": 4,
"order": 14,
"section_id": "source-module-manifest-result-is-the-boundary-gate",
"source_line_end": 88,
"source_line_start": 62,
"title": "`_source_module_manifest_result` is the boundary gate"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "negative-cases",
"source_line_end": 90,
"source_line_start": 89,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run loads eight named negative fixtures and checks that each produces its expected error code, listed in EXPECTEDNEGATIVECASES. run loads eight named negative fixtures and checks that each produces its expected error code, listed in EXPECTEDNEGATIVECASES. privatemodelweightsexport must raise INTERPRETABILITYPRIVATEWEIGHTSFORBIDDEN. rawactivationdump must raise INTERPRETABILITYRAWACTIVATIONDUMPFORBIDDEN. proprietarypromptexport must raise INTERPRETABILITYPROPRIETARYPROMPTFORBIDDEN. hiddenchainofthoughtexport must raise INTERPRETABILITYHIDDENCOTFORBIDDEN. unverifiablefeaturename must raise INTERPRETABILITYFEATURENAMEUNVERIFIABLE. graphscreenshotwithoutedges must raise INTERPRETABILITYMACHINER\u2026",
"level": 4,
"order": 16,
"section_id": "run-loads-eight-named-negative-fixtures",
"source_line_end": 95,
"source_line_start": 91,
"title": "`run` loads eight named negative fixtures"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 97,
"source_line_start": 96,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The accounting shape borrows from the mechanistic interpretability tradition, the study of the internal circuits of neural networks (Anthropic, Transformer Circuits): circuits should be graph-structured, features should be identifiable, causal language should be backed by interventions, and faithfulness language should be bounded. The accounting shape borrows from the mechanistic interpretability tradition, the study of the internal circuits of neural networks (Anthropic, Transformer Circuits): circuits should be graph-structured, features should be identifiable, causal language should be backed by interventions, and faithfulness language should be bounded. This component does not reproduce\u2026",
"level": 4,
"order": 18,
"section_id": "the-accounting-shape-borrows-from-the-mechanistic-interpretability-tradition",
"source_line_end": 102,
"source_line_start": 98,
"title": "The accounting shape borrows from the mechanistic interpretability tradition"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from the microcosm-substrate/ public root:",
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 111,
"source_line_start": 103,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The focused test exercises the fixture and bundle: recomputation, stale-row rejection, the graph and weight gates, source-body import checks, and cached-card reuse. The focused test exercises the fixture and bundle: recomputation, stale-row rejection, the graph and weight gates, source-body import checks, and cached-card reuse. A pass is reader-verifiable evidence only. It excludes launch, external model access, source-file changes, or whole-system correctness.",
"level": 4,
"order": 20,
"section_id": "the-focused-test-exercises-the-fixture-and-bundle",
"source_line_end": 116,
"source_line_start": 112,
"title": "The focused test exercises the fixture and bundle"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 118,
"source_line_start": 117,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim the evidence supports is a public, cold-replayable circuit-attribution accounting for the named fixture and exported bundle. The strongest claim the evidence supports is a public, cold-replayable circuit-attribution accounting for the named fixture and exported bundle. Feature ids are tied to machine-readable graph edges with a traversable path to a public error node. Transparency and faithfulness language passes only when it carries a causal-intervention result record ref and an explicit limit ref. The toy-transformer declaration is coupled to recomputed forward, gradient, and ablation results, and a stale declaration is rejected. Copied source bodies are verified by ma\u2026",
"level": 4,
"order": 22,
"section_id": "the-bounded-attribution-replay-claim",
"source_line_end": 123,
"source_line_start": 119,
"title": "The bounded attribution-replay claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The proof boundary is that fixture and bundle. This component does not access a live model, does not export private weights, raw activations, proprietary prompts, or hidden reasoning, and is not a model-transparency product. The proof boundary is that fixture and bundle. This component does not access a live model, does not export private weights, raw activations, proprietary prompts, or hidden reasoning, and is not a model-transparency product. It carries no benchmark, public sharing, hosting, or launch-scope decision. The scope limit is a public circuit-attribution runtime record and copied source-body evidence, nothing wider.",
"level": 4,
"order": 23,
"section_id": "where-the-replay-proof-boundary-stops",
"source_line_end": 128,
"source_line_start": 124,
"title": "Where the replay proof boundary stops"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/mechanistic_interpretability_circuit_attribution_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Mechanistic interpretability replay validates public circuit-attribution result record contracts without live model access or private activation export."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "mechanistic_interpretability_circuit_attribution_replay.py",
"path": "src/microcosm_core/organs/mechanistic_interpretability_circuit_attribution_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public circuit-attribution replay validation, source-module digest checks, secret-exclusion policy, negative cases, and scope limits.",
"symbols": [
"run",
"run_attribution_bundle",
"_replay_policy_findings",
"_source_module_manifest_result",
"_source_open_body_import_summary",
"_build_result",
"_write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/mechanistic_interpretability_circuit_attribution_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/mechanistic_interpretability_circuit_attribution_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/mechanistic_interpretability_circuit_attribution_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Mechanistic Interpretability Circuit Attribution Replay checks that public circuit-attribution replay rows carry feature-to-edge links, intervention deltas, sufficiency and faithfulness limits, target/source refs, source-module digest evidence, secret-exclusion scans, negative cases, and scope limits. It validates the result record contract only, not live model transparency, private weights, raw activations, proprietary prompts, hidden reasoning, external model access, benchmark claims, launch, public sharing, or whole-system correctness.",
"summary_line": "Mechanistic interpretability replay validates public circuit-attribution result record contracts without live model access or private activation export.",
"title": "Mechanistic Interpretability Circuit Attribution Replay"
},
{
"anchor": "paper-module-spatial-world-model-counterfactual-simulation-replay",
"authority_ceiling": "Declared public synthetic spatial counterfactual-replay metadata and source-module import evidence only; no robot or AV operation, real-world geographic accuracy, simulator product validation, generated-video authority, benchmark claims, external model access, hosting, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
It asks whether a declared spatial counterfactual row binds a scene, event, and predicted outcome that survive independent recomputation, rather than a shape that merely looks right. Spatial world-model demos are easy to oversell. A plausible video, or a row that simply asserts "the model predicted the next state correctly", can pass for understanding without anything being checked. This component answers one narrow question: does a declared spatial counterfactual row bind a source scene, an event, and a predicted outcome that survive an independent recomputation, or is it only a shape that looks right?
\n
The runtime distrusts the predicted state
\n
The predicted actor count, transition delta, event label, and spawn cells are derived from the inputs, so a stale or hand-edited prediction stops matching and the row blocks.
\n
The result is a metadata-only record of refs, digests, and counts over synthetic metadata, running no robot, no vehicle, and no real map. The result is a metadata-only record: refs, digests, counts, and verdicts, with no video bodies, raw sensor payloads, or location data. It checks synthetic metadata. It does not run a robot, drive a vehicle, or measure a real map.
\n
Purpose
\n
Synthetic robotics and self-driving scenes
\n
The input is synthetic robotics and self-driving scenes, each with an action and a declared next scene, from occluded forklifts to late-turning oncoming vehicles. The input is a set of synthetic scenes drawn from robotics and self-driving settings: a forklift entering an aisle from behind an occlusion, a small pedestrian stepping into a crosswalk, a gust pushing a drone off a corridor, a reflective floor read as free space, a pallet stack sliding into a lane, and an oncoming vehicle accelerating late into a turn. Each scene has an action and a declared next scene.
\n
Scenes force state-difference accounting
\n
A loose check would just accept the declared next scene, so the scenes exist to force state-difference accounting rather than to approximate the real world. A loose check would read the declared next scene and accept it. That is exactly the failure the component avoids, because a declared prediction can be edited to say anything. The scenes exist to force state-difference accounting, not because the gridworld approximates the real world.
\n
How it works
\n
Fixture and bundle entry points
\n
run validates the fixture with eight negative cases and run_simulation_bundle the exported bundle without them, both calling _build_result then _write_receipts. Two entry points drive the same pipeline. run validates the fixture with the eight negative cases included. run_simulation_bundle validates an exported bundle with negatives excluded, and can reuse a prior result when the input digest is unchanged. Both call _build_result, then _write_receipts.
\n
Six inputs through the ordered check stages
\n
_build_result loads six input files, checks the protocol and metadata-only policy, then runs the replay, recomputation, source-module, and private-string stages, passing only when all hold._build_result loads six input files (a protocol, a replay policy, scene states, counterfactual replays, state transitions, and a source-module manifest). It checks that the protocol selects this component and that the replay policy declares the metadata-only scope limit, tested by _required_policy_ok, which requires metadata_projection_only true and every other ceiling flag false. It runs each replay row through _replay_policy_findings, runs the recomputation in _state_transition_analysis, checks the copied source modules in _source_module_manifest_result, and scans every input and source path for private strings. The run passes only if all of those hold.
\n
Required fields and forbidden payloads per row
\n
_replay_policy_findings requires each row's declared refs and labels and emits a typed error for any forbidden flag, private-video body, raw-sensor payload, or non-public paths._replay_policy_findings checks one replay row. It requires the declared fields (scene ref, action-trace ref, predicted-state ref, transition-diff ref, oracle-check ref, sensor-packet refs, limitation labels, and the payload boundary). It emits a typed error when a forbidden flag is set to true, for example SPATIAL_PRIVATE_VIDEO_FORBIDDEN or SPATIAL_LIVE_OPERATION_FORBIDDEN, and it rejects any row that carries a private-video body, a raw-sensor payload, or a non-public paths string.
\n
Recomputing the gridworld step
\n
_state_transition_analysis builds an 8 by 8 gridworld from the actor count and steps it with a deterministic event action, computing an input-driven actor-count delta rather than copying the expected row._state_transition_analysis is the recomputation. For each replay it resolves exactly one matching transition row, builds an 8 by 8 gridworld from the source scene's actor count with _gridworld_initial_state, then steps it with _gridworld_step. The step maps the event label to a deterministic action in EVENT_GRIDWORLD_ACTIONS, computes an actor-count delta with _gridworld_actor_count_delta, and places new actors with _gridworld_spawn_cells. The delta is input-driven, not copied from the expected row:
_gridworld_spawn_cells hashes the inputs to place actors, then the analysis compares every declared field against what the gridworld produced, and one mismatch blocks the whole result. Spawn cells are also input-derived. _gridworld_spawn_cells hashes the event, replay id, scene ref, topology ref, sensor-packet refs, consistency budget, limitation labels, and source actor count, then walks the bounded grid from the declared event cell. The analysis then compares the declared predicted actor count, transition delta, event label, spawn cell or cells, and the three refs against what the gridworld actually produced. Any mismatch records SPATIAL_STATE_TRANSITION_SIMULATION_MISMATCH on that row, and one bad row blocks the whole result.
\n
Copied source bodies verified by digest
\n
_source_module_manifest_result verifies each copied module by digest, anchors, and line count, rejecting any manifest that places body text into a result._source_module_manifest_result handles the copied source bodies the exported bundle may carry. It verifies each module by its sha256 digest, its required text anchors, and its line count against the file on disk. It rejects any manifest that tries to place body text into a result, emitting SPATIAL_SOURCE_BODY_TEXT_IN_RECEIPT_FORBIDDEN or SPATIAL_SOURCE_MODULE_BODY_TEXT_IN_RECEIPT_FORBIDDEN. _source_open_body_import_summary folds the verified modules into a metadata-only summary of refs, digests, and counts.
\n
Records and the compact result card
\n
_write_receipts writes the result, board, validation, and sign-off records, and result_card projects a compact card of status, counts, and command without the bulky payload._write_receipts writes the result, board, validation, and sign-off records and returns the result with the board attached. result_card projects the result into a compact card that omits the bulky payload keys and keeps the status, counts, and command.
\nDiagram of the mechanism (9 steps).Diagram source & refs
flowchart TD Scene["Scene-state row actor count + topology"] --> Step["Deterministic gridworld step 8x8 bounded recomputation"] Replay["Counterfactual replay row event + sensor refs + budget"] --> Step Step --> Actual["Actual next state actor delta + spawn cells"] Replay --> Declared["Declared predicted state transition diff + oracle check"] Actual --> Compare{"Actual matches declared?"} Declared --> Compare Compare -->|yes| Result record["metadata-only pass record counts + refs + digests"] Compare -->|no| Block["Typed mismatch, blocked"] Replay --> Boundary{"Forbidden payload or claim?"} Boundary -->|yes| Block
\n
Negative cases
\n
Eight rejections each bound to a code
\n
EXPECTED_NEGATIVE_CASES names eight rows that must be rejected with their codes, and a fixture run passes only when every expected code is observed.EXPECTED_NEGATIVE_CASES names eight rows that must be rejected, each bound to its error code: private video export, real-world location claim, live robot or AV operation, raw sensor export, simulator-product claim, generated-video-only authority, geographic-accuracy claim, and benchmark claims without a state diff (SPATIAL_BENCHMARK_SCORE_REQUIRES_STATE_DIFF). A fixture run passes only when every expected code is observed. If a negative row slipped through, its code would be missing from the observed set and the run would block.
\n
Prior Art Grounding
\n
Borrowed from the world-models line
\n
The replay borrows the counterfactual-rollout shape from the World Models line over synthetic metadata, while the gridworld here is a hand-written toy, not a learned model. This replay exercises a spatial world model under counterfactual interventions. It draws on the world-models line of work (Ha and Schmidhuber, World Models), where an agent learns a compressed model of its environment and rolls it forward under hypothetical actions. The component borrows the counterfactual-rollout shape over synthetic metadata. The gridworld here is a hand-written toy, not a learned model.
\n
Validation Result record Path
\n
Run from microcosm-substrate:
\n
What a pass over the fixture proves
\n
A pass means the six predicted states matched the gridworld step, all eight negative codes were observed, the source modules verified by digest, and no input leaked a private string. A pass means the six declared predicted states match the gridworld's actual step, the eight negative codes were all observed, the copied source modules verified by digest, and no input path leaked a private string. These validation result records prove the synthetic replay and source-module import boundary only.
\n
Scope boundary
\n
Scope limit
\n
The fixture-bound strongest claim
\n
The strongest claim is fixture-bound: over public synthetic inputs the component recomputed each state transition and produced the metadata-only records, and recomputation is why the counts are evidence. The strongest claim the evidence supports is fixture-bound: the component ran over public synthetic inputs, recomputed each state transition, and produced the metadata-only records and projections named above, reproducible by the commands on this page. The recomputation is why the counts are evidence rather than assertion.
\n
A toy gridworld, not perception or control
\n
The boundary stops at a toy 8 by 8 gridworld of actor counts and spawn cells, modelling no perception, control, or physics, and authorizing no robot operation, simulator product, or launch. The proof boundary stops there. It is a public metadata consistency check over a toy 8 by 8 gridworld that models actor counts and spawn cells only. It does not model perception, control, physics, sensor calibration, camera or lidar geometry, maps, vehicle dynamics, or human behavior. The scope limit, carried in AUTHORITY_CEILING, allows synthetic scene, action, prediction, diff, and oracle refs, source-open public sensor refs, and copied-body digests. It excludes robot or AV operation, real-world geographic accuracy, a simulator product, generated-video authority, benchmark claims, external model access, hosting, public sharing, launch, or whole-system correctness.
\n",
"components": [
{
"anchor": "component-spatial_world_model_counterfactual_simulation_replay",
"component_id": "spatial_world_model_counterfactual_simulation_replay",
"public_label": "Gridworld Counterfactual State Replay"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 7
},
"link_count": 19,
"links": [
{
"id": "spatial_world_model_counterfactual_simulation_replay",
"kind": "organ",
"kind_label": "component",
"label": "Gridworld Counterfactual State Replay",
"relation": "explains",
"url": "component-spatial-world-model-counterfactual-simulation-replay.html"
},
{
"id": "mechanism.spatial_world_model_counterfactual_simulation_replay.validates_public_spatial_world_model_counterfactual_simulation_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public spatial world model counterfactual simulation replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-spatial-world-model-counterfactual-simulation-replay-validates-public-spatial-world-model-counterfactual-simulation-replay"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
},
{
"id": "paper_module.research_replication_rubric_artifact_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Research Replication Rubric Artifact Replay",
"relation": "depends on",
"url": "paper-module-research-replication-rubric-artifact-replay.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
},
{
"id": "paper_module.materials_chemistry_closed_loop_lab_safety_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Materials Chemistry Closed-Loop Lab-Safety Replay",
"relation": "depends on",
"url": "paper-module-materials-chemistry-closed-loop-lab-safety-replay.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.spatial_world_model_counterfactual_simulation_replay",
"reader_packet": {
"abstract": "Spatial world-model replay validates synthetic counterfactual metadata without robot operation, simulator product claims, or real-world accuracy claims.",
"authority_ceiling": "Declared public synthetic spatial counterfactual-replay metadata and source-module import evidence only; no robot or AV operation, real-world geographic accuracy, simulator product validation, generated-video authority, benchmark claims, external model access, hosting, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.spatial_world_model_counterfactual_simulation_replay.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-spatial-world-model-counterfactual-simulation-replay-diagram-1",
"html_url": "docs/paper-module-spatial-world-model-counterfactual-simulation-replay.html#paper-module-spatial-world-model-counterfactual-simulation-replay-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Scene[\"Scene-state row actor count + topology\"] --> Step[\"Deterministic gridworld step 8x8 bounded recomputation\"] Replay[\"Counterfactual replay row event + sensor refs + budget\"] --> Step Step --> Actual[\"Actual next state actor delta + spawn cells\"] Replay --> Declared[\"Declared predicted state transition diff + oracle check\"] Actual --> Compare{\"Actual matches declared?\"} Declared --> Compare Compare -->|yes| Result record[\"metadata-only pass record counts + refs + digests\"] Compare -->|no| Block[\"Typed mismatch, blocked\"] Replay --> Boundary{\"Forbidden payload or claim?\"} Boundary -->|yes| Block",
"source_format": "mermaid",
"source_hash": "sha256:6571f599247951cda300afaafc3699fafaad127a382506901ee24736d2f04486",
"source_line_end": 158,
"source_line_start": 146
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.research_replication_rubric_artifact_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Research Replication Rubric Artifact Replay",
"relation": "depends on",
"url": "paper-module-research-replication-rubric-artifact-replay.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
},
{
"id": "paper_module.materials_chemistry_closed_loop_lab_safety_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Materials Chemistry Closed-Loop Lab-Safety Replay",
"relation": "depends on",
"url": "paper-module-materials-chemistry-closed-loop-lab-safety-replay.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "spatial_world_model_counterfactual_simulation_replay",
"kind": "organ",
"kind_label": "component",
"label": "Gridworld Counterfactual State Replay",
"relation": "explains",
"url": "component-spatial-world-model-counterfactual-simulation-replay.html"
},
{
"id": "mechanism.spatial_world_model_counterfactual_simulation_replay.validates_public_spatial_world_model_counterfactual_simulation_replay",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public spatial world model counterfactual simulation replay",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-spatial-world-model-counterfactual-simulation-replay-validates-public-spatial-world-model-counterfactual-simulation-replay"
}
],
"relation": "explains"
},
{
"count": 8,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.spatial_world_model_counterfactual_simulation_replay",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 241,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/spatial_world_model_counterfactual_simulation_replay.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Ha and Schmidhuber, World Models",
"section_id": "borrowed-from-the-world-models-line",
"url": "https://arxiv.org/abs/1803.10122"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It asks whether a declared spatial counterfactual row binds a scene, event, and predicted outcome that survive independent recomputation, rather than a shape that merely looks right.** Spatial world-model demos are easy to oversell. A plausible video, or a row that simply asserts \"the model predicted the next state correctly\", can pass for understanding without anything being checked. This component answers one narrow question: does a declared spatial counterfactual row bind a source scene, an event, and a predicted outcome that survive an independent recomputation, or is it only a shape that looks right?",
"level": 4,
"order": 1,
"section_id": "it-asks-whether-a-declared-spatial-counterfactual-row-binds-a-scene",
"source_line_end": 13,
"source_line_start": 3,
"title": "It asks whether a declared spatial counterfactual row binds a scene"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "from the scene and event and stepping it itself, so a stale or hand-edited prediction stops matching and blocks.** The unusual part is that the runtime does not trust the predicted state in the input. event, steps it forward itself, and compares its own result against what the row claims. The predicted actor count, transition delta, event label, and spawn cells are derived from the inputs, so a stale or hand-edited prediction stops matching and the row blocks.",
"level": 4,
"order": 2,
"section_id": "the-runtime-distrusts-the-predicted-state",
"source_line_end": 24,
"source_line_start": 14,
"title": "The runtime distrusts the predicted state"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The result is a metadata-only record of refs, digests, and counts over synthetic metadata, running no robot, no vehicle, and no real map.** The result is a metadata-only record: refs, digests, counts, and verdicts, with no video bodies, raw sensor payloads, or location data. It checks synthetic metadata. It does not run a robot, drive a vehicle, or measure a real map.",
"level": 4,
"order": 3,
"section_id": "the-result-is-a-body-free-record-of-refs",
"source_line_end": 31,
"source_line_start": 25,
"title": "The result is a metadata-only record of refs"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 33,
"source_line_start": 32,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The input is synthetic robotics and self-driving scenes, each with an action and a declared next scene, from occluded forklifts to late-turning oncoming vehicles.** The input is a set of synthetic scenes drawn from robotics and self-driving settings: a forklift entering an aisle from behind an occlusion, a small pedestrian stepping into a crosswalk, a gust pushing a drone off a corridor, a reflective floor read as free space, a pallet stack sliding into a lane, and an oncoming vehicle accelerating late into a turn. Each scene has an action and a declared next scene.",
"level": 4,
"order": 5,
"section_id": "synthetic-robotics-and-self-driving-scenes",
"source_line_end": 44,
"source_line_start": 34,
"title": "Synthetic robotics and self-driving scenes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A loose check would just accept the declared next scene, so the scenes exist to force state-difference accounting rather than to approximate the real world.** A loose check would read the declared next scene and accept it. That is exactly the failure the component avoids, because a declared prediction can be edited to say anything. The scenes exist to force state-difference accounting, not because the gridworld approximates the real world.",
"level": 4,
"order": 6,
"section_id": "scenes-force-state-difference-accounting",
"source_line_end": 52,
"source_line_start": 45,
"title": "Scenes force state-difference accounting"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 54,
"source_line_start": 53,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run validates the fixture with eight negative cases and runsimulation_bundle the exported bundle without them, both calling buildresult then writereceipts.** Two entry points drive the same pipeline. run validates the fixture with the eight negative cases included. runsimulationbundle validates an exported bundle with negatives excluded, and can reuse a prior result when the input digest is unchanged. Both call buildresult, then writereceipts.",
"level": 4,
"order": 8,
"section_id": "fixture-and-bundle-entry-points",
"source_line_end": 63,
"source_line_start": 55,
"title": "Fixture and bundle entry points"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildresult loads six input files, checks the protocol and metadata-only policy, then runs the replay, recomputation, source-module, and private-string stages, passing only when all hold. build_result loads six input files (a protocol, a replay policy, scene states, counterfactual replays, state transitions, and a source-module manifest). It checks that the protocol selects this component and that the replay policy declares the metadata-only scope limit, tested by requiredpolicyok, which requires metadataprojection_only true and every other ceiling flag false. It runs each replay row through replaypolicy_findings, runs the recomputation in statetransition_analysis, checks the copied source\u2026",
"level": 4,
"order": 9,
"section_id": "six-inputs-through-the-ordered-check-stages",
"source_line_end": 77,
"source_line_start": 64,
"title": "Six inputs through the ordered check stages"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "replaypolicy_findings requires each row's declared refs and labels and emits a typed error for any forbidden flag, private-video body, raw-sensor payload, or non-public paths. replaypolicyfindings checks one replay row. It requires the declared fields (scene ref, action-trace ref, predicted-state ref, transition-diff ref, oracle-check ref, sensor-packet refs, limitation labels, and the payload boundary). It emits a typed error when a forbidden flag is set to true, for example SPATIALPRIVATEVIDEOFORBIDDEN or SPATIALLIVEOPERATIONFORBIDDEN, and it rejects any row that carries a private-video body, a raw-sensor payload, or a non-public paths string.",
"level": 4,
"order": 10,
"section_id": "required-fields-and-forbidden-payloads-per-row",
"source_line_end": 89,
"source_line_start": 78,
"title": "Required fields and forbidden payloads per row"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "statetransition_analysis builds an 8 by 8 gridworld from the actor count and steps it with a deterministic event action, computing an input-driven actor-count delta rather than copying the expected row. statetransitionanalysis is the recomputation. For each replay it resolves exactly one matching transition row, builds an 8 by 8 gridworld from the source scene's actor count with gridworldinitial_state, then steps it with gridworldstep. The step maps the event label to a deterministic action in EVENTGRIDWORLDACTIONS, computes an actor-count delta with gridworldactorcountdelta, and places new actors with gridworldspawn_cells. The delta is input-driven, not copied from the expected row:",
"level": 4,
"order": 11,
"section_id": "recomputing-the-gridworld-step",
"source_line_end": 110,
"source_line_start": 90,
"title": "Recomputing the gridworld step"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "gridworldspawn_cells hashes the inputs to place actors, then the analysis compares every declared field against what the gridworld produced, and one mismatch blocks the whole result. Spawn cells are also input-derived. gridworldspawncells hashes the event, replay id, scene ref, topology ref, sensor-packet refs, consistency budget, limitation labels, and source actor count, then walks the bounded grid from the declared event cell. The analysis then compares the declared predicted actor count, transition delta, event label, spawn cell or cells, and the three refs against what the gridworld actually produced. Any mismatch records SPATIALSTATETRANSITIONSIMULATIONMISMATCH on that row, and one ba\u2026",
"level": 4,
"order": 12,
"section_id": "input-derived-spawn-cells-and-the-comparison",
"source_line_end": 123,
"source_line_start": 111,
"title": "Input-derived spawn cells and the comparison"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourcemodulemanifestresult verifies each copied module by digest, anchors, and line count, rejecting any manifest that places body text into a result.** sourcemodulemanifestresult handles the copied source bodies the exported bundle may carry. It verifies each module by its sha256 digest, its required text anchors, and its line count against the file on disk. It rejects any manifest that tries to place body text into a result, emitting SPATIALSOURCEBODYTEXTINRECEIPTFORBIDDEN or SPATIALSOURCEMODULEBODYTEXTINRECEIPT_FORBIDDEN. sourceopenbodyimport_summary folds the verified modules into a metadata-only summary of refs, digests, and counts.",
"level": 4,
"order": 13,
"section_id": "copied-source-bodies-verified-by-digest",
"source_line_end": 136,
"source_line_start": 124,
"title": "Copied source bodies verified by digest"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "writereceipts writes the result, board, validation, and sign-off records, and result_card projects a compact card of status, counts, and command without the bulky payload. write_receipts writes the result, board, validation, and sign-off records and returns the result with the board attached. result_card projects the result into a compact card that omits the bulky payload keys and keeps the status, counts, and command.",
"level": 4,
"order": 14,
"section_id": "records-and-the-compact-result-card",
"source_line_end": 159,
"source_line_start": 137,
"title": "Records and the compact result card"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "negative-cases",
"source_line_end": 161,
"source_line_start": 160,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVE_CASES names eight rows that must be rejected with their codes, and a fixture run passes only when every expected code is observed.** EXPECTEDNEGATIVECASES names eight rows that must be rejected, each bound to its error code: private video export, real-world location claim, live robot or AV operation, raw sensor export, simulator-product claim, generated-video-only authority, geographic-accuracy claim, and benchmark claims without a state diff (SPATIALBENCHMARKSCOREREQUIRESSTATE_DIFF). A fixture run passes only when every expected code is observed. If a negative row slipped through, its code would be missing from the observed set and the run would block.",
"level": 4,
"order": 16,
"section_id": "eight-rejections-each-bound-to-a-code",
"source_line_end": 173,
"source_line_start": 162,
"title": "Eight rejections each bound to a code"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 175,
"source_line_start": 174,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The replay borrows the counterfactual-rollout shape from the World Models line over synthetic metadata, while the gridworld here is a hand-written toy, not a learned model.** This replay exercises a spatial world model under counterfactual interventions. It draws on the world-models line of work (Ha and Schmidhuber, World Models), where an agent learns a compressed model of its environment and rolls it forward under hypothetical actions. The component borrows the counterfactual-rollout shape over synthetic metadata. The gridworld here is a hand-written toy, not a learned model.",
"level": 4,
"order": 18,
"section_id": "borrowed-from-the-world-models-line",
"source_line_end": 186,
"source_line_start": 176,
"title": "Borrowed from the world-models line"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from microcosm-substrate:",
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 205,
"source_line_start": 187,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the six predicted states matched the gridworld step, all eight negative codes were observed, the source modules verified by digest, and no input leaked a private string.** A pass means the six declared predicted states match the gridworld's actual step, the eight negative codes were all observed, the copied source modules verified by digest, and no input path leaked a private string. These validation result records prove the synthetic replay and source-module import boundary only.",
"level": 4,
"order": 20,
"section_id": "what-a-pass-over-the-fixture-proves",
"source_line_end": 214,
"source_line_start": 206,
"title": "What a pass over the fixture proves"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 216,
"source_line_start": 215,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is fixture-bound: over public synthetic inputs the component recomputed each state transition and produced the metadata-only records, and recomputation is why the counts are evidence.** The strongest claim the evidence supports is fixture-bound: the component ran over public synthetic inputs, recomputed each state transition, and produced the metadata-only records and projections named above, reproducible by the commands on this page. The recomputation is why the counts are evidence rather than assertion.",
"level": 4,
"order": 22,
"section_id": "the-fixture-bound-strongest-claim",
"source_line_end": 226,
"source_line_start": 217,
"title": "The fixture-bound strongest claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The boundary stops at a toy 8 by 8 gridworld of actor counts and spawn cells, modelling no perception, control, or physics, and authorizing no robot operation, simulator product, or launch.** The proof boundary stops there. It is a public metadata consistency check over a toy 8 by 8 gridworld that models actor counts and spawn cells only. It does not model perception, control, physics, sensor calibration, camera or lidar geometry, maps, vehicle dynamics, or human behavior. The scope limit, carried in scope_limit, allows synthetic scene, action, prediction, diff, and oracle refs, source-open public sensor refs, and copied-body digests. It excludes robot or AV operation, real-world geograph\u2026",
"level": 4,
"order": 23,
"section_id": "a-toy-gridworld-not-perception-or-control",
"source_line_end": 240,
"source_line_start": 227,
"title": "A toy gridworld, not perception or control"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/spatial_world_model_counterfactual_simulation_replay.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Spatial world-model replay validates synthetic counterfactual metadata without robot operation, simulator product claims, or real-world accuracy claims."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "spatial_world_model_counterfactual_simulation_replay.py",
"path": "src/microcosm_core/organs/spatial_world_model_counterfactual_simulation_replay.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for synthetic spatial counterfactual replay validation, transition analysis, source-module digest checks, negative cases, and scope limits.",
"symbols": [
"run",
"run_simulation_bundle",
"_state_transition_analysis",
"_replay_policy_findings",
"_source_module_manifest_result",
"_source_open_body_import_summary",
"_build_result",
"_write_receipts"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/spatial_world_model_counterfactual_simulation_replay.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/spatial_world_model_counterfactual_simulation_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/spatial_world_model_counterfactual_simulation_replay.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Spatial World Model Counterfactual Simulation Replay validates public synthetic counterfactual rows: scene-state refs, action traces, predicted-state refs, transition diffs, oracle-check refs, public sensor-packet refs, limitation labels, source-module digest evidence, negative cases, and scope limits. It keeps generated-video and metadata evidence below real-world geographic accuracy, robot or AV operation, simulator product claims, benchmark claims, external model access, hosting, launch, public sharing, and whole-system correctness.",
"summary_line": "Spatial world-model replay validates synthetic counterfactual metadata without robot operation, simulator product claims, or real-world accuracy claims.",
"title": "Spatial World Model Counterfactual Simulation Replay"
},
{
"anchor": "paper-module-prediction-oracle-reconciliation",
"authority_ceiling": "Synthetic invented prediction packet and source-module import evidence only; no forecasting correctness or accuracy, no trading, financial, or investment-related actions, no live market data, no external model access, no prediction public sharing, no performance track record, no non-public data import, no launch-scope decision, no publishing-scope decision, and no whole-system correctness.",
"body_html": "
prediction_oracle_reconciliation audits one synthetic forecast packet on invented fixture data, with no live feed, external model access, trading, or performance claim.prediction_oracle_reconciliation takes one synthetic forecast packet and checks whether its reasoning would survive an honest audit. It runs entirely on invented fixture data. There is no live market feed, no external model access, no trading, and no performance claim.
\n
It answers one narrow question about whether the packet keeps its evidence honest and grading recomputable, emitting metadata-only result records of counts, verdicts, and rejected cases. The component answers one narrow question: does this prediction packet keep its evidence honest and its grading recomputable, or does it cut a corner? It emits a set of result records that name the counts, verdicts, and rejected cases. The packet body, the claimed numbers, and any copied source bodies stay out of those records.
\n
The component distrusts the packet's own numbers, recomputing error and direction from raw prices and rejecting any claimed value that disagrees. The unusual choice is that the component does not trust the numbers the packet reports. For every graded row it recomputes the error and the direction from the raw prices, then rejects any claimed value that disagrees with the recompute.
\n
Purpose
\n
Two quiet ways a forecast is wrong
\n
A forecast can hide a large numeric miss behind a correct direction or lean on evidence that arrived after the outcome, and this component surfaces both on a synthetic packet. A forecast can be wrong in two quiet ways. It can call the direction right and still be far off on the number, so a correct arrow hides a large miss. And it can look accurate only because it used evidence that arrived after the outcome it was meant to predict. A loose checker that reads the packet's own scores back to it catches neither. This component exists to make both failures visible on a synthetic packet before any reasoning is dressed up as a track record.
\n
Every value is a fixture
\n
Prices, targets, branches, and realized values are all fixtures, so any direction hit or numeric miss speaks only about the fixture and the grading mechanics. The prices, targets, branches, and realized values are all fixtures. A direction hit or a numeric miss inside a result record is a statement about the fixture and the grading mechanics, nothing more.
\n
How it works
\n
Entry points and _build_result
\n
run and run_prediction_bundle both call _build_result, which loads the packet, secret-scans the inputs, validates the main packet and each negative case, and folds the parts into one status. The entry points are run for the first-wave fixture and run_prediction_bundle for the exported bundle. Both call _build_result, which loads the packet, runs a secret scan over the inputs through scan_paths, validates the main packet, validates each negative-case input, and folds the parts into one status.
\n
Density and authority checks in validate_reconciliation_packet
\n
validate_reconciliation_packet checks packet density, rejects any forbidden authority flag through _authority_overclaim, then runs and folds five gates.validate_reconciliation_packet is the core. It first checks packet density: at least five pattern ids, four source refs, one projection result record, and three public runtime refs. It calls _authority_overclaim, which raises PREDICTION_TRADING_OR_ADVICE_OVERCLAIM if any flag in FORBIDDEN_AUTHORITY_FLAGS (trading, financial or investment-related actions, live market data, external model access, public sharing, launch, non-public data equivalence) is set true. Then it runs five gates and folds their findings.
\n
CP1 branches must retain the losing side
\n
_validate_cp1 requires every branch to name its chosen side, carry rationale, and keep the record of why the opposite side lost, with equity-lane branches also needing a confirmation bit._validate_cp1 reads cp1_branches. Every branch must name the side it chose, carry rationale refs, and keep an opposite_side_invalidation_ref, the record of why the losing side lost. A branch that asserts a winner without retaining the discarded alternative raises PREDICTION_CP1_BIFURCATION_UNRESOLVED. An equity, market, or finance lane branch also needs an explicit confirmation bit, or it raises PREDICTION_EQUITY_CONFIRMATION_REQUIRED.
\n
CP2 fixes the universe and forbids post-target evidence
\n
_validate_cp2 requires each prediction's target inside the declared universe and every evidence ref to be pre-target, the gate that stops a packet from grading itself with hindsight._validate_cp2 reads cp2_predictions. Each prediction's target_id must sit inside the declared target universe, so the set of things being predicted is fixed before the outcome; a target outside it raises PREDICTION_CP2_TARGET_OUT_OF_UNIVERSE. Every evidence ref must be pre-target: _evidence_is_pre_t accepts a ref only if it carries a T- time prefix, and a ref that points past the target window raises PREDICTION_ORACLE_POST_T_EVIDENCE_FORBIDDEN. This is the gate that stops a packet from grading itself with hindsight.
\n
Recomputing every graded row from raw prices
\n
The numeric functions recompute error and direction from raw prices, contradict any mismatched claim, surface large misses, refuse to grade missing or degraded rows, and preserve the STOCK/ETF split._numeric_grading_inputs and _numeric_reconciliation_rows do the arithmetic. For each graded oracle-diff row the component recomputes the absolute delta as the distance between predicted and realized price, the percent delta against the snapshot price, and the direction hit as predicted direction against realized direction. If the row also reports its own abs_error, pred_error_pct, or direction_hit, the claimed value must match the recompute, or the row raises a contradiction finding. A row with a correct direction whose error clears a floor (ten in absolute terms or five percent) is surfaced as PREDICTION_ORACLE_NUMERIC_LARGE_MISS_SURFACED. A row with no realized price is not fabricated into a graded row; it raises PREDICTION_ORACLE_MISSING_REALIZED_TRUTH_NOT_GRADED. A row marked degraded is gated out of grading rather than scored. STOCK and ETF rows are counted separately, and a run that collapses the split raises PREDICTION_ORACLE_ASSET_CLASS_SPLIT_PRESERVED.
\n
Dossier edits held to bounded deltas
\n
_validate_mutations allows an edit only to add a contradiction, revise confidence, or retire a claim against a known prediction, with high-severity mutations needing two refs and an allowlist._validate_mutations reads dossier_mutations. An edit to the running record may only add a contradiction, revise a confidence band, or retire a claim, and it must target a known prediction. A high-severity mutation needs two evidence refs and an explicit public-delta allowlist, or it raises PREDICTION_DOSSIER_MUTATION_UNSAFE.
\n
The count floor for a pass
\n
A pass requires no positive finding and a minimum census of CP1 branches, CP2 predictions, graded rows, STOCK and ETF numeric rows, and one bounded mutation. The packet passes only when no positive finding fires and the counts hold: at least two CP1 branches, two CP2 predictions, two graded oracle-diff rows, two numeric rows spanning both STOCK and ETF, and one bounded mutation.
\n
The bundle path verifies copied source bodies
\n
The bundle path adds validate_source_module_imports to confirm each copied source body is metadata-only and digest-matched, then writes the result record files and summary card. The bundle path adds validate_source_module_imports, which reads source_module_manifest.json, confirms each copied source body declares copied_non_secret_macro_body, keeps its body text out of result records, states an exact-copy or source-faithful relation, and matches the recorded sha256 through _sha256. _source_open_body_import_summary reduces that to a metadata-only summary of digests and ids. write_receipts then writes the result record, board, validation, and sign-off files, and result_card builds the command-line summary card.
\n
The two constant tables that fix the boundary
\n
EXPECTED_NEGATIVE_CASES maps each negative input to its error code so a silenced gate is caught, and AUTHORITY_CEILING holds every authority flag false in every record. The two constant tables fix the boundary. EXPECTED_NEGATIVE_CASES maps each named negative input to the error code it must produce, so a gate that stops firing is caught as a missing case. AUTHORITY_CEILING holds every authority flag at false and is copied into every result record.
\nDiagram of the mechanism (6 steps).Diagram source & refs
flowchart TD Packet["Synthetic prediction packet targets, CP1 branches, CP2 predictions, oracle diff, numeric rows, mutations"] CP1["validate_cp1 chosen side + why the opposite side lost; equity lane needs confirmation"] CP2["validate_cp2 target inside universe; evidence must be pre-target (T-)"] Numeric["numeric grading recompute abs error, percent error, direction; claimed values must match"] Mutation["validate_mutations bounded dossier deltas only"] Result["build_result counts, verdicts, authority booleans; packet body stays out"] Packet --> CP1 Packet --> CP2 Packet --> Numeric Packet --> Mutation CP1 --> Result CP2 --> Result Numeric --> Result Mutation --> Result
\n
Negative cases
\n
Ten inputs the component must catch
\n
EXPECTED_NEGATIVE_CASES names ten inputs to reject or surface, each mapped to an error code, and _build_result fails if any expected case does not appear.EXPECTED_NEGATIVE_CASES names ten inputs the component must reject or surface: a CP2 target outside the universe, an unresolved CP1 bifurcation, post-target evidence, an unconfirmed equity-lane claim, an unsafe high-severity dossier mutation, a trading or advice overclaim, a direction hit that is still a large numeric miss, a row missing realized truth, a degraded feed row, and a run that drops the STOCK and ETF split. Each input maps to a specific error code, and _build_result fails if any expected case does not appear.
\n
Prior Art Grounding
\n
Forecast scoring and market-scoring lineage
\n
The component draws on probabilistic forecast evaluation and prediction-market scoring, borrowing the reconciliation pattern of declaring the universe up front, separating pre- and post-target evidence, and bounding dossier edits. This component draws on probabilistic forecast evaluation and prediction-market infrastructure. The Brier score is an early probability-forecast verification anchor, proper-scoring-rule work such as Gneiting and Raftery motivates incentive-compatible scoring, and Hanson's logarithmic market scoring rule grounds the idea that forecasts are updated and evaluated through explicit scoring. Forecasting-tournament work on calibration also motivates keeping prediction evidence separate from post-outcome explanation. What this component borrows is the reconciliation pattern: declare the target universe before the outcome, keep pre-target evidence apart from post-target evidence, grade against a synthetic oracle, and hold dossier edits to declared deltas.
A passing run means the synthetic packet cleared every gate, all ten negative cases fired, the scan was clean, and the digests matched, proving nothing about a real forecast. A passing run means the synthetic packet cleared every gate, all ten negative cases fired, the secret scan was clean, and the bundle digests matched. It proves nothing about a real forecast.
\n
Scope boundary
\n
Scope limit
\n
The strongest claim the evidence supports
\n
On a synthetic fixture, the component recomputes grading, refuses post-target evidence, requires resolved CP1 branches inside the universe, bounds dossier edits, and rejects every forbidden authority flag. The strongest claim the evidence supports is narrow. On a synthetic fixture, the component recomputes numeric grading rather than trusting the packet, refuses post-target evidence, requires resolved CP1 branches inside a declared target universe, holds dossier edits to bounded deltas, and rejects any attempt to raise a trading, advice, provider, live-market, public sharing, or launch flag. That is the proof boundary: the fixture, its gates, and its result records.
\n
What a green result is not
\n
A green result makes no forecast, gives no advice, calls no provider, imports no non-public data, and claims no track record, with AUTHORITY_CEILING holding every flag false. It refuses everything past that. It makes no forecast, gives no financial or investment-related actions, calls no live provider, imports no non-public data, and claims no track record. A green result is not launch-scope decision and not whole-system correctness. The scope limit in AUTHORITY_CEILING keeps every such flag false, and the validation result records under Validation Result record Path are the only evidence this module offers.
\n",
"components": [
{
"anchor": "component-prediction_oracle_reconciliation",
"component_id": "prediction_oracle_reconciliation",
"public_label": "Prediction Oracle Reconciliation"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 14,
"links": [
{
"id": "prediction_oracle_reconciliation",
"kind": "organ",
"kind_label": "component",
"label": "Prediction Oracle Reconciliation",
"relation": "explains",
"url": "component-prediction-oracle-reconciliation.html"
},
{
"id": "mechanism.prediction_oracle_reconciliation.validates_public_prediction_oracle_reconciliation",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public prediction oracle reconciliation",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-prediction-oracle-reconciliation-validates-public-prediction-oracle-reconciliation"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
},
{
"id": "paper_module.finance_forecast_evaluation_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Finance Forecast Evaluation Spine",
"relation": "depends on",
"url": "paper-module-finance-forecast-evaluation-spine.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
},
{
"id": "paper_module.research_replication_rubric_artifact_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Research Replication Rubric Artifact Replay",
"relation": "depends on",
"url": "paper-module-research-replication-rubric-artifact-replay.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch12_market_dashboard_read_model_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 12 Market Dashboard Read-Model Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[64:paper_module.batch12_market_dashboard_read_model_capsule].depends_on",
"url": "paper-module-batch12-market-dashboard-read-model-capsule.html"
},
{
"id": "paper_module.batch12_prediction_market_board_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 12 Prediction Market Board Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[65:paper_module.batch12_prediction_market_board_capsule].depends_on",
"url": "paper-module-batch12-prediction-market-board-capsule.html"
},
{
"id": "paper_module.batch8_structural_theses_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Structural Theses Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[61:paper_module.batch8_structural_theses_capsule].depends_on",
"url": "paper-module-batch8-structural-theses-capsule.html"
}
],
"paper_module_id": "paper_module.prediction_oracle_reconciliation",
"reader_packet": {
"abstract": "Prediction Oracle Reconciliation exercises synthetic forecast reconciliation gates without forecasting, trading, provider, or live-market authority.",
"authority_ceiling": "Synthetic invented prediction packet and source-module import evidence only; no forecasting correctness or accuracy, no trading, financial, or investment-related actions, no live market data, no external model access, no prediction public sharing, no performance track record, no non-public data import, no launch-scope decision, no publishing-scope decision, and no whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.prediction_oracle_reconciliation.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-prediction-oracle-reconciliation-diagram-1",
"html_url": "docs/paper-module-prediction-oracle-reconciliation.html#paper-module-prediction-oracle-reconciliation-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Packet[\"Synthetic prediction packet targets, CP1 branches, CP2 predictions, oracle diff, numeric rows, mutations\"] CP1[\"validate_cp1 chosen side + why the opposite side lost; equity lane needs confirmation\"] CP2[\"validate_cp2 target inside universe; evidence must be pre-target (T-)\"] Numeric[\"numeric grading recompute abs error, percent error, direction; claimed values must match\"] Mutation[\"validate_mutations bounded dossier deltas only\"] Result[\"build_result counts, verdicts, authority booleans; packet body stays out\"] Packet --> CP1 Packet --> CP2 Packet --> Numeric Packet --> Mutation CP1 --> Result CP2 --> Result Numeric --> Result Mutation --> Result",
"source_format": "mermaid",
"source_hash": "sha256:51b0e142997564a8da37ef010b7f7ca38bdafa289d04575f083a7908e42e110e",
"source_line_end": 79,
"source_line_start": 63
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.finance_forecast_evaluation_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Finance Forecast Evaluation Spine",
"relation": "depends on",
"url": "paper-module-finance-forecast-evaluation-spine.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
},
{
"id": "paper_module.research_replication_rubric_artifact_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Research Replication Rubric Artifact Replay",
"relation": "depends on",
"url": "paper-module-research-replication-rubric-artifact-replay.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "prediction_oracle_reconciliation",
"kind": "organ",
"kind_label": "component",
"label": "Prediction Oracle Reconciliation",
"relation": "explains",
"url": "component-prediction-oracle-reconciliation.html"
},
{
"id": "mechanism.prediction_oracle_reconciliation.validates_public_prediction_oracle_reconciliation",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public prediction oracle reconciliation",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-prediction-oracle-reconciliation-validates-public-prediction-oracle-reconciliation"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.prediction_oracle_reconciliation",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 124,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/prediction_oracle_reconciliation.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Brier score",
"section_id": "forecast-scoring-and-market-scoring-lineage",
"url": "https://journals.ametsoc.org/doi/10.1175/1520-0493%281950%29078%3C0001%3AVOFEIT%3E2.0.CO%3B2"
},
{
"kind": "external_reference",
"label": "Gneiting and Raftery",
"section_id": "forecast-scoring-and-market-scoring-lineage",
"url": "https://sites.stat.washington.edu/people/raftery/Research/PDF/Gneiting2007jasa.pdf"
},
{
"kind": "external_reference",
"label": "logarithmic market scoring rule",
"section_id": "forecast-scoring-and-market-scoring-lineage",
"url": "https://hanson.gmu.edu/mktscore.pdf"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "predictionoraclereconciliation audits one synthetic forecast packet on invented fixture data, with no live feed, external model access, trading, or performance claim. predictionoraclereconciliation takes one synthetic forecast packet and checks whether its reasoning would survive an honest audit. It runs entirely on invented fixture data. There is no live market feed, no external model access, no trading, and no performance claim.",
"level": 4,
"order": 1,
"section_id": "prediction-oracle-reconciliation-audits-one-synthetic-forecast-packet-on-invented-fixture-data",
"source_line_end": 6,
"source_line_start": 3,
"title": "`prediction_oracle_reconciliation` audits one synthetic forecast packet on invented fixture data"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It answers one narrow question about whether the packet keeps its evidence honest and grading recomputable, emitting metadata-only result records of counts, verdicts, and rejected cases. The component answers one narrow question: does this prediction packet keep its evidence honest and its grading recomputable, or does it cut a corner? It emits a set of result records that name the counts, verdicts, and rejected cases. The packet body, the claimed numbers, and any copied source bodies stay out of those records.",
"level": 4,
"order": 2,
"section_id": "it-answers-one-narrow-question-about-whether-the-packet-keeps-its-evidence-honest",
"source_line_end": 10,
"source_line_start": 7,
"title": "It answers one narrow question about whether the packet keeps its evidence honest"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component distrusts the packet's own numbers, recomputing error and direction from raw prices and rejecting any claimed value that disagrees. The unusual choice is that the component does not trust the numbers the packet reports. For every graded row it recomputes the error and the direction from the raw prices, then rejects any claimed value that disagrees with the recompute.",
"level": 4,
"order": 3,
"section_id": "the-component-distrusts-the-packet-s-own-numbers",
"source_line_end": 14,
"source_line_start": 11,
"title": "The component distrusts the packet's own numbers"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A forecast can hide a large numeric miss behind a correct direction or lean on evidence that arrived after the outcome, and this component surfaces both on a synthetic packet. A forecast can be wrong in two quiet ways. It can call the direction right and still be far off on the number, so a correct arrow hides a large miss. And it can look accurate only because it used evidence that arrived after the outcome it was meant to predict. A loose checker that reads the packet's own scores back to it catches neither. This component exists to make both failures visible on a synthetic packet before any reasoning is dressed up as a track record.",
"level": 4,
"order": 5,
"section_id": "two-quiet-ways-a-forecast-is-wrong",
"source_line_end": 20,
"source_line_start": 17,
"title": "Two quiet ways a forecast is wrong"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Prices, targets, branches, and realized values are all fixtures, so any direction hit or numeric miss speaks only about the fixture and the grading mechanics. The prices, targets, branches, and realized values are all fixtures. A direction hit or a numeric miss inside a result record is a statement about the fixture and the grading mechanics, nothing more.",
"level": 4,
"order": 6,
"section_id": "every-value-is-a-fixture",
"source_line_end": 24,
"source_line_start": 21,
"title": "Every value is a fixture"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run and runpredictionbundle both call buildresult, which loads the packet, secret-scans the inputs, validates the main packet and each negative case, and folds the parts into one status. The entry points are run for the first-wave fixture and runpredictionbundle for the exported bundle. Both call buildresult, which loads the packet, runs a secret scan over the inputs through scan_paths, validates the main packet, validates each negative-case input, and folds the parts into one status.",
"level": 4,
"order": 8,
"section_id": "entry-points-and-build-result",
"source_line_end": 30,
"source_line_start": 27,
"title": "Entry points and `_build_result`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatereconciliationpacket checks packet density, rejects any forbidden authority flag through authorityoverclaim, then runs and folds five gates. validatereconciliationpacket is the core. It first checks packet density: at least five pattern ids, four source refs, one projection result record, and three public runtime refs. It calls authorityoverclaim, which raises PREDICTIONTRADINGORADVICEOVERCLAIM if any flag in FORBIDDENAUTHORITYFLAGS (trading, financial or investment-related actions, live market data, external model access, public sharing, launch, non-public data equivalence) is set true. Then it runs five gates and folds their findings.",
"level": 4,
"order": 9,
"section_id": "density-and-authority-checks-in-validate-reconciliation-packet",
"source_line_end": 34,
"source_line_start": 31,
"title": "Density and authority checks in `validate_reconciliation_packet`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatecp1 requires every branch to name its chosen side, carry rationale, and keep the record of why the opposite side lost, with equity-lane branches also needing a confirmation bit. validatecp1 reads cp1branches. Every branch must name the side it chose, carry rationale refs, and keep an oppositesideinvalidationref, the record of why the losing side lost. A branch that asserts a winner without retaining the discarded alternative raises PREDICTIONCP1BIFURCATIONUNRESOLVED. An equity, market, or finance lane branch also needs an explicit confirmation bit, or it raises PREDICTIONEQUITYCONFIRMATION_REQUIRED.",
"level": 4,
"order": 10,
"section_id": "cp1-branches-must-retain-the-losing-side",
"source_line_end": 38,
"source_line_start": 35,
"title": "CP1 branches must retain the losing side"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatecp2 requires each prediction's target inside the declared universe and every evidence ref to be pre-target, the gate that stops a packet from grading itself with hindsight. validatecp2 reads cp2predictions. Each prediction's targetid must sit inside the declared target universe, so the set of things being predicted is fixed before the outcome; a target outside it raises PREDICTIONCP2TARGETOUTOFUNIVERSE. Every evidence ref must be pre-target: evidenceispret accepts a ref only if it carries a T- time prefix, and a ref that points past the target window raises PREDICTIONORACLEPOSTTEVIDENCE_FORBIDDEN. This is the gate that stops a packet from grading itself with hindsight.",
"level": 4,
"order": 11,
"section_id": "cp2-fixes-the-universe-and-forbids-post-target-evidence",
"source_line_end": 42,
"source_line_start": 39,
"title": "CP2 fixes the universe and forbids post-target evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The numeric functions recompute error and direction from raw prices, contradict any mismatched claim, surface large misses, refuse to grade missing or degraded rows, and preserve the STOCK/ETF split. numericgradinginputs and numericreconciliationrows do the arithmetic. For each graded oracle-diff row the component recomputes the absolute delta as the distance between predicted and realized price, the percent delta against the snapshot price, and the direction hit as predicted direction against realized direction. If the row also reports its own abserror, prederrorpct, or directionhit, the claimed value must match the recompute, or the row raises a contradiction finding. A row with a correct\u2026",
"level": 4,
"order": 12,
"section_id": "recomputing-every-graded-row-from-raw-prices",
"source_line_end": 46,
"source_line_start": 43,
"title": "Recomputing every graded row from raw prices"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatemutations allows an edit only to add a contradiction, revise confidence, or retire a claim against a known prediction, with high-severity mutations needing two refs and an allowlist. validatemutations reads dossiermutations. An edit to the running record may only add a contradiction, revise a confidence band, or retire a claim, and it must target a known prediction. A high-severity mutation needs two evidence refs and an explicit public-delta allowlist, or it raises PREDICTIONDOSSIERMUTATION_UNSAFE.",
"level": 4,
"order": 13,
"section_id": "dossier-edits-held-to-bounded-deltas",
"source_line_end": 50,
"source_line_start": 47,
"title": "Dossier edits held to bounded deltas"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass requires no positive finding and a minimum census of CP1 branches, CP2 predictions, graded rows, STOCK and ETF numeric rows, and one bounded mutation. The packet passes only when no positive finding fires and the counts hold: at least two CP1 branches, two CP2 predictions, two graded oracle-diff rows, two numeric rows spanning both STOCK and ETF, and one bounded mutation.",
"level": 4,
"order": 14,
"section_id": "the-count-floor-for-a-pass",
"source_line_end": 54,
"source_line_start": 51,
"title": "The count floor for a pass"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The bundle path adds validatesourcemoduleimports to confirm each copied source body is metadata-only and digest-matched, then writes the result record files and summary card. The bundle path adds validatesourcemoduleimports, which reads sourcemodulemanifest.json, confirms each copied source body declares copiednonsecretmacrobody, keeps its body text out of result records, states an exact-copy or source-faithful relation, and matches the recorded sha256 through sha256. sourceopenbodyimportsummary reduces that to a metadata-only summary of digests and ids. writereceipts then writes the result record, board, validation, and sign-off files, and resultcard builds the command-line summary card.",
"level": 4,
"order": 15,
"section_id": "the-bundle-path-verifies-copied-source-bodies",
"source_line_end": 58,
"source_line_start": 55,
"title": "The bundle path verifies copied source bodies"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVECASES maps each negative input to its error code so a silenced gate is caught, and AUTHORITYCEILING holds every authority flag false in every record. The two constant tables fix the boundary. EXPECTEDNEGATIVECASES maps each named negative input to the error code it must produce, so a gate that stops firing is caught as a missing case. AUTHORITYCEILING holds every authority flag at false and is copied into every result record.",
"level": 4,
"order": 16,
"section_id": "the-two-constant-tables-that-fix-the-boundary",
"source_line_end": 80,
"source_line_start": 59,
"title": "The two constant tables that fix the boundary"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "negative-cases",
"source_line_end": 82,
"source_line_start": 81,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVECASES names ten inputs to reject or surface, each mapped to an error code, and buildresult fails if any expected case does not appear. EXPECTEDNEGATIVECASES names ten inputs the component must reject or surface: a CP2 target outside the universe, an unresolved CP1 bifurcation, post-target evidence, an unconfirmed equity-lane claim, an unsafe high-severity dossier mutation, a trading or advice overclaim, a direction hit that is still a large numeric miss, a row missing realized truth, a degraded feed row, and a run that drops the STOCK and ETF split. Each input maps to a specific error code, and buildresult fails if any expected case does not appear.",
"level": 4,
"order": 18,
"section_id": "ten-inputs-the-component-must-catch",
"source_line_end": 86,
"source_line_start": 83,
"title": "Ten inputs the component must catch"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "prior-art-grounding",
"source_line_end": 88,
"source_line_start": 87,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component draws on probabilistic forecast evaluation and prediction-market scoring, borrowing the reconciliation pattern of declaring the universe up front, separating pre- and post-target evidence, and bounding dossier edits. This component draws on probabilistic forecast evaluation and prediction-market infrastructure. The Brier score is an early probability-forecast verification anchor, proper-scoring-rule work such as Gneiting and Raftery motivates incentive-compatible scoring, and Hanson's logarithmic market scoring rule grounds the idea that forecasts are updated and evaluated through explicit scoring. Forecasting-tournament work on calibration also motivates keeping prediction ev\u2026",
"level": 4,
"order": 20,
"section_id": "forecast-scoring-and-market-scoring-lineage",
"source_line_end": 92,
"source_line_start": 89,
"title": "Forecast scoring and market-scoring lineage"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from microcosm-substrate:",
"level": 2,
"order": 21,
"section_id": "validation-receipt-path",
"source_line_end": 109,
"source_line_start": 93,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A passing run means the synthetic packet cleared every gate, all ten negative cases fired, the scan was clean, and the digests matched, proving nothing about a real forecast. A passing run means the synthetic packet cleared every gate, all ten negative cases fired, the secret scan was clean, and the bundle digests matched. It proves nothing about a real forecast.",
"level": 4,
"order": 22,
"section_id": "what-a-passing-run-means",
"source_line_end": 113,
"source_line_start": 110,
"title": "What a passing run means"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 23,
"section_id": "claim-ceiling",
"source_line_end": 115,
"source_line_start": 114,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "On a synthetic fixture, the component recomputes grading, refuses post-target evidence, requires resolved CP1 branches inside the universe, bounds dossier edits, and rejects every forbidden authority flag. The strongest claim the evidence supports is narrow. On a synthetic fixture, the component recomputes numeric grading rather than trusting the packet, refuses post-target evidence, requires resolved CP1 branches inside a declared target universe, holds dossier edits to bounded deltas, and rejects any attempt to raise a trading, advice, provider, live-market, public sharing, or launch flag. That is the proof boundary: the fixture, its gates, and its result records.",
"level": 4,
"order": 24,
"section_id": "the-strongest-claim-the-evidence-supports",
"source_line_end": 119,
"source_line_start": 116,
"title": "The strongest claim the evidence supports"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A green result makes no forecast, gives no advice, calls no provider, imports no non-public data, and claims no track record, with AUTHORITYCEILING holding every flag false. It refuses everything past that. It makes no forecast, gives no financial or investment-related actions, calls no live provider, imports no non-public data, and claims no track record. A green result is not launch-scope decision and not whole-system correctness. The scope limit in AUTHORITYCEILING keeps every such flag false, and the validation result records under Validation Result record Path are the only evidence this module offers.",
"level": 4,
"order": 25,
"section_id": "what-a-green-result-is-not",
"source_line_end": 123,
"source_line_start": 120,
"title": "What a green result is not"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/prediction_oracle_reconciliation.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Prediction Oracle Reconciliation exercises synthetic forecast reconciliation gates without forecasting, trading, provider, or live-market authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "prediction_oracle_reconciliation.py",
"path": "src/microcosm_core/organs/prediction_oracle_reconciliation.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for synthetic prediction reconciliation validation, source-module digest checks, oracle-diff and numeric-row gates, negative cases, and scope limits.",
"symbols": [
"run",
"run_prediction_bundle",
"validate_source_module_imports",
"validate_reconciliation_packet",
"_source_open_body_import_summary",
"_build_result",
"write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/prediction_oracle_reconciliation.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/prediction_oracle_reconciliation.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/prediction_oracle_reconciliation.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Prediction Oracle Reconciliation validates synthetic prediction packets through CP1 fork preservation, CP2 target-universe checks, pre-target evidence limits, oracle-diff grading, bounded dossier edits, numeric reconciliation rows, source-module imports, negative cases, and scope limits. It is a projection-mechanics replay, not a forecasting correctness claim, investment or trading decisions, live market data call, external model access, non-public data equivalence, public sharing, launch, or whole-system correctness.",
"summary_line": "Prediction Oracle Reconciliation exercises synthetic forecast reconciliation gates without forecasting, trading, provider, or live-market authority.",
"title": "Prediction Oracle Reconciliation"
},
{
"anchor": "paper-module-provider-context-recipe-budget",
"authority_ceiling": "Context-budget projection fixture and source-body import evidence only; no provider or API call authorization, no Lean or Lake execution, no proof or oracle truth-side material, no theorem or domain-level conclusions, no launch-scope decision, no publishing-scope decision, and no whole-system correctness.",
"body_html": "
Six fixed context-budget recipes
\n
provider_context_recipe_budget_policy measures the byte cost of an agent's assembled context across six fixed recipes, filling each in order until the budget runs out and naming what did not fit, without calling any model.provider_context_recipe_budget_policy checks the size accounting of the context an agent would assemble before it calls an outside model. It does not call one. It takes six fixed recipe shapes, measures how many bytes each section costs, fills each recipe in declared order until the budget runs out, and names any section that did not fit.
\n
Context windows as allocated budgets
\n
The defended idea is that a context window is a budget to spend rather than a dump, so an over-budget recipe must name the sections it dropped instead of silently truncating. The idea it defends is short: a context window is a budget to spend, not a place to dump everything retrieved. The board records this as context_is_budget_not_dump. A recipe that goes over budget must say which sections it dropped, not silently truncate them.
\n
The output is context-shape metadata per recipe: the byte ceiling, admitted and omitted section ids, the allowed deliverable route, and authority claims that stay false, with no model call and no answer. The output is metadata about context shape. For each recipe it reports the byte ceiling, which section ids were admitted, which were omitted, the deliverable route the recipe is allowed to produce, and a set of authority claims that stay false. No model is called and no answer is produced.
\n
Purpose
\n
Two failure modes of a loose context builder
\n
A loose context builder can overflow a byte limit into arbitrary truncation or quietly carry answer-side material into a prompt, and this component makes both visible as plain accounting. A loose context builder has two failure modes. It can overflow a real byte limit and get truncated at an arbitrary point, and it can quietly carry answer-side material into a prompt: a proof body, an oracle-only premise list, an ideal answer, a provider output body. This component makes both failures visible as plain accounting before any model call is in scope.
\n
Six recipes pinned to fixed budgets and routes
\n
The six recipes are a closed set in EXPECTED_RECIPE_BUDGETS, each pinned to one deliverable route in EXPECTED_DELIVERABLES so it cannot change the kind of output it claims. The six recipes are a closed set, listed in EXPECTED_RECIPE_BUDGETS: minimal_4kb at 4096 bytes, premise_16kb at 16384, skill_32kb and repair_32kb at 32768, fewshot_64kb at 65536, and strategy_classification_4kb at 4096. Each recipe is pinned to one deliverable route in EXPECTED_DELIVERABLES, so a recipe cannot change what kind of output it claims to produce.
\n
Byte sizes measured, not asserted
\n
Byte sizes are measured by running the copied benchmark harness's real _provider_context_pack when present, falling back to declared sizes and recording that in real_section_body_status rather than guessing. The byte sizes are measured, not asserted. When the copied benchmark harness is present the validator imports it and runs its real _provider_context_pack over each recipe, then measures the packed section bodies. If the harness cannot load, the component falls back to the declared byte sizes and says so in real_section_body_status rather than guessing.
\n
How it works
\n
Two entrypoints share _build_result
\n
run loads the fixture with seven negative cases and run_budget_bundle the exported shape without them, both calling _build_result to load recipes, sections, copied bodies, and the scan first.run and run_budget_bundle are the two entrypoints. run loads the fixture inputs with the seven negative cases enabled. run_budget_bundle loads the exported bundle shape without them. Both call _build_result, which loads the recipe rows, section rows, copied source-module bodies, and the non-public-state scan before it builds anything.
\n
Deterministic ordered fill in _recipe_projection
\n
_recipe_projection walks a recipe's ordered sections, admitting each only while the running total stays within byte_budget and omitting the first over-budget section and every one after it. Recipe projection is deterministic. _recipe_projection walks a recipe's ordered section ids, measures each section with _byte_size and _section_accounting, and admits a section only while the running total stays at or under the recipe's byte_budget. The first section that would push the total over budget goes to the omitted list instead, and so does every section after it. The projection returns the included and omitted section ids, the included byte count, an approximate token count of ceil(bytes / 4), the graph role, and the deliverable type.
\n
The validator then runs three independent checks and turns each failure into a structured finding rather than a silent pass. The validator then runs three independent checks and turns each failure into a structured finding rather than a silent pass.
\nDiagram of the mechanism (7 steps).Diagram source & refs
flowchart TD Inputs["recipe rows + section rows + copied source bodies"] --> Project["_recipe_projection ordered byte-budget fill"] Project --> RecipeCheck["_recipe_findings budget, route, manifest"] Inputs --> SectionCheck["_section_findings source ref + anchor backing"] Inputs --> SourceCheck["_source_module_findings digest + anchor of copied bodies"] RecipeCheck --> Result["_build_result status + context packets"] SectionCheck --> Result SourceCheck --> Result Result --> Result records["_write_receipts metadata-only result records"]
\n
What _recipe_findings rejects
\n
_recipe_findings rejects a mismatched or oversized budget, a forbidden truth-side section, a proof-body or provider-call authorization, a drifted deliverable route, and an over-budget projection with no omitted-sections manifest._recipe_findings rejects a recipe whose budget does not match the pinned value or exceeds 65536, a recipe that lists a forbidden truth-side section id, a recipe that sets proof_bodies_allowed or carries a forbidden body field, a recipe that sets provider_calls_authorized, a recipe whose deliverable_type drifts from its route, and an over-budget projection with no omitted-sections manifest.
\n
Section refs and anchors in _section_findings
\n
_section_findings requires each public section to cite an allowed source ref and anchor that truly appears in a copied body, rejecting forbidden ids, forbidden body keys, outside refs, and synthetic material._section_findings requires each public section to cite a source ref from the allowed set and at least one source anchor, then confirms the anchor text actually appears in one of the cited copied bodies. It rejects a section id in the forbidden set, a section carrying a forbidden body key, a section citing a ref outside the imported bodies, and a section marked synthetic, placeholder, fabricated, or fixture-only by _synthetic_section_material_marker.
\n
Digest and anchor checks over eight copied bodies
\n
_source_module_findings covers eight copied bodies, checking for each that the file exists, the manifest and source digests match the target, and every required anchor string is present._source_module_findings covers the eight copied bodies: the graph benchmark harness, the formal-ladder eval, the result record reducer, the set calibration report, and the compute-provider, provider-adapter, transform-job, and navigation-transform-result record standards. For each it checks that the row declares a copied body kept out of result records, that the target file exists, that the manifest digest matches the target and the source digest matches the target, and that every required anchor string is present.
\n
metadata-only result records and the compact card
\n
_write_receipts writes metadata-only records in both modes and result_card emits a compact card, keeping counts, ids, digests, routes, and verdicts but never proof bodies or model answers._write_receipts keeps the output metadata-only. Fixture mode writes a result, a board, a validation record, and an sign-off record. Bundle mode writes one exported-bundle validation record. result_card emits a compact card and drops the context packets, source refs, result record paths, private-scan hits, and the scope boundary payload. The records keep counts, ids, digests, routes, and verdicts. They never keep proof bodies or model answers.
\n
The recipe and its route:
\n
Recipe
Budget
Deliverable route
minimal_4kb
4 KiB
environment_metadata
premise_16kb
16 KiB
ranked_premise_ids
skill_32kb
32 KiB
strategy_metadata
repair_32kb
32 KiB
failure_classification
fewshot_64kb
64 KiB
redacted_synthesis_advisory
strategy_classification_4kb
4 KiB
strategy_id_classification
\n
Negative cases
\n
Seven cases the validator must reject
\n
Seven fixture cases must each be rejected, every one keyed to an expected error code in EXPECTED_NEGATIVE_CASES. The fixture carries seven cases the validator must reject, each keyed to an expected error code in EXPECTED_NEGATIVE_CASES.
\n
budget_overflow_recipe rejects a recipe above its public byte ceiling.
truth_side_section rejects an oracle-only section id.
proof_body_leakage rejects proof or model body fields.
provider_call_authorized rejects a fixture that authorizes a model call.
deliverable_type_route_mismatch rejects a recipe whose route changed.
omitted_sections_suppressed rejects over-budget context with no omitted-sections manifest.
synthetic_section_materials rejects section material with no allowed source ref or anchor, or marked synthetic.
\n
A missing case blocks the result
\n
A pass requires all seven cases observed and no floor check fired, and a missing case blocks the result. A pass means all seven were observed and no floor check fired. A missing case blocks the result.
\n
Prior Art Grounding
\n
Retrieval-augmented generation and CLI restraint
\n
The recipe budget sits in the retrieval-augmented-generation lineage, narrowed to metadata that budgets, orders, and drops sections, and borrows the CLI principle of saying enough but not too much. The recipe budget sits in the lineage of retrieval-augmented generation and context packing. Lewis et al.'s Retrieval-Augmented Generation is the research anchor for conditioning generation on retrieved supporting material rather than model parameters alone. This component narrows that idea to metadata: retrieved sections are budgeted, ordered, and dropped explicitly before any model call is in scope. The command-facing style borrows the Command Line Interface Guidelines principle of saying enough but not too much, expressed here as fixed byte ceilings and explicit omitted-section manifests.
A green result proves only public context-recipe metadata, with fixture mode observing every negative case and writing four records and bundle mode validating the exported shape into one metadata-only record. A green result proves only public context-recipe metadata: byte ceilings, ordered admission, omitted sections, deliverable routes, the digest and anchor checks over the copied bodies, and the seven negative cases. Fixture mode must observe every negative case and write the four records. Bundle mode must validate the exported shape and write one metadata-only record.
\n
Scope boundary
\n
Scope limit
\n
The strongest claim supported
\n
The strongest claim is that a closed set of recipes stays inside its byte ceilings, fills in order, names what it drops, and carries no answer-side material, all as accounting before any model call. The strongest claim this component supports is that a small, closed set of context recipes stays inside its declared byte ceilings, fills sections in declared order, names what it drops, and carries no answer-side material, all visible as accounting before any model call. The proof boundary is exactly that metadata plus the copied bodies it checks by digest and anchor.
\n
What the boundary refuses
\n
The boundary calls no model, runs no Lean or Lake, proves no theorem, and exports no proof or answer body, excluding launch, public sharing, and any treatment of context metadata as proof authority. It refuses everything past that boundary. It does not call a model, run Lean or Lake, prove a theorem, export a proof body, or reveal oracle-only truth-side material. Its scope limit excludes model answer bodies, launch-scope decision, publishing-scope decision, and treating context metadata as proof authority. The strategy_classification_4kb route emits only strategy_id_classification and cannot carry a model answer body.
\n",
"components": [
{
"anchor": "component-provider_context_recipe_budget_policy",
"component_id": "provider_context_recipe_budget_policy",
"public_label": "Provider Context Recipe Budget Policy"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 6
},
"link_count": 18,
"links": [
{
"id": "provider_context_recipe_budget_policy",
"kind": "organ",
"kind_label": "component",
"label": "Provider Context Recipe Budget Policy",
"relation": "explains",
"url": "component-provider-context-recipe-budget-policy.html"
},
{
"id": "mechanism.provider_context_recipe_budget_policy.validates_public_context_budget_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public context budget boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-provider-context-recipe-budget-policy-validates-public-context-budget-boundary"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.formal_math_readiness_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Readiness Gate",
"relation": "depends on",
"url": "paper-module-formal-math-readiness-gate.html"
},
{
"id": "paper_module.undeclared_library_prior_classifier",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Undeclared Library Prior Classifier",
"relation": "depends on",
"url": "paper-module-undeclared-library-prior-classifier.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.bounded_autonomy_campaign_packet",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Bounded Autonomy Campaign Packet",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[45:paper_module.bounded_autonomy_campaign_packet].depends_on",
"url": "paper-module-bounded-autonomy-campaign-packet.html"
},
{
"id": "paper_module.tool_server_pressure_inventory",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tool Server Pressure Inventory",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[50:paper_module.tool_server_pressure_inventory].depends_on",
"url": "paper-module-tool-server-pressure-inventory.html"
}
],
"paper_module_id": "paper_module.provider_context_recipe_budget",
"reader_packet": {
"abstract": "Provider Context Recipe Budget validates context-budget projection mechanics without authorizing external model access or truth-side material.",
"authority_ceiling": "Context-budget projection fixture and source-body import evidence only; no provider or API call authorization, no Lean or Lake execution, no proof or oracle truth-side material, no theorem or domain-level conclusions, no launch-scope decision, no publishing-scope decision, and no whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.provider_context_recipe_budget.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-provider-context-recipe-budget-diagram-1",
"html_url": "docs/paper-module-provider-context-recipe-budget.html#paper-module-provider-context-recipe-budget-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Inputs[\"recipe rows + section rows + copied source bodies\"] --> Project[\"_recipe_projection ordered byte-budget fill\"] Project --> RecipeCheck[\"_recipe_findings budget, route, manifest\"] Inputs --> SectionCheck[\"_section_findings source ref + anchor backing\"] Inputs --> SourceCheck[\"_source_module_findings digest + anchor of copied bodies\"] RecipeCheck --> Result[\"_build_result status + context packets\"] SectionCheck --> Result SourceCheck --> Result Result --> Result records[\"_write_receipts metadata-only result records\"]",
"source_format": "mermaid",
"source_hash": "sha256:b4986e845d1359af06957ad1e9e0f25b414b00ca8fe8f7858a82bf1aed6747ff",
"source_line_end": 113,
"source_line_start": 103
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.formal_math_readiness_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Readiness Gate",
"relation": "depends on",
"url": "paper-module-formal-math-readiness-gate.html"
},
{
"id": "paper_module.undeclared_library_prior_classifier",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Undeclared Library Prior Classifier",
"relation": "depends on",
"url": "paper-module-undeclared-library-prior-classifier.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "provider_context_recipe_budget_policy",
"kind": "organ",
"kind_label": "component",
"label": "Provider Context Recipe Budget Policy",
"relation": "explains",
"url": "component-provider-context-recipe-budget-policy.html"
},
{
"id": "mechanism.provider_context_recipe_budget_policy.validates_public_context_budget_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public context budget boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-provider-context-recipe-budget-policy-validates-public-context-budget-boundary"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-16",
"kind": "principle",
"kind_label": "principle",
"label": "Bind authority to transaction scope",
"relation": "governed by",
"url": "doctrine.html#dcard-p-16"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.provider_context_recipe_budget",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 273,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/provider_context_recipe_budget.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Retrieval-Augmented Generation",
"section_id": "retrieval-augmented-generation-and-cli-restraint",
"url": "https://arxiv.org/abs/2005.11401"
},
{
"kind": "external_reference",
"label": "Command Line Interface Guidelines",
"section_id": "retrieval-augmented-generation-and-cli-restraint",
"url": "https://clig.dev/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "providercontextrecipebudget_policy measures the byte cost of an agent's assembled context across six fixed recipes, filling each in order until the budget runs out and naming what did not fit, without calling any model.** providercontextrecipebudgetpolicy checks the size accounting of the context an agent would assemble before it calls an outside model. It does not call one. It takes six fixed recipe shapes, measures how many bytes each section costs, fills each recipe in declared order until the budget runs out, and names any section that did not fit.",
"level": 4,
"order": 1,
"section_id": "six-fixed-context-budget-recipes",
"source_line_end": 13,
"source_line_start": 3,
"title": "Six fixed context-budget recipes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The defended idea is that a context window is a budget to spend rather than a dump, so an over-budget recipe must name the sections it dropped instead of silently truncating.** The idea it defends is short: a context window is a budget to spend, not a place to dump everything retrieved. The board records this as contextisbudgetnotdump. A recipe that goes over budget must say which sections it dropped, not silently truncate them.",
"level": 4,
"order": 2,
"section_id": "context-windows-as-allocated-budgets",
"source_line_end": 23,
"source_line_start": 14,
"title": "Context windows as allocated budgets"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The output is context-shape metadata per recipe: the byte ceiling, admitted and omitted section ids, the allowed deliverable route, and authority claims that stay false, with no model call and no answer.** The output is metadata about context shape. For each recipe it reports the byte ceiling, which section ids were admitted, which were omitted, the deliverable route the recipe is allowed to produce, and a set of authority claims that stay false. No model is called and no answer is produced.",
"level": 4,
"order": 3,
"section_id": "the-output-is-context-shape-metadata-per-recipe",
"source_line_end": 33,
"source_line_start": 24,
"title": "The output is context-shape metadata per recipe"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 35,
"source_line_start": 34,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A loose context builder can overflow a byte limit into arbitrary truncation or quietly carry answer-side material into a prompt, and this component makes both visible as plain accounting.** A loose context builder has two failure modes. It can overflow a real byte limit and get truncated at an arbitrary point, and it can quietly carry answer-side material into a prompt: a proof body, an oracle-only premise list, an ideal answer, a provider output body. This component makes both failures visible as plain accounting before any model call is in scope.",
"level": 4,
"order": 5,
"section_id": "two-failure-modes-of-a-loose-context-builder",
"source_line_end": 46,
"source_line_start": 36,
"title": "Two failure modes of a loose context builder"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The six recipes are a closed set in EXPECTEDRECIPE_BUDGETS, each pinned to one deliverable route in EXPECTED_DELIVERABLES so it cannot change the kind of output it claims.** The six recipes are a closed set, listed in EXPECTEDRECIPEBUDGETS: minimal4kb at 4096 bytes, premise16kb at 16384, skill_32kb and repair32kb at 32768, fewshot64kb at 65536, and strategyclassification4kb at 4096. Each recipe is pinned to one deliverable route in EXPECTED_DELIVERABLES, so a recipe cannot change what kind of output it claims to produce.",
"level": 4,
"order": 6,
"section_id": "six-recipes-pinned-to-fixed-budgets-and-routes",
"source_line_end": 58,
"source_line_start": 47,
"title": "Six recipes pinned to fixed budgets and routes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Byte sizes are measured by running the copied benchmark harness's real providercontext_pack when present, falling back to declared sizes and recording that in realsectionbodystatus rather than guessing. The byte sizes are measured, not asserted. When the copied benchmark harness is present the validator imports it and runs its real providercontext_pack over each recipe, then measures the packed section bodies. If the harness cannot load, the component falls back to the declared byte sizes and says so in realsectionbody_status rather than guessing.",
"level": 4,
"order": 7,
"section_id": "byte-sizes-measured-not-asserted",
"source_line_end": 69,
"source_line_start": 59,
"title": "Byte sizes measured, not asserted"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 71,
"source_line_start": 70,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run loads the fixture with seven negative cases and runbudget_bundle the exported shape without them, both calling buildresult to load recipes, sections, copied bodies, and the scan first.** run and runbudgetbundle are the two entrypoints. run loads the fixture inputs with the seven negative cases enabled. runbudgetbundle loads the exported bundle shape without them. Both call buildresult, which loads the recipe rows, section rows, copied source-module bodies, and the non-public-state scan before it builds anything.",
"level": 4,
"order": 9,
"section_id": "two-entrypoints-share-build-result",
"source_line_end": 82,
"source_line_start": 72,
"title": "Two entrypoints share `_build_result`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "recipeprojection walks a recipe's ordered sections, admitting each only while the running total stays within byte_budget and omitting the first over-budget section and every one after it.** Recipe projection is deterministic. recipeprojection walks a recipe's ordered section ids, measures each section with bytesize and sectionaccounting, and admits a section only while the running total stays at or under the recipe's byte_budget. The first section that would push the total over budget goes to the omitted list instead, and so does every section after it. The projection returns the included and omitted section ids, the included byte count, an approximate token count of ceil(bytes / 4), the gr\u2026",
"level": 4,
"order": 10,
"section_id": "deterministic-ordered-fill-in-recipe-projection",
"source_line_end": 96,
"source_line_start": 83,
"title": "Deterministic ordered fill in `_recipe_projection`"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "The validator then runs three independent checks and turns each failure into a structured finding rather than a silent pass. The validator then runs three independent checks and turns each failure into a structured finding rather than a silent pass.",
"level": 4,
"order": 11,
"section_id": "the-validator-then-runs-three-independent-checks",
"source_line_end": 114,
"source_line_start": 97,
"title": "The validator then runs three independent checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "recipefindings rejects a mismatched or oversized budget, a forbidden truth-side section, a proof-body or provider-call authorization, a drifted deliverable route, and an over-budget projection with no omitted-sections manifest.** recipefindings rejects a recipe whose budget does not match the pinned value or exceeds 65536, a recipe that lists a forbidden truth-side section id, a recipe that sets proofbodiesallowed or carries a forbidden body field, a recipe that sets providercallsauthorized, a recipe whose deliverable_type drifts from its route, and an over-budget projection with no omitted-sections manifest.",
"level": 4,
"order": 12,
"section_id": "what-recipe-findings-rejects",
"source_line_end": 127,
"source_line_start": 115,
"title": "What `_recipe_findings` rejects"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sectionfindings requires each public section to cite an allowed source ref and anchor that truly appears in a copied body, rejecting forbidden ids, forbidden body keys, outside refs, and synthetic material.** sectionfindings requires each public section to cite a source ref from the allowed set and at least one source anchor, then confirms the anchor text actually appears in one of the cited copied bodies. It rejects a section id in the forbidden set, a section carrying a forbidden body key, a section citing a ref outside the imported bodies, and a section marked synthetic, placeholder, fabricated, or fixture-only by syntheticsectionmaterialmarker.",
"level": 4,
"order": 13,
"section_id": "section-refs-and-anchors-in-section-findings",
"source_line_end": 139,
"source_line_start": 128,
"title": "Section refs and anchors in `_section_findings`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourcemodule_findings covers eight copied bodies, checking for each that the file exists, the manifest and source digests match the target, and every required anchor string is present.** sourcemodule_findings covers the eight copied bodies: the graph benchmark harness, the formal-ladder eval, the result record reducer, the set calibration report, and the compute-provider, provider-adapter, transform-job, and navigation-transform-result record standards. For each it checks that the row declares a copied body kept out of result records, that the target file exists, that the manifest digest matches the target and the source digest matches the target, and that every required anchor string is pr\u2026",
"level": 4,
"order": 14,
"section_id": "digest-and-anchor-checks-over-eight-copied-bodies",
"source_line_end": 152,
"source_line_start": 140,
"title": "Digest and anchor checks over eight copied bodies"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "writereceipts writes metadata-only records in both modes and result_card emits a compact card, keeping counts, ids, digests, routes, and verdicts but never proof bodies or model answers.** writereceipts keeps the output metadata-only. Fixture mode writes a result, a board, a validation record, and an sign-off record. Bundle mode writes one exported-bundle validation record. result_card emits a compact card and drops the context packets, source refs, result record paths, private-scan hits, and the scope boundary payload. The records keep counts, ids, digests, routes, and verdicts. They never keep proof bodies or model answers. The recipe and its route:",
"level": 4,
"order": 15,
"section_id": "body-free-receipts-and-the-compact-card",
"source_line_end": 175,
"source_line_start": 153,
"title": "metadata-only result records and the compact card"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "negative-cases",
"source_line_end": 177,
"source_line_start": 176,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**Seven fixture cases must each be rejected, every one keyed to an expected error code in EXPECTEDNEGATIVECASES.** The fixture carries seven cases the validator must reject, each keyed to an expected error code in EXPECTEDNEGATIVECASES. budgetoverflowrecipe rejects a recipe above its public byte ceiling. truthsidesection rejects an oracle-only section id. proofbodyleakage rejects proof or model body fields. providercallauthorized rejects a fixture that authorizes a model call. deliverabletyperoute_mismatch rejects a recipe whose route changed. omittedsectionssuppressed rejects over-budget context with no omitted-sections manifest. syntheticsectionmaterials rejects section material with no a\u2026",
"level": 4,
"order": 17,
"section_id": "seven-cases-the-validator-must-reject",
"source_line_end": 194,
"source_line_start": 178,
"title": "Seven cases the validator must reject"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass requires all seven cases observed and no floor check fired, and a missing case blocks the result.** A pass means all seven were observed and no floor check fired. A missing case blocks the result.",
"level": 4,
"order": 18,
"section_id": "a-missing-case-blocks-the-result",
"source_line_end": 201,
"source_line_start": 195,
"title": "A missing case blocks the result"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "prior-art-grounding",
"source_line_end": 203,
"source_line_start": 202,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The recipe budget sits in the retrieval-augmented-generation lineage, narrowed to metadata that budgets, orders, and drops sections, and borrows the CLI principle of saying enough but not too much.** The recipe budget sits in the lineage of retrieval-augmented generation and context packing. Lewis et al.'s Retrieval-Augmented Generation is the research anchor for conditioning generation on retrieved supporting material rather than model parameters alone. This component narrows that idea to metadata: retrieved sections are budgeted, ordered, and dropped explicitly before any model call is in scope. The command-facing style borrows the Command Line Interface Guidelines principle of saying e\u2026",
"level": 4,
"order": 20,
"section_id": "retrieval-augmented-generation-and-cli-restraint",
"source_line_end": 219,
"source_line_start": 204,
"title": "Retrieval-augmented generation and CLI restraint"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from microcosm-substrate:",
"level": 2,
"order": 21,
"section_id": "validation-receipt-path",
"source_line_end": 236,
"source_line_start": 220,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A green result proves only public context-recipe metadata, with fixture mode observing every negative case and writing four records and bundle mode validating the exported shape into one metadata-only record.** A green result proves only public context-recipe metadata: byte ceilings, ordered admission, omitted sections, deliverable routes, the digest and anchor checks over the copied bodies, and the seven negative cases. Fixture mode must observe every negative case and write the four records. Bundle mode must validate the exported shape and write one metadata-only record.",
"level": 4,
"order": 22,
"section_id": "what-a-green-result-proves",
"source_line_end": 247,
"source_line_start": 237,
"title": "What a green result proves"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 23,
"section_id": "claim-ceiling",
"source_line_end": 249,
"source_line_start": 248,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is that a closed set of recipes stays inside its byte ceilings, fills in order, names what it drops, and carries no answer-side material, all as accounting before any model call.** The strongest claim this component supports is that a small, closed set of context recipes stays inside its declared byte ceilings, fills sections in declared order, names what it drops, and carries no answer-side material, all visible as accounting before any model call. The proof boundary is exactly that metadata plus the copied bodies it checks by digest and anchor.",
"level": 4,
"order": 24,
"section_id": "the-strongest-claim-supported",
"source_line_end": 260,
"source_line_start": 250,
"title": "The strongest claim supported"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The boundary calls no model, runs no Lean or Lake, proves no theorem, and exports no proof or answer body, excluding launch, public sharing, and any treatment of context metadata as proof authority.** It refuses everything past that boundary. It does not call a model, run Lean or Lake, prove a theorem, export a proof body, or reveal oracle-only truth-side material. Its scope limit excludes model answer bodies, launch-scope decision, public sharing decision, and treating context metadata as proof authority. The strategyclassification4kb route emits only strategyidclassification and cannot carry a model answer body.",
"level": 4,
"order": 25,
"section_id": "what-the-boundary-refuses",
"source_line_end": 272,
"source_line_start": 261,
"title": "What the boundary refuses"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/provider_context_recipe_budget.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Provider Context Recipe Budget validates context-budget projection mechanics without authorizing external model access or truth-side material."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "provider_context_recipe_budget_policy.py",
"path": "src/microcosm_core/organs/provider_context_recipe_budget_policy.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for provider-context budget fixture validation, section fill checks, omitted-section manifests, source-body digest checks, negative cases, and scope limits.",
"symbols": [
"run",
"run_budget_bundle",
"_recipe_projection",
"_recipe_findings",
"_section_findings",
"_build_result",
"_write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/provider_context_recipe_budget_policy.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/provider_context_recipe_budget.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/provider_context_recipe_budget.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Provider Context Recipe Budget validates provider-context recipe mechanics: fixed byte ceilings, ordered section fill, omitted-section manifests, deliverable routing, digest-checked source-body imports, forbidden-body rejection, negative cases, and scope limits. It emits context metadata and verdicts only, without provider/API authorization, Lean/Lake execution, proof or oracle truth-side material, formal-result correctness, domain-level conclusions, launch, public sharing, or whole-system correctness.",
"summary_line": "Provider Context Recipe Budget validates context-budget projection mechanics without authorizing external model access or truth-side material.",
"title": "Provider Context Recipe Budget"
},
{
"anchor": "paper-module-undeclared-library-prior-classifier",
"authority_ceiling": "Symbol-boundary classification fixture over copied Lean/Std premise rows and pre-extracted symbol observations only; no proof-source reads, Lean or Lake execution, formal-result correctness, whole-library implicit allowlist, Mathlib availability, external model access, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
Declared versus undeclared library priors
\n
It answers one question about a Lean-accepted proof: did the candidate cite a known library symbol outside the premises its recipe allowed? A theorem prover can return a proof that Lean accepts and still break the rules of the evaluation it ran under. The usual way this happens is quiet: the proof reaches for a library lemma the recipe never put on the table. The symbol is real and the proof is sound, but the run used a fact it was not allowed to assume. This component answers one question about that gap. Given the premises a candidate was allowed to use and the qualified symbols it actually reached for, did it cite a known library symbol outside the allowed set?
\n
It works only from a copied premise index and pre-extracted observations, never running Lean or reading the proof, with the allowed set closed by construction from premise rows. It works only from a copied premise index and a list of symbol observations that were extracted beforehand. It does not run Lean, it does not read the proof body, and it does not treat the standard library as an implicit allowlist. The allowed set is closed by construction: a symbol is admissible because a premise row names it, never because it happens to exist in Lean's standard library.
\n
The public fixture's 11 premises and three observations classify to one undeclared-prior escalation, one budget-violation retry, and one advisory none, each named with its rule and class. The public fixture carries 11 premise rows across four namespaces (Bool, Iff, List, Nat) and three symbol observations. Running it classifies one observation as UNDECLARED_LIBRARY_PRIOR routed to bridge_escalate, one as PREMISE_BUDGET_VIOLATION routed to retry, and one as an advisory NONE. The written records name each symbol, the premise rule it came from, and the class that was computed.
\n
Purpose
\n
Compiling does not mean in-recipe
\n
A checked proof is not a proof that stayed inside the allowed building blocks, and a loose check cannot tell an in-recipe proof from one that borrowed an out-of-recipe lemma. The problem is that "the proof checked" does not mean "the proof stayed inside the allowed building blocks." A loose check that trusts the prover's sign-off, or that admits any symbol found in the standard library, cannot tell an in-recipe proof from one that quietly borrowed an out-of-recipe lemma. Both compile. Both look sound.
\n
Budget breach kept apart from residual breach
\n
An explicit budget breach from a disallowed premise id is kept distinct from a residual breach using an undeclared known symbol, with the budget case taking precedence. The component also keeps two failure modes apart. An explicit budget breach, where the candidate names a premise id the recipe did not allow, is not the same as a residual breach, where the candidate used a known symbol that turns out to be undeclared. The first is settled directly from the cited ids and takes precedence. The second is what the symbol comparison is for. Folding both into one class would either over-escalate honest retries or let genuine out-of-recipe library use pass as a budget note.
\n
How it works
\n
Two entry points into _build_result
\n
run builds the fixture result with negatives included, and run_symbol_bundle runs the same builder over an exported source-body bundle and can reuse a fresh prior record. The fixture entrypoint is run. It calls _build_result with the fixture input mode and negative cases included, then _write_receipts. The exported-bundle entrypoint is run_symbol_bundle, which runs the same builder over an exported source-body bundle and can reuse a fresh prior record. Inside _build_result the stages run in this order.
\n
Loading inputs and the secret scan
\n
_load_payloads reads the four inputs plus seven negative-case files, and scan_paths runs a secret-exclusion pass before the four validators run._load_payloads reads the four input files (projection_protocol.json, premise_index.json, symbol_observations.json, classifier_policy.json) plus the seven negative-case files. scan_paths runs a secret-exclusion pass over those inputs and records a blocking hit count. Then four validators run.
\n
Validating projection, premises, and policy
\n
Three validators check the projection's refs and digests, require every premise row to carry an id, qualified name, and source ref, and pin the symbol regex and review outcomes.validate_projection_protocol checks that the projection cites real source refs, source pattern ids, target refs, and source digests, and that any omitted proof or provider material carries an omission record. validate_premise_index reads the premise rows, requires each to have a premise id, a qualified theorem_or_def_name, and a public source ref, and reports the premise and namespace counts. validate_classifier_policy requires the declared qualified-symbol pattern to equal QUALIFIED_SYMBOL_RE (the regex over the Nat, List, Bool, Iff, and Eq namespaces), the undeclared review outcome to be bridge_escalate, the budget review outcome to be retry, and it rejects any policy that sets proof_bodies_allowed true.
\n
_classify_row computes the class with precedence
\n
_classify_row maps cited premises to symbols, finds the known ones outside the allowed and cited sets, and applies budget-over-residual precedence, flagging any drift from the asserted class.validate_symbol_observations is the classifier. For each observation it calls _classify_row, which is where the rule lives. _classify_row reads allowed_premise_ids and cited_unallowed_premise_ids, maps them through the premise index to their symbols, extracts the observed qualified refs, and keeps the known ones (those a premise row names). The undeclared set is the known symbols that are neither allowed nor already cited as unallowed. Precedence is explicit: if cited_unallowed_premise_ids is present the row is PREMISE_BUDGET_VIOLATION with retry; otherwise a non-empty undeclared set is UNDECLARED_LIBRARY_PRIOR with bridge_escalate; otherwise the row is NONE and advisory. _classify_row then compares the observation's own asserted class against the recomputed one and records a finding on any mismatch, so the public assertion cannot drift from the premise-index computation.
\n
The exported bundle's provenance check
\n
For the exported bundle validate_source_module_manifest rejects in-result record bodies and verifies each source body against its digests, counts, and anchors as provenance, not a second proof. For the exported bundle, validate_source_module_manifest checks source_module_manifest.json, rejects manifest-level or row-level body_in_receipt: true, and verifies each declared source-body import against its source and target digests, line counts, byte counts, and required anchors. This is a provenance check on copied bodies, not a second proof.
\n
Result records and the compact card
\n
_write_receipts writes the result, board, validation, and sign-off records, and result_card projects a pass or fail card that omits the heavier evidence._write_receipts writes the result, board, validation, and sign-off records. result_card projects a compact pass or fail card that omits the premise rows, classification rows, source digests, secret-scan detail, and scope limit body, so the card can be shown without exporting the heavier evidence.
\n
Function
Role
run
Fixture entrypoint; builds the result with negative cases and writes records
Runs the secret scan and four validators, merges findings, sets pass or blocked
validate_premise_index
Builds the closed allowlist from copied premise rows
validate_classifier_policy
Pins the qualified-symbol regex and the two review outcomes
validate_symbol_observations
Iterates observations and negative cases into classifications
_classify_row
Computes the class and applies the budget-over-residual precedence
validate_source_module_manifest
Verifies exported source bodies by digest, size, and anchor
_write_receipts
Emits result, board, validation, and sign-off records
result_card
Projects the compact pass or fail card with heavy keys omitted
\nDiagram of the mechanism (10 steps).Diagram source & refs
Source refs
Budget
cited_unallowed_premise_ids present
flowchart TD inputs["Copied premise index and symbol observations"] policy["Classifier policy regex and outcomes"] classify["_classify_row compute allowed vs undeclared"] budget["cited_unallowed_premise_ids present"] residual["Known symbol outside allowed set"] clean["Allowed or no known undeclared symbol"] retry["PREMISE_BUDGET_VIOLATION route retry"] escalate["UNDECLARED_LIBRARY_PRIOR route bridge_escalate"] advisory["NONE route accept_as_advisory"] records["Result, board, validation, sign-off records"] inputs --> classify policy --> classify classify --> budget classify --> residual classify --> clean budget --> retry residual --> escalate clean --> advisory retry --> records escalate --> records advisory --> records
\n
Negative cases
\n
Seven negative cases and their codes
\n
Seven negative cases each carry a code the run must observe, from leaked proof bodies and non-public refs to correctness overclaim, missing escalation, precedence violations, and false positives. The fixture ships seven negative cases in EXPECTED_NEGATIVE_CASES, each with an error code the run must observe. proof_body_leakage must raise SYMBOL_CLASSIFIER_PROOF_BODY_FORBIDDEN: observations may carry hashes and extracted refs, bounded evidence bodies. private_source_ref_leakage must raise SYMBOL_CLASSIFIER_PRIVATE_SOURCE_REF_FORBIDDEN. theorem_correctness_overclaim must raise SYMBOL_CLASSIFIER_THEOREM_CORRECTNESS_OVERCLAIM, because a library-prior classifier is not theorem-correctness authority. missing_escalation_for_undeclared_symbol must raise SYMBOL_CLASSIFIER_UNDECLARED_LIBRARY_PRIOR_NOT_ESCALATED when an undeclared symbol is not routed to bridge_escalate. premise_budget_precedence_violation must raise SYMBOL_CLASSIFIER_PREMISE_BUDGET_PRECEDENCE when a cited-unallowed row is instead escalated as a residual. allowed_symbol_false_positive must raise SYMBOL_CLASSIFIER_ALLOWED_SYMBOL_FALSE_POSITIVE when a symbol already admitted by allowed_premise_ids is quarantined. unqualified_symbol_overclaim must raise SYMBOL_CLASSIFIER_UNQUALIFIED_SYMBOL_OVERCLAIM when an unqualified token is used to support the qualified library-prior class. A run that fails to observe any expected case is blocked.
\n
Prior Art Grounding
\n
Premise control and library-aware proof search
\n
It sits in the Sledgehammer and Mathlib lineage of premise control, turning the observation that accepted proofs lean on large libraries into a boundary check on out-of-recipe symbols. This check sits in the formal-methods lineage of premise control and library-aware proof search. Isabelle's Sledgehammer makes relevant-fact selection an explicit part of automated proof search, and Lean and Mathlib practice makes clear that accepted proofs can depend on a large library context. The component turns that observation into a boundary check: an accepted artifact is not enough if it quietly used symbols outside the declared premise set. It classifies the symbol-budget breach without judging whether the theorem is true.
\n
Isabelle Sledgehammer and relevant-fact selection: https://isabelle.in.tum.de/doc/sledgehammer.pdf
Lean community Mathlib overview: https://leanprover-community.github.io/mathlib-overview.html
Lean 4 tactic and proof environment context: https://lean-lang.org/theorem_proving_in_lean4/Tactics/
\n
Validation Result record Path
\n
Run from microcosm-substrate:
\n
What a pass certifies
\n
A pass means the fixture still classifies 11 premises and three observations into the expected three classes, the negatives are observed, the secret scan is clean, and the projection edges hold. A pass means the fixture still classifies 11 premises and 3 observations into one undeclared-library prior, one premise-budget retry, and one advisory case, the seven negative cases are observed, the secret-exclusion scan has no blocking hits, and the generated projection reports Mermaid available_from_capsule_edges, Atlas linked_from_capsule_edges, and 19 relationship edges. The validation result records written by _write_receipts carry the same counts and the authority flags. A pass is bounded classifier evidence, not launch or whole-system correctness.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
Over a copied premise index and pre-extracted observations it computes which cite a known symbol outside the allowed set, separating that residual breach from a precedence-taking budget breach. The strongest claim this supports is narrow: over a copied, digest-bearing premise index and pre-extracted qualified symbol observations, it computes which observations cite a known library symbol outside the allowed premise set, and it separates that residual breach from an explicit premise-budget breach with the budget case taking precedence. That is the proof boundary. Everything the classifier reports is recomputed from premise rows and observations, not echoed from the observation's own assertion.
\n
What the scope limit excludes
\n
The ceiling excludes running Lean, proving correctness, reading proof bodies, importing non-public refs, or claiming Mathlib availability, and treats no standard-library declaration as an implicit allowed prior. The scope limit excludes the rest. It does not run Lean or Lake, prove formal-result correctness, read or export proof bodies, use external model services, import non-public source refs, or claim Mathlib availability. It does not treat every standard library or Mathlib declaration as an allowed prior, and it excludes launch or public sharing. Unknown or unqualified tokens stay outside the positive library-prior class unless the public observation and closed premise index make the boundary explicit. The exported source-body check is provenance by digest, size, and anchor, not semantic equivalence for any private source it was copied from.
\n",
"components": [
{
"anchor": "component-undeclared_library_prior_symbol_classifier",
"component_id": "undeclared_library_prior_symbol_classifier",
"public_label": "Undeclared Library Prior Symbol Classifier"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 6
},
"link_count": 18,
"links": [
{
"id": "undeclared_library_prior_symbol_classifier",
"kind": "organ",
"kind_label": "component",
"label": "Undeclared Library Prior Symbol Classifier",
"relation": "explains",
"url": "component-undeclared-library-prior-symbol-classifier.html"
},
{
"id": "mechanism.undeclared_library_prior_symbol_classifier.validates_public_symbol_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public symbol boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-undeclared-library-prior-symbol-classifier-validates-public-symbol-boundary"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.corpus_readiness_mathlib_absence_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Corpus Readiness Mathlib Absence Gate",
"relation": "depends on",
"url": "paper-module-corpus-readiness-mathlib-absence-gate.html"
},
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depends on",
"url": "paper-module-tactic-portfolio-availability.html"
},
{
"id": "paper_module.lean_std_premise_index",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Lean/Std Premise Index",
"relation": "depends on",
"url": "paper-module-lean-std-premise-index.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.provider_context_recipe_budget",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Provider Context Recipe Budget",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[54:paper_module.provider_context_recipe_budget].depends_on",
"url": "paper-module-provider-context-recipe-budget.html"
}
],
"paper_module_id": "paper_module.undeclared_library_prior_classifier",
"reader_packet": {
"abstract": "Undeclared Library Prior Classifier scores extracted Lean symbol observations against an allowed premise set without running Lean or treating libraries as implicit allowlists.",
"authority_ceiling": "Symbol-boundary classification fixture over copied Lean/Std premise rows and pre-extracted symbol observations only; no proof-source reads, Lean or Lake execution, formal-result correctness, whole-library implicit allowlist, Mathlib availability, external model access, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.undeclared_library_prior_classifier.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-undeclared-library-prior-classifier-diagram-1",
"html_url": "docs/paper-module-undeclared-library-prior-classifier.html#paper-module-undeclared-library-prior-classifier-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD inputs[\"Copied premise index and symbol observations\"] policy[\"Classifier policy regex and outcomes\"] classify[\"_classify_row compute allowed vs undeclared\"] budget[\"cited_unallowed_premise_ids present\"] residual[\"Known symbol outside allowed set\"] clean[\"Allowed or no known undeclared symbol\"] retry[\"PREMISE_BUDGET_VIOLATION route retry\"] escalate[\"UNDECLARED_LIBRARY_PRIOR route bridge_escalate\"] advisory[\"NONE route accept_as_advisory\"] records[\"Result, board, validation, sign-off records\"] inputs --> classify policy --> classify classify --> budget classify --> residual classify --> clean budget --> retry residual --> escalate clean --> advisory retry --> records escalate --> records advisory --> records",
"source_format": "mermaid",
"source_hash": "sha256:2b9462f2dfb7e3fab81336758827c2fcbd8ff821a61ca8d43e1d439ea77f4264",
"source_line_end": 167,
"source_line_start": 144
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.corpus_readiness_mathlib_absence_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Corpus Readiness Mathlib Absence Gate",
"relation": "depends on",
"url": "paper-module-corpus-readiness-mathlib-absence-gate.html"
},
{
"id": "paper_module.tactic_portfolio_availability",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Tactic Portfolio Availability",
"relation": "depends on",
"url": "paper-module-tactic-portfolio-availability.html"
},
{
"id": "paper_module.lean_std_premise_index",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Lean/Std Premise Index",
"relation": "depends on",
"url": "paper-module-lean-std-premise-index.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "undeclared_library_prior_symbol_classifier",
"kind": "organ",
"kind_label": "component",
"label": "Undeclared Library Prior Symbol Classifier",
"relation": "explains",
"url": "component-undeclared-library-prior-symbol-classifier.html"
},
{
"id": "mechanism.undeclared_library_prior_symbol_classifier.validates_public_symbol_boundary",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public symbol boundary",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-undeclared-library-prior-symbol-classifier-validates-public-symbol-boundary"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.undeclared_library_prior_classifier",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 274,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/undeclared_library_prior_classifier.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It answers one question about a Lean-accepted proof: did the candidate cite a known library symbol outside the premises its recipe allowed?** A theorem prover can return a proof that Lean accepts and still break the rules of the evaluation it ran under. The usual way this happens is quiet: the proof reaches for a library lemma the recipe never put on the table. The symbol is real and the proof is sound, but the run used a fact it was not allowed to assume. This component answers one question about that gap. Given the premises a candidate was allowed to use and the qualified symbols it actually reached for, did it cite a known library symbol outside the allowed set?",
"level": 4,
"order": 1,
"section_id": "declared-versus-undeclared-library-priors",
"source_line_end": 13,
"source_line_start": 3,
"title": "Declared versus undeclared library priors"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It works only from a copied premise index and pre-extracted observations, never running Lean or reading the proof, with the allowed set closed by construction from premise rows.** It works only from a copied premise index and a list of symbol observations that were extracted beforehand. It does not run Lean, it does not read the proof body, and it does not treat the standard library as an implicit allowlist. The allowed set is closed by construction: a symbol is admissible because a premise row names it, never because it happens to exist in Lean's standard library.",
"level": 4,
"order": 2,
"section_id": "it-works-only-from-a-copied-premise-index-and-pre-extracted-observations",
"source_line_end": 23,
"source_line_start": 14,
"title": "It works only from a copied premise index and pre-extracted observations"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The public fixture's 11 premises and three observations classify to one undeclared-prior escalation, one budget-violation retry, and one advisory none, each named with its rule and class.** The public fixture carries 11 premise rows across four namespaces (Bool, Iff, List, Nat) and three symbol observations. Running it classifies one observation as UNDECLAREDLIBRARYPRIOR routed to bridge_escalate, one as PREMISEBUDGETVIOLATION routed to retry, and one as an advisory NONE. The written records name each symbol, the premise rule it came from, and the class that was computed.",
"level": 4,
"order": 3,
"section_id": "the-public-fixture-s-11-premises",
"source_line_end": 34,
"source_line_start": 24,
"title": "The public fixture's 11 premises"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 36,
"source_line_start": 35,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A checked proof is not a proof that stayed inside the allowed building blocks, and a loose check cannot tell an in-recipe proof from one that borrowed an out-of-recipe lemma.** The problem is that \"the proof checked\" does not mean \"the proof stayed inside the allowed building blocks.\" A loose check that trusts the prover's sign-off, or that admits any symbol found in the standard library, cannot tell an in-recipe proof from one that quietly borrowed an out-of-recipe lemma. Both compile. Both look sound.",
"level": 4,
"order": 5,
"section_id": "compiling-does-not-mean-in-recipe",
"source_line_end": 46,
"source_line_start": 37,
"title": "Compiling does not mean in-recipe"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**An explicit budget breach from a disallowed premise id is kept distinct from a residual breach using an undeclared known symbol, with the budget case taking precedence.** The component also keeps two failure modes apart. An explicit budget breach, where the candidate names a premise id the recipe did not allow, is not the same as a residual breach, where the candidate used a known symbol that turns out to be undeclared. The first is settled directly from the cited ids and takes precedence. The second is what the symbol comparison is for. Folding both into one class would either over-escalate honest retries or let genuine out-of-recipe library use pass as a budget note.",
"level": 4,
"order": 6,
"section_id": "budget-breach-kept-apart-from-residual-breach",
"source_line_end": 58,
"source_line_start": 47,
"title": "Budget breach kept apart from residual breach"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 60,
"source_line_start": 59,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run builds the fixture result with negatives included, and runsymbol_bundle runs the same builder over an exported source-body bundle and can reuse a fresh prior record. The fixture entrypoint is run. It calls build_result with the fixture input mode and negative cases included, then writereceipts. The exported-bundle entrypoint is runsymbolbundle, which runs the same builder over an exported source-body bundle and can reuse a fresh prior record. Inside buildresult the stages run in this order.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-into-build-result",
"source_line_end": 70,
"source_line_start": 61,
"title": "Two entry points into `_build_result`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "loadpayloads reads the four inputs plus seven negative-case files, and scanpaths runs a secret-exclusion pass before the four validators run. loadpayloads reads the four input files (projection_protocol.json, premiseindex.json, symbolobservations.json, classifier_policy.json) plus the seven negative-case files. scan_paths runs a secret-exclusion pass over those inputs and records a blocking hit count. Then four validators run.",
"level": 4,
"order": 9,
"section_id": "loading-inputs-and-the-secret-scan",
"source_line_end": 79,
"source_line_start": 71,
"title": "Loading inputs and the secret scan"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Three validators check the projection's refs and digests, require every premise row to carry an id, qualified name, and source ref, and pin the symbol regex and review outcomes. validateprojection_protocol checks that the projection cites real source refs, source pattern ids, target refs, and source digests, and that any omitted proof or provider material carries an omission record. validatepremiseindex reads the premise rows, requires each to have a premise id, a qualified theoremordef_name, and a public source ref, and reports the premise and namespace counts. validateclassifierpolicy requires the declared qualified-symbol pattern to equal QUALIFIEDSYMBOLRE (the regex over the Nat, List\u2026",
"level": 4,
"order": 10,
"section_id": "validating-projection-premises-and-policy",
"source_line_end": 94,
"source_line_start": 80,
"title": "Validating projection, premises, and policy"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "classifyrow maps cited premises to symbols, finds the known ones outside the allowed and cited sets, and applies budget-over-residual precedence, flagging any drift from the asserted class. validatesymbol_observations is the classifier. For each observation it calls classifyrow, which is where the rule lives. classifyrow reads allowedpremiseids and citedunallowedpremise_ids, maps them through the premise index to their symbols, extracts the observed qualified refs, and keeps the known ones (those a premise row names). The undeclared set is the known symbols that are neither allowed nor already cited as unallowed. Precedence is explicit: if citedunallowedpremise_ids is present the row is PRE\u2026",
"level": 4,
"order": 11,
"section_id": "classify-row-computes-the-class-with-precedence",
"source_line_end": 111,
"source_line_start": 95,
"title": "`_classify_row` computes the class with precedence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "For the exported bundle validatesourcemodulemanifest rejects in-result record bodies and verifies each source body against its digests, counts, and anchors as provenance, not a second proof. For the exported bundle, validatesourcemodulemanifest checks sourcemodulemanifest.json, rejects manifest-level or row-level bodyinreceipt: true, and verifies each declared source-body import against its source and target digests, line counts, byte counts, and required anchors. This is a provenance check on copied bodies, not a second proof.",
"level": 4,
"order": 12,
"section_id": "the-exported-bundle-s-provenance-check",
"source_line_end": 121,
"source_line_start": 112,
"title": "The exported bundle's provenance check"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "writereceipts writes the result, board, validation, and sign-off records, and resultcard projects a pass or fail card that omits the heavier evidence. writereceipts writes the result, board, validation, and sign-off records. result_card projects a compact pass or fail card that omits the premise rows, classification rows, source digests, secret-scan detail, and scope limit body, so the card can be shown without exporting the heavier evidence.",
"level": 4,
"order": 13,
"section_id": "receipts-and-the-compact-card",
"source_line_end": 168,
"source_line_start": 122,
"title": "Result records and the compact card"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 170,
"source_line_start": 169,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Seven negative cases each carry a code the run must observe, from leaked proof bodies and non-public refs to correctness overclaim, missing escalation, precedence violations, and false positives. The fixture ships seven negative cases in EXPECTEDNEGATIVE_CASES, each with an error code the run must observe. proofbodyleakage must raise SYMBOLCLASSIFIERPROOFBODYFORBIDDEN: observations may carry hashes and extracted refs, bounded evidence bodies. privatesourceref_leakage must raise SYMBOLCLASSIFIERPRIVATESOURCEREFFORBIDDEN. theoremcorrectness_overclaim must raise SYMBOLCLASSIFIERTHEOREMCORRECTNESSOVERCLAIM, because a library-prior classifier is not theorem-correctness authority. missingescala\u2026",
"level": 4,
"order": 15,
"section_id": "seven-negative-cases-and-their-codes",
"source_line_end": 192,
"source_line_start": 171,
"title": "Seven negative cases and their codes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 194,
"source_line_start": 193,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**It sits in the Sledgehammer and Mathlib lineage of premise control, turning the observation that accepted proofs lean on large libraries into a boundary check on out-of-recipe symbols.** This check sits in the formal-methods lineage of premise control and library-aware proof search. Isabelle's Sledgehammer makes relevant-fact selection an explicit part of automated proof search, and Lean and Mathlib practice makes clear that accepted proofs can depend on a large library context. The component turns that observation into a boundary check: an accepted artifact is not enough if it quietly used symbols outside the declared premise set. It classifies the symbol-budget breach without judging wh\u2026",
"level": 4,
"order": 17,
"section_id": "premise-control-and-library-aware-proof-search",
"source_line_end": 213,
"source_line_start": 195,
"title": "Premise control and library-aware proof search"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from microcosm-substrate:",
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 232,
"source_line_start": 214,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the fixture still classifies 11 premises and three observations into the expected three classes, the negatives are observed, the secret scan is clean, and the projection edges hold.** A pass means the fixture still classifies 11 premises and 3 observations into one undeclared-library prior, one premise-budget retry, and one advisory case, the seven negative cases are observed, the secret-exclusion scan has no blocking hits, and the generated projection reports Mermaid availablefromcapsule_edges, Atlas linkedfromcapsule_edges, and 19 relationship edges. The validation result records written by writereceipts carry the same counts and the authority flags. A pass is bounded class\u2026",
"level": 4,
"order": 19,
"section_id": "what-a-pass-certifies",
"source_line_end": 245,
"source_line_start": 233,
"title": "What a pass certifies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 247,
"source_line_start": 246,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Over a copied premise index and pre-extracted observations it computes which cite a known symbol outside the allowed set, separating that residual breach from a precedence-taking budget breach.** The strongest claim this supports is narrow: over a copied, digest-bearing premise index and pre-extracted qualified symbol observations, it computes which observations cite a known library symbol outside the allowed premise set, and it separates that residual breach from an explicit premise-budget breach with the budget case taking precedence. That is the proof boundary. Everything the classifier reports is recomputed from premise rows and observations, not echoed from the observation's own asse\u2026",
"level": 4,
"order": 21,
"section_id": "the-strongest-supported-claim",
"source_line_end": 259,
"source_line_start": 248,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The ceiling excludes running Lean, proving correctness, reading proof bodies, importing non-public refs, or claiming Mathlib availability, and treats no standard-library declaration as an implicit allowed prior.** The scope limit excludes the rest. It does not run Lean or Lake, prove formal-result correctness, read or export proof bodies, use external model services, import non-public source refs, or claim Mathlib availability. It does not treat every standard library or Mathlib declaration as an allowed prior, and it excludes launch or public sharing. Unknown or unqualified tokens stay outside the positive library-prior class unless the public observation and closed premise index make th\u2026",
"level": 4,
"order": 22,
"section_id": "what-the-authority-ceiling-excludes",
"source_line_end": 273,
"source_line_start": 260,
"title": "What the scope limit excludes"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/undeclared_library_prior_classifier.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Undeclared Library Prior Classifier scores extracted Lean symbol observations against an allowed premise set without running Lean or treating libraries as implicit allowlists."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "undeclared_library_prior_symbol_classifier.py",
"path": "src/microcosm_core/organs/undeclared_library_prior_symbol_classifier.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for undeclared-library prior classification, premise-index validation, source-module digest checks, symbol observation classification, negative cases, and scope limits.",
"symbols": [
"run",
"run_symbol_bundle",
"validate_projection_protocol",
"validate_premise_index",
"validate_classifier_policy",
"validate_source_module_manifest",
"validate_symbol_observations",
"_classify_row"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/undeclared_library_prior_symbol_classifier.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/undeclared_library_prior_classifier.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/undeclared_library_prior_classifier.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Undeclared Library Prior Classifier validates copied Lean/Std premise rows and pre-extracted symbol observations, classifying undeclared library priors and premise-budget violations with route outcomes, source-module digest checks, secret-exclusion scans, negative cases, and scope limits. It does not read proof source, run Lean or Lake, prove formal-result correctness, treat the whole standard library as an implicit allowlist, claim Mathlib availability, use external model services, launch, publish, or prove whole-system correctness.",
"summary_line": "Undeclared Library Prior Classifier scores extracted Lean symbol observations against an allowed premise set without running Lean or treating libraries as implicit allowlists.",
"title": "Undeclared Library Prior Classifier"
},
{
"anchor": "paper-module-voice-to-doctrine-self-improvement-loop",
"authority_ceiling": "Declared lesson-propagation fixture only; no source notes export, non-public body export, model-output data export, source or doctrine mutation authority, global-promotion authority, live work log mutation, publishing-scope decision, launch-scope decision, external model access, whole-system equivalence, or whole-system correctness.",
"body_html": "
It decides from the files alone whether each lesson in a self-improvement pass actually moved a named surface or only left a note claiming it did. This component checks whether a claim of self-improvement is backed by evidence. The system it comes from turns a local lesson into a durable change: a piece of pressure is sensed, classified, assigned to an owner surface, changed or captured there, validated, and closed with a condition for revisiting it later. This component takes a folder describing one pass of that loop and decides, from the files alone, whether each lesson really moved a named surface or only produced a note that says it did.
\n
It runs on local files and emits metadata-only records, requiring every lesson's refs to resolve to real files and reporting only ids, counts, hashes, and status. It runs on local files and writes metadata-only result records. Every lesson names the surface it changed, the action it took, and the refs that show the change held. Each ref must resolve to a real file. The result records report ids, counts, hashes, and pass or fail status, never the private text behind them.
\n
The run entry point validates the test fixture, while run_voice_to_doctrine_bundle validates the exported example bundle instead. The entry point is run in src/microcosm_core/organs/voice_to_doctrine_self_improvement_loop.py. A second entry point, run_voice_to_doctrine_bundle, validates the exported example bundle instead of the test fixture.
\n
Purpose
\n
Learning is easy to assert, hard to back
\n
A claim that the system learned is easy to assert, so this component recomputes the verdict from evidence and ignores any label a lesson gives itself. "The system learned from its work" is easy to assert and hard to back. A log line, a closed ticket, and a confident summary all read as progress. If a check accepts that language at face value, a run can pass by describing success rather than showing it. This component refuses that. It recomputes the verdict from the evidence and ignores any label a lesson gives itself.
\n
Three trusts a loose version would make
\n
A loose check would trust a lesson's stated status, an unverified ref, and a refinement that changed nothing, and the checks below close all three gaps. A loose version would trust three things this one does not: the lesson's own stated status, a ref that names a file without checking the file exists, and a refinement that changed nothing. Each of those lets an unbacked claim through. The checks below close all three.
\n
How it works
\n
run loads inputs and five validators
\n
run loads four input files plus six negative-case files, then runs five validators in order and merges their findings.run loads four input files from the chosen directory: projection_protocol.json, propagation_policy.json, owner_surfaces.json, and local_lessons.json, plus six negative-case files. It then runs five validators in order and merges their findings.
\n
validate_projection_protocol checks three source refs
\n
validate_projection_protocol requires the three source pattern refs the loop is built from, recording any missing one as a required-source-pattern-ref-missing code.validate_projection_protocol checks that the protocol declares the three source pattern refs the loop is built from: recursive_self_improvement_operating_loop, doctrine_population_loop, and local_to_general_propagation. A missing ref is recorded as VOICE_DOCTRINE_REQUIRED_MACRO_PATTERN_REF_MISSING.
\n
validate_policy requires the seven-step sequence
\n
validate_policy requires the full seven-step sequence from sense_local_pressure through publish_reentry_condition and rejects any policy allowing result record-only progress.validate_policy checks that the policy carries the full seven-step sequence, from sense_local_pressure through publish_reentry_condition, and rejects any policy that authorizes result record-only progress.
\n
validate_owner_surfaces maps surface authority
\n
validate_owner_surfaces builds a map where each surface needs a stable id, a public ref, and a declared mutation authority to own a change.validate_owner_surfaces builds a map of owner surfaces. Each surface needs a stable id, a public ref, and a declared mutation authority. A surface that names none of these cannot own a change.
\n
validate_lessons resolves every ref
\n
validate_lessons, the core, requires each lesson's fields, a mapped owner surface, one of four typed outcomes, and every ref resolving through _resolve_ref to an anchor.validate_lessons is the core. For each lesson it checks the required fields are present, the selected owner surface exists in the map, and the outcome is one of four: refined_existing_surface, workitem_captured, nothing_to_refine, or already_propagated_verified. It then resolves every ref the lesson carries (changed_surface_ref, evidence_ref, evidence_refs, validation_ref, closeout_ref) through _resolve_ref, which reads the target file and confirms the anchor text after :: is present. A refinement with no changed surface is result record-only progress and fails. A capture with no re-entry condition fails. A nothing_to_refine with no stewardship and next-best-lane check fails.
\n
Asserted labels cannot rescue a row
\n
A lesson's expected_label or expected_verdict is counted then ignored when the row is not genuinely backed, so a fixture cannot pass by declaring its own success. The defense against self-grading lives here too. A lesson may carry expected_label, expected_status, expected_outcome, or expected_verdict. validate_lessons counts those keys, then computes whether the lesson is genuinely backed from the other checks. If the lesson is not backed, the asserted label is recorded as ignored under VOICE_DOCTRINE_BAKED_EXPECTED_LABEL_IGNORED and cannot rescue the row. A fixture cannot pass by declaring its own success.
_source_module_result confirms each of eight copied source bodies by hash, byte and line count, and anchors, allowing an exact copy unless a row declares a verified light edit._source_module_result validates the copied source bodies listed in source_module_manifest.json. For each of the eight rows it confirms the target file exists, its sha256, byte count, and line count match the manifest, every required anchor is present in both the copied body and the live source file, and the copy is exact unless the row declares a verified light edit. Seven bodies are exact copies; one, the task-ledger skill body, is a verified light edit with the operator's raw gesture removed.
\n
This page is a checked surface
\n
This page is itself a checked surface, with lesson refs pointing at literal anchors here; deleting one leaves a ref unresolved and fails the run, though the fixture resolves nineteen. This page is one of the surfaces the fixture checks. Its lesson rows point evidence and completion refs at named anchors in this file, including Public Mechanics, Shape, Reader Proof Boundary, Body Handling, and Scope boundary. _resolve_ref looks for each anchor as literal text; if an anchor is deleted, the ref does not resolve, validate_lessons records VOICE_DOCTRINE_SURFACE_REF_UNRESOLVED, and the run fails. The fixture resolves nineteen refs with none unresolved.
\n
Secret scan, gating, and record writers
\n
After the validators run secret-scans inputs and copied bodies and passes only on clean findings, then writes three metadata-only records, with the bundle entry requiring the source manifest. After the validators, run runs a secret-exclusion scan over the input paths and copied bodies, merges the observed negative cases, and passes only when there are no blocking findings, no missing negative cases, a clean scan, and verified source modules. _build_board and the result and validation writers then emit voice_to_doctrine_self_improvement_loop_result.json, ..._board.json, and ..._validation_receipt.json, each carrying body_in_receipt: false. run_voice_to_doctrine_bundle calls run with negative cases off and the source-module manifest required, then stamps the bundle id.
\n
What the fixture describes
\n
The fixture spans four lessons over five owner surfaces: two refinements, one deferred capture with a re-entry condition, and one verified no-op. The fixture describes four lessons across five owner surfaces: two refine an existing surface, one captures deferred work with a re-entry condition, and one records a verified no-op.
\n
Shape
\nDiagram of the mechanism (9 steps).Diagram source & refs
Source refs
3 source pattern refs
validate_projection_protocol
flowchart TD Load["Load inputs protocol, policy, owner surfaces, lessons"] Protocol["validate_projection_protocol 3 source pattern refs"] Policy["validate_policy 7-step sequence"] Owners["validate_owner_surfaces id, public ref, mutation authority"] Lessons["validate_lessons fields, owner, outcome, resolve every ref"] Recompute["Ignore expected label recompute from evidence"] Negative["validate_negative_cases 6 refusals"] Source["_source_module_result 8 copied bodies: hash, anchors, exact copy"] Result records["metadata-only result records result, board, validation"] Load --> Protocol Load --> Policy Load --> Owners Owners --> Lessons Lessons --> Recompute Load --> Negative Load --> Source Protocol --> Result records Policy --> Result records Recompute --> Result records Negative --> Result records Source --> Result records
\n
Negative cases
\n
validate_negative_cases reads six files, each a shape the loop must refuse, and records a fixed code for each:validate_negative_cases reads six files, each a shape the loop must refuse, and records a fixed code for each:
\n
raw_operator_voice_export carrying a forbidden body key gives VOICE_DOCTRINE_RAW_OPERATOR_BODY_FORBIDDEN.
doctrine_node_hand_edit that edits a principle, concept, mechanism, or axiom by direct file write gives VOICE_DOCTRINE_DIRECT_NODE_EDIT_FORBIDDEN.
consume_without_deposit with no deposit outcome gives VOICE_DOCTRINE_CONSUME_WITHOUT_DEPOSIT.
pattern_receipt_only_progress whose owner action is append-result record-only gives VOICE_DOCTRINE_RECEIPT_ONLY_PROGRESS.
global_promotion_without_owner_validation that requests promotion without owner validation gives VOICE_DOCTRINE_GLOBAL_PROMOTION_WITHOUT_OWNER_VALIDATION.
private_thread_body_export carrying a forbidden body key gives VOICE_DOCTRINE_PRIVATE_THREAD_BODY_FORBIDDEN.
\n
Fixture, bundle, and tampering modes
\n
In fixture mode all six cases must be observed without blocking, in bundle mode none is expected, and source-module tampering each drops the verified-body count and adds a blocking code. In fixture mode these six cases are expected: they must all be observed for the run to pass, and they do not block. In bundle mode the negative inputs are absent and none is expected. The source-module tests also exercise tampering: a changed hash, a rehashed body edit, a wrong source ref, and a missing live source each drop the verified-body count and add a blocking code.
\n
Prior Art Grounding
\n
After-action review and pattern language
\n
It follows after-action review and pattern-language practice, but a lesson becomes durable only with evidence, an owner surface, a validation ref, and a bounded re-entry path. This component follows after-action review, lessons-learned practice, and pattern-language work. NASA's Lessons Learned Information System is a public example of preserving operational lessons so later work can reuse them, and pattern-language practice gives a vocabulary for turning repeated local solutions into named reusable forms. This component keeps that direction without collapsing a local lesson into doctrine: a lesson becomes durable only when it has evidence, an owner surface, a validation ref, and a bounded re-entry path.
\n
NASA Lessons Learned Information System: https://llis.nasa.gov/
Pattern language background: https://hillside.net/patterns/
\n
Validation Result record Path
\n
Run from microcosm-substrate:
\n
What a green record means
\n
A green record means the lessons named a real owner surface, resolved every ref, observed the negative cases, and passed the secret scan, and nothing beyond that. A green fixture or bundle record means the lessons named a real owner surface, resolved every ref, observed the negative cases, and passed the secret scan. It means nothing beyond that.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
On a local folder it can require each lesson to name a surface, resolve every ref, land a typed outcome, and survive six refusals, with the bundle adding an eight-body copy floor. The strongest claim this component supports: on a local folder, it can require that each lesson name an owner surface, resolve every evidence and completion ref to a real file, land one of four typed outcomes, and survive the six refusals, while ignoring any label the lesson asserts about itself. The exported bundle adds a body floor: eight copied source bodies verified by hash, byte count, line count, and anchor, with the raw source text kept out of the result records.
\n
What it refuses
\n
It exports no raw voice, thread bodies, or payloads and grants no doctrine mutation, promotion, or launch, keeping the proof boundary at the fixture, bundle, and copied-body manifest. It refuses the rest. It does not export source notes, private thread bodies, or model-output data. It does not grant doctrine mutation, direct edits to doctrine nodes, or global promotion. It does not perform live mutation of tracked work, external model access, source-file changes, launch, or public sharing. A pass is bounded validator evidence, bounded evidence that future lessons will generalize correctly and not whole-system correctness. The proof boundary and scope limit are the fixture, the exported bundle, the copied-body manifest, and the checks named above.
\n",
"components": [
{
"anchor": "component-voice_to_doctrine_self_improvement_loop",
"component_id": "voice_to_doctrine_self_improvement_loop",
"public_label": "Voice To Doctrine Self Improvement Loop"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 2,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 15,
"links": [
{
"id": "voice_to_doctrine_self_improvement_loop",
"kind": "organ",
"kind_label": "component",
"label": "Voice To Doctrine Self Improvement Loop",
"relation": "explains",
"url": "component-voice-to-doctrine-self-improvement-loop.html"
},
{
"id": "mechanism.voice_to_doctrine_self_improvement_loop.validates_public_voice_to_doctrine_self_improvement_loop",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public voice to doctrine self improvement loop",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-voice-to-doctrine-self-improvement-loop-validates-public-voice-to-doctrine-self-improvement-loop"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-7",
"kind": "principle",
"kind_label": "principle",
"label": "Track known unknowns without claiming the unknown is mapped",
"relation": "governed by",
"url": "doctrine.html#dcard-p-7"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
},
{
"id": "concept.voice_to_doctrine_self_improvement_loop_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "local pressure rows as typed candidates for owner-surface refinement",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-voice-to-doctrine-self-improvement-loop-bundle"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.pattern_assimilation",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Assimilation",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[71:paper_module.pattern_assimilation].depends_on",
"url": "paper-module-pattern-assimilation.html"
}
],
"paper_module_id": "paper_module.voice_to_doctrine_self_improvement_loop",
"reader_packet": {
"abstract": "Voice to Doctrine Self-Improvement Loop validates lesson propagation without exporting source notes or granting doctrine mutation authority.",
"authority_ceiling": "Declared lesson-propagation fixture only; no source notes export, non-public body export, model-output data export, source or doctrine mutation authority, global-promotion authority, live work log mutation, publishing-scope decision, launch-scope decision, external model access, whole-system equivalence, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.voice_to_doctrine_self_improvement_loop.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-voice-to-doctrine-self-improvement-loop-diagram-1",
"html_url": "docs/paper-module-voice-to-doctrine-self-improvement-loop.html#paper-module-voice-to-doctrine-self-improvement-loop-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Load[\"Load inputs protocol, policy, owner surfaces, lessons\"] Protocol[\"validate_projection_protocol 3 source pattern refs\"] Policy[\"validate_policy 7-step sequence\"] Owners[\"validate_owner_surfaces id, public ref, mutation authority\"] Lessons[\"validate_lessons fields, owner, outcome, resolve every ref\"] Recompute[\"Ignore expected label recompute from evidence\"] Negative[\"validate_negative_cases 6 refusals\"] Source[\"_source_module_result 8 copied bodies: hash, anchors, exact copy\"] Result records[\"metadata-only result records result, board, validation\"] Load --> Protocol Load --> Policy Load --> Owners Owners --> Lessons Lessons --> Recompute Load --> Negative Load --> Source Protocol --> Result records Policy --> Result records Recompute --> Result records Negative --> Result records Source --> Result records",
"source_format": "mermaid",
"source_hash": "sha256:8578741b5d16b562d7c4c2a137cfffb1c93428c4b9577b7ac018a1f19e091e6a",
"source_line_end": 93,
"source_line_start": 69
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "voice_to_doctrine_self_improvement_loop",
"kind": "organ",
"kind_label": "component",
"label": "Voice To Doctrine Self Improvement Loop",
"relation": "explains",
"url": "component-voice-to-doctrine-self-improvement-loop.html"
},
{
"id": "mechanism.voice_to_doctrine_self_improvement_loop.validates_public_voice_to_doctrine_self_improvement_loop",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public voice to doctrine self improvement loop",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-voice-to-doctrine-self-improvement-loop-validates-public-voice-to-doctrine-self-improvement-loop"
}
],
"relation": "explains"
},
{
"count": 6,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-7",
"kind": "principle",
"kind_label": "principle",
"label": "Track known unknowns without claiming the unknown is mapped",
"relation": "governed by",
"url": "doctrine.html#dcard-p-7"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "concept.voice_to_doctrine_self_improvement_loop_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "local pressure rows as typed candidates for owner-surface refinement",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-voice-to-doctrine-self-improvement-loop-bundle"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.voice_to_doctrine_self_improvement_loop",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 154,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/voice_to_doctrine_self_improvement_loop.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "It decides from the files alone whether each lesson in a self-improvement pass actually moved a named surface or only left a note claiming it did. This component checks whether a claim of self-improvement is backed by evidence. The system it comes from turns a local lesson into a durable change: a piece of pressure is sensed, classified, assigned to an owner surface, changed or captured there, validated, and closed with a condition for revisiting it later. This component takes a folder describing one pass of that loop and decides, from the files alone, whether each lesson really moved a named surface or only produced a note that says it did.",
"level": 4,
"order": 1,
"section_id": "it-decides-from-the-files-alone-whether-each-lesson-in",
"source_line_end": 6,
"source_line_start": 3,
"title": "It decides from the files alone whether each lesson in"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It runs on local files and emits metadata-only records, requiring every lesson's refs to resolve to real files and reporting only ids, counts, hashes, and status. It runs on local files and writes metadata-only result records. Every lesson names the surface it changed, the action it took, and the refs that show the change held. Each ref must resolve to a real file. The result records report ids, counts, hashes, and pass or fail status, never the private text behind them.",
"level": 4,
"order": 2,
"section_id": "it-runs-on-local-files-and-emits-body-free-records",
"source_line_end": 10,
"source_line_start": 7,
"title": "It runs on local files and emits metadata-only records"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The run entry point validates the test fixture, while runvoicetodoctrinebundle validates the exported example bundle instead. The entry point is run in src/microcosmcore/components/voicetodoctrineselfimprovementloop.py. A second entry point, runvoicetodoctrinebundle, validates the exported example bundle instead of the test fixture.",
"level": 4,
"order": 3,
"section_id": "the-run-entry-point-validates-the-test-fixture",
"source_line_end": 14,
"source_line_start": 11,
"title": "The `run` entry point validates the test fixture"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A claim that the system learned is easy to assert, so this component recomputes the verdict from evidence and ignores any label a lesson gives itself. \"The system learned from its work\" is easy to assert and hard to back. A log line, a closed ticket, and a confident summary all read as progress. If a check accepts that language at face value, a run can pass by describing success rather than showing it. This component refuses that. It recomputes the verdict from the evidence and ignores any label a lesson gives itself.",
"level": 4,
"order": 5,
"section_id": "learning-is-easy-to-assert-hard-to-back",
"source_line_end": 20,
"source_line_start": 17,
"title": "Learning is easy to assert, hard to back"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A loose check would trust a lesson's stated status, an unverified ref, and a refinement that changed nothing, and the checks below close all three gaps. A loose version would trust three things this one does not: the lesson's own stated status, a ref that names a file without checking the file exists, and a refinement that changed nothing. Each of those lets an unbacked claim through. The checks below close all three.",
"level": 4,
"order": 6,
"section_id": "three-trusts-a-loose-version-would-make",
"source_line_end": 24,
"source_line_start": 21,
"title": "Three trusts a loose version would make"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run loads four input files plus six negative-case files, then runs five validators in order and merges their findings. run loads four input files from the chosen directory: projectionprotocol.json, propagationpolicy.json, ownersurfaces.json, and locallessons.json, plus six negative-case files. It then runs five validators in order and merges their findings.",
"level": 4,
"order": 8,
"section_id": "run-loads-inputs-and-five-validators",
"source_line_end": 30,
"source_line_start": 27,
"title": "`run` loads inputs and five validators"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateprojectionprotocol requires the three source pattern refs the loop is built from, recording any missing one as a required-source-pattern-ref-missing code. validateprojectionprotocol checks that the protocol declares the three source pattern refs the loop is built from: recursiveselfimprovementoperatingloop, doctrinepopulationloop, and localtogeneralpropagation. A missing ref is recorded as VOICEDOCTRINEREQUIREDMACROPATTERNREF_MISSING.",
"level": 4,
"order": 9,
"section_id": "validate-projection-protocol-checks-three-macro-refs",
"source_line_end": 34,
"source_line_start": 31,
"title": "`validate_projection_protocol` checks three source refs"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatepolicy requires the full seven-step sequence from senselocalpressure through publishreentrycondition and rejects any policy allowing result record-only progress. validatepolicy checks that the policy carries the full seven-step sequence, from senselocalpressure through publishreentrycondition, and rejects any policy that authorizes result record-only progress.",
"level": 4,
"order": 10,
"section_id": "validate-policy-requires-the-seven-step-sequence",
"source_line_end": 38,
"source_line_start": 35,
"title": "`validate_policy` requires the seven-step sequence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validateownersurfaces builds a map where each surface needs a stable id, a public ref, and a declared mutation authority to own a change. validateownersurfaces builds a map of owner surfaces. Each surface needs a stable id, a public ref, and a declared mutation authority. A surface that names none of these cannot own a change.",
"level": 4,
"order": 11,
"section_id": "validate-owner-surfaces-maps-surface-authority",
"source_line_end": 42,
"source_line_start": 39,
"title": "`validate_owner_surfaces` maps surface authority"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatelessons, the core, requires each lesson's fields, a mapped owner surface, one of four typed outcomes, and every ref resolving through resolveref to an anchor. validatelessons is the core. For each lesson it checks the required fields are present, the selected owner surface exists in the map, and the outcome is one of four: refinedexistingsurface, workitemcaptured, nothingtorefine, or alreadypropagatedverified. It then resolves every ref the lesson carries (changedsurfaceref, evidenceref, evidencerefs, validationref, completionref) through resolveref, which reads the target file and confirms the anchor text after :: is present. A refinement with no changed surface is result record-on\u2026",
"level": 4,
"order": 12,
"section_id": "validate-lessons-resolves-every-ref",
"source_line_end": 46,
"source_line_start": 43,
"title": "`validate_lessons` resolves every ref"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A lesson's expectedlabel or expectedverdict is counted then ignored when the row is not genuinely backed, so a fixture cannot pass by declaring its own success. The defense against self-grading lives here too. A lesson may carry expectedlabel, expectedstatus, expectedoutcome, or expectedverdict. validatelessons counts those keys, then computes whether the lesson is genuinely backed from the other checks. If the lesson is not backed, the asserted label is recorded as ignored under VOICEDOCTRINEBAKEDEXPECTEDLABELIGNORED and cannot rescue the row. A fixture cannot pass by declaring its own success.",
"level": 4,
"order": 13,
"section_id": "asserted-labels-cannot-rescue-a-row",
"source_line_end": 50,
"source_line_start": 47,
"title": "Asserted labels cannot rescue a row"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourcemoduleresult confirms each of eight copied source bodies by hash, byte and line count, and anchors, allowing an exact copy unless a row declares a verified light edit. sourcemoduleresult validates the copied source bodies listed in sourcemodule_manifest.json. For each of the eight rows it confirms the target file exists, its sha256, byte count, and line count match the manifest, every required anchor is present in both the copied body and the live source file, and the copy is exact unless the row declares a verified light edit. Seven bodies are exact copies; one, the task-ledger skill body, is a verified light edit with the operator's raw gesture removed.",
"level": 4,
"order": 14,
"section_id": "source-module-result-verifies-eight-copied-bodies",
"source_line_end": 54,
"source_line_start": 51,
"title": "`_source_module_result` verifies eight copied bodies"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "This page is itself a checked surface, with lesson refs pointing at literal anchors here; deleting one leaves a ref unresolved and fails the run, though the fixture resolves nineteen. This page is one of the surfaces the fixture checks. Its lesson rows point evidence and completion refs at named anchors in this file, including Public Mechanics, Shape, Reader Proof Boundary, Body Handling, and Scope boundary. resolveref looks for each anchor as literal text; if an anchor is deleted, the ref does not resolve, validatelessons records VOICEDOCTRINESURFACEREF_UNRESOLVED, and the run fails. The fixture resolves nineteen refs with none unresolved.",
"level": 4,
"order": 15,
"section_id": "this-page-is-a-checked-surface",
"source_line_end": 58,
"source_line_start": 55,
"title": "This page is a checked surface"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "After the validators run secret-scans inputs and copied bodies and passes only on clean findings, then writes three metadata-only records, with the bundle entry requiring the source manifest. After the validators, run runs a secret-exclusion scan over the input paths and copied bodies, merges the observed negative cases, and passes only when there are no blocking findings, no missing negative cases, a clean scan, and verified source modules. buildboard and the result and validation writers then emit voicetodoctrineselfimprovementloopresult.json, ...board.json, and ...validationreceipt.json, each carrying bodyinreceipt: false. runvoicetodoctrine_bundle calls run with negative cases off and t\u2026",
"level": 4,
"order": 16,
"section_id": "secret-scan-gating-and-record-writers",
"source_line_end": 62,
"source_line_start": 59,
"title": "Secret scan, gating, and record writers"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The fixture spans four lessons over five owner surfaces: two refinements, one deferred capture with a re-entry condition, and one verified no-op. The fixture describes four lessons across five owner surfaces: two refine an existing surface, one captures deferred work with a re-entry condition, and one records a verified no-op.",
"level": 4,
"order": 17,
"section_id": "what-the-fixture-describes",
"source_line_end": 66,
"source_line_start": 63,
"title": "What the fixture describes"
},
{
"block_kinds": [
"diagram"
],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "shape",
"source_line_end": 94,
"source_line_start": 67,
"title": "Shape"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "negative-cases",
"source_line_end": 96,
"source_line_start": 95,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "validatenegativecases reads six files, each a shape the loop must refuse, and records a fixed code for each: validatenegativecases reads six files, each a shape the loop must refuse, and records a fixed code for each: rawoperatorvoiceexport carrying a forbidden body key gives VOICEDOCTRINERAWOPERATORBODYFORBIDDEN. doctrinenodehandedit that edits a principle, concept, mechanism, or axiom by direct file write gives VOICEDOCTRINEDIRECTNODEEDITFORBIDDEN. consumewithoutdeposit with no deposit outcome gives VOICEDOCTRINECONSUMEWITHOUTDEPOSIT. patternreceiptonlyprogress whose owner action is append-result record-only gives VOICEDOCTRINERECEIPTONLY_PROGRESS. globalpromotionwithoutownervalidation th\u2026",
"level": 4,
"order": 20,
"section_id": "validate-negative-cases-reads-six-files",
"source_line_end": 108,
"source_line_start": 97,
"title": "`validate_negative_cases` reads six files"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "In fixture mode all six cases must be observed without blocking, in bundle mode none is expected, and source-module tampering each drops the verified-body count and adds a blocking code. In fixture mode these six cases are expected: they must all be observed for the run to pass, and they do not block. In bundle mode the negative inputs are absent and none is expected. The source-module tests also exercise tampering: a changed hash, a rehashed body edit, a wrong source ref, and a missing live source each drop the verified-body count and add a blocking code.",
"level": 4,
"order": 21,
"section_id": "fixture-bundle-and-tampering-modes",
"source_line_end": 112,
"source_line_start": 109,
"title": "Fixture, bundle, and tampering modes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "prior-art-grounding",
"source_line_end": 114,
"source_line_start": 113,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "It follows after-action review and pattern-language practice, but a lesson becomes durable only with evidence, an owner surface, a validation ref, and a bounded re-entry path. This component follows after-action review, lessons-learned practice, and pattern-language work. NASA's Lessons Learned Information System is a public example of preserving operational lessons so later work can reuse them, and pattern-language practice gives a vocabulary for turning repeated local solutions into named reusable forms. This component keeps that direction without collapsing a local lesson into doctrine: a lesson becomes durable only when it has evidence, an owner surface, a validation ref, and a bounded\u2026",
"level": 4,
"order": 23,
"section_id": "after-action-review-and-pattern-language",
"source_line_end": 121,
"source_line_start": 115,
"title": "After-action review and pattern language"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from microcosm-substrate:",
"level": 2,
"order": 24,
"section_id": "validation-receipt-path",
"source_line_end": 139,
"source_line_start": 122,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A green record means the lessons named a real owner surface, resolved every ref, observed the negative cases, and passed the secret scan, and nothing beyond that. A green fixture or bundle record means the lessons named a real owner surface, resolved every ref, observed the negative cases, and passed the secret scan. It means nothing beyond that.",
"level": 4,
"order": 25,
"section_id": "what-a-green-record-means",
"source_line_end": 143,
"source_line_start": 140,
"title": "What a green record means"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 26,
"section_id": "claim-ceiling",
"source_line_end": 145,
"source_line_start": 144,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "On a local folder it can require each lesson to name a surface, resolve every ref, land a typed outcome, and survive six refusals, with the bundle adding an eight-body copy floor. The strongest claim this component supports: on a local folder, it can require that each lesson name an owner surface, resolve every evidence and completion ref to a real file, land one of four typed outcomes, and survive the six refusals, while ignoring any label the lesson asserts about itself. The exported bundle adds a body floor: eight copied source bodies verified by hash, byte count, line count, and anchor, with the raw source text kept out of the result records.",
"level": 4,
"order": 27,
"section_id": "the-strongest-supported-claim",
"source_line_end": 149,
"source_line_start": 146,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It exports no raw voice, thread bodies, or payloads and grants no doctrine mutation, promotion, or launch, keeping the proof boundary at the fixture, bundle, and copied-body manifest. It refuses the rest. It does not export source notes, private thread bodies, or model-output data. It does not grant doctrine mutation, direct edits to doctrine nodes, or global promotion. It does not perform live mutation of tracked work, external model access, source-file changes, launch, or public sharing. A pass is bounded validator evidence, bounded evidence that future lessons will generalize correctly and not whole-system correctness. The proof boundary and scope limit are the fixture, the exported bund\u2026",
"level": 4,
"order": 28,
"section_id": "what-it-refuses",
"source_line_end": 153,
"source_line_start": 150,
"title": "What it refuses"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/voice_to_doctrine_self_improvement_loop.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Voice to Doctrine Self-Improvement Loop validates lesson propagation without exporting source notes or granting doctrine mutation authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "voice_to_doctrine_self_improvement_loop.py",
"path": "src/microcosm_core/organs/voice_to_doctrine_self_improvement_loop.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for lesson propagation validation, owner-surface checks, negative cases, and scope limits.",
"symbols": [
"run",
"run_voice_to_doctrine_bundle",
"validate_projection_protocol",
"validate_policy",
"validate_owner_surfaces",
"validate_lessons",
"validate_negative_cases",
"_build_board"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/voice_to_doctrine_self_improvement_loop.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/voice_to_doctrine_self_improvement_loop.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/voice_to_doctrine_self_improvement_loop.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Voice to Doctrine Self-Improvement Loop validates whether declared lessons refined a named owner surface with validation or were captured with a re-entry condition. It checks projection protocol, policy, owner surfaces, lesson rows, negative cases, source-open body imports, and scope limits while rejecting source notes export, private thread bodies, model-output data, direct doctrine-node edits, unvalidated global promotion, live work log mutation, public sharing, launch, and whole-system correctness.",
"summary_line": "Voice to Doctrine Self-Improvement Loop validates lesson propagation without exporting source notes or granting doctrine mutation authority.",
"title": "Voice to Doctrine Self-Improvement Loop"
},
{
"anchor": "paper-module-routing-anti-patterns-registry",
"authority_ceiling": "Declared public routing anti-pattern registry fixture and copied source body evidence only; no route source authority, no route mutation, no private routing-note export, no external model access, no launch-scope decision, no publishing-scope decision, and no whole-system correctness.",
"body_html": "
It checks a public list of repeated navigation mistakes, each a typed row with a stable id and a plain explanation. The system keeps a public list of navigation mistakes that agents repeat: reaching for grep before asking the kernel for a route, sending work to a bridge before the scope is chosen, or switching execution mode in a chat reply without updating the disk contract. Each mistake is one typed row with a stable id and a plain explanation. This component checks that list.
\n
It confirms the rows hold their shape and the exported copy matches the source byte for byte, emitting counts, ids, and digests without copying the body or editing a route. It reads a copy of the routing registry, confirms the rows hold their declared shape, and confirms that the exported copy still matches the source byte for byte. It emits a result record carrying counts, ids, digests, and verdicts. It never copies the registry body text into that record, and it never edits a route.
\n
Fixture checker, not a live router
\n
It is a checker, not a router, and rejects any row that declares route or source authority even while describing a public anti-pattern.routing_anti_patterns_registry is a checker, not a router. A page that lists routing failures is easy to mistake for the thing that decides routes. This component refuses that role in code. A row may describe a public anti-pattern, but a row that declares route or source authority is rejected.
\n
Purpose
\n
Navigation fails quietly until named
\n
Navigation can fail silently, as when an agent greps a wide tree instead of taking a kernel route, so naming the failures as typed rows lets a reader recognise them. A navigation system can fail quietly. An agent greps a wide tree when one kernel route would have narrowed it first, and nothing complains until the work is already off course. Naming these failures as typed rows lets a reader recognise them instead of rediscovering them.
\n
A loose list is not trustworthy evidence
\n
Duplicate ids, missing explanations, self-graded labels, or claimed route authority would break the list, so each failure mode becomes a named rejection. A loose list is not enough. If rows can carry duplicate ids, skip their explanations, grade themselves with a pre-written label, or quietly claim route authority, the list stops being trustworthy evidence. The component turns each of those failure modes into a named rejection, so a passing run means something specific.
\n
How it works
\n
Fixture and bundle entrypoints over one core
\n
run checks the fixture with its negative cases and run-bundle checks the exported bundle with a source-copy manifest, both calling _build_result in a fixed order. Two entrypoints share one validation core. run checks the first-wave fixture, including its negative cases. run_routing_anti_patterns_bundle (the run-bundle command) checks the exported bundle, where a source-copy manifest is required and the negative cases are not present. Both call _build_result, which runs the checks in a fixed order.
\n
Function
Role
_load_payloads
Loads the positive registry copy and, for the fixture, the six negative-case files.
validate_copied_macro_registry_rows
Checks row shape and derives each row's repair state from its id and text.
_negative_findings
Confirms each declared negative case is rejected with its expected error code.
_source_module_manifest_result
Verifies the exported copy's sha256 digests, anchors, and metadata-only boundary.
scan_paths
Scans the inputs for forbidden private classes and secrets.
_write_receipts and result_card
Emit the result record and the compact card without the copied body.
\n
Row shape, anchors, and forbidden fields
\n
_payload_findings requires the routing_anti_patterns kind, uniquely-id'd rows with text and the three named anchors, and rejects authority overclaims, authority roles, and private fields. The row check is validate_copied_macro_registry_rows, which wraps _payload_findings. _payload_findings requires kind: routing_anti_patterns, a positive integer version, and at least one anti_patterns row. Each row must carry a stable unique id and explanatory text. When the positive fixture is checked, the three named anchors kernel_before_grep, bridge_before_scope, and mode_in_chat_only must all appear. It then walks every nested object and rejects three things: authority overclaims such as release_authorized, provider_calls_authorized, source_mutation_authorized, and route_policy_mutation_authorized; forbidden authority roles, where a row claims source_authority, route_authority, control_plane_authority, or a similar role; and private fields such as private_source_body, raw_seed_body, provider_payload_body, and secret_value.
\n
Repair state derived from text, not labels
\n
_route_repair_state_for_row derives each row's repair state from its own id and text rather than a baked label, keeping baked_expected_labels_sufficient false. One design choice sits in _route_repair_state_for_row. Rather than trust a label written into the row, the checker derives each row's repair state from the row's own id and text. kernel_before_grep only earns kernel_first_navigation if its text actually mentions grep, kernel, and route. A row that ships a pre-written label is recorded, and the field baked_expected_labels_sufficient stays false. The meaning stays grounded in text a reader can see, not in a self-asserted grade.
\n
Source-copy digest and anchor accountability
\n
_source_module_manifest_result streams a sha256 over each copied module, requires all declared digests and anchors to match, and blocks the run on any mismatch. The exported-bundle path adds source-copy accountability through _source_module_manifest_result. It requires source_module_manifest.json, checks that the import is classified copied_non_secret_macro_body, and confirms the manifest keeps body text out of the result record. For each listed module it streams a sha256 over the copied target with _sha256, requires the declared sha256, source_sha256, and target_sha256 to all match that digest, checks that the declared anchors appear in the copied body under source_modules/, and re-runs the row check over that copied body. A digest mismatch or a missing anchor blocks the run.
\n
Secret scan and metadata-only result records
\n
_build_result passes only when rows, negative cases, the secret scan, and the source-copy check all clear, then writes metadata-only records and a compact card._build_result also runs a secret-exclusion scan over the input paths. It reports a pass only when the positive rows have no findings, no expected negative case is missing, the scan has no blocking hit, and the source-copy check is either clean or absent. _write_receipts then writes the result, board, and validation records, and result_card projects a compact card whose omission list names the full-payload keys it left out.
\nDiagram of the mechanism (8 steps).Diagram source & refs
Source refs
shape + derived repair state
validate_copied_macro_registry_rows
flowchart TD Run["run / run-bundle"] Build["_build_result"] Rows["validate_copied_macro_registry_rows shape + derived repair state"] Neg["_negative_findings six named rejections"] Src["_source_module_manifest_result sha256 + anchors"] Scan["scan_paths secret exclusion"] Status["status = pass or blocked"] Out["_write_receipts / result_card metadata-only records"] Run --> Build Build --> Rows Build --> Neg Build --> Src Build --> Scan Rows --> Status Neg --> Status Src --> Status Scan --> Status Status --> Out
\n
Negative cases
\n
The first-wave fixture ships six perturbations. Each must be rejected with its own error code: The first-wave fixture ships six perturbations. Each must be rejected with its own error code:
\n
missing_kind rejected with ROUTING_ANTI_PATTERN_KIND_REQUIRED
duplicate_id rejected with ROUTING_ANTI_PATTERN_DUPLICATE_ID
missing_text rejected with ROUTING_ANTI_PATTERN_TEXT_REQUIRED
authority_overclaim rejected with ROUTING_ANTI_PATTERN_AUTHORITY_OVERCLAIM
source_authority_masquerade rejected with ROUTING_ANTI_PATTERN_SOURCE_AUTHORITY_FORBIDDEN
private_source_leakage rejected with ROUTING_ANTI_PATTERN_PRIVATE_SOURCE_FORBIDDEN
\n
Negative cases as proof obligations
\n
A run passes only when every expected case appears with its code and missing_negative_cases is empty, making the negatives proof obligations rather than illustrations. A run passes only when every expected case appears with its expected code and missing_negative_cases is empty. That makes the negative cases proof obligations, not illustrations.
\n
Prior Art Grounding
\n
From pattern catalogs to CLI usability
\n
The registry follows the pattern and anti-pattern catalog tradition and CLI usability practice, requiring stable ids and explanatory text while staying below route-source authority. This registry follows the pattern and anti-pattern catalog tradition: name recurring failure shapes so future operators can recognise and avoid them. The Hillside patterns library is the positive pattern-language ancestor, and the software anti-pattern literature supplies the inverse move of documenting practices that look useful but produce bad outcomes. The routing-specific presentation borrows from CLI usability practice: the Command Line Interface Guidelines press for discoverability and clear errors, and this component applies that pressure to navigation failures by requiring stable ids and explanatory text while staying below route-source authority.
\n
Validation Result record Path
\n
From microcosm-substrate, check the public routing-registry diagnostic without writing tracked records:
\n
What a pass proves and does not
\n
A pass proves the public fixture, the six negative cases, and the copied-body digest floor, and does not make the registry a route source. A pass proves the public fixture, the six negative cases, and the copied-body digest floor. It does not make the registry a route source.
\n
Scope boundary
\n
Scope limit
\n
The one supported claim
\n
The evidence supports one claim: the rows hold their shape, carry stable ids, keep the three anchors, derive repair state from text, and the exported copy matches the source byte for byte. The evidence here supports one claim: the public anti-pattern rows hold their declared shape, carry stable ids and explanatory text, keep the three named anchors, derive each repair state from row text rather than a baked label, and the exported copy matches the source registry byte for byte with no secret leak. The validation result records that back that claim are the records listed under Validation Result record Path.
\n
Projection-only, never route authority
\n
The boundary stops there: the component decides and mutates no routes, exposes no private notes, and excludes external model access or launch, staying projection-only. The proof boundary stops there. This component is a checker, not route source authority. It does not decide routes, mutate routes, prove route freshness, expose private routing notes, authorize external model access, include launch operations or public sharing, prove maturity or hosted-public posture, or prove whole-system correctness. Its scope limit is projection-only: it can describe how navigation goes wrong without becoming the thing that decides how navigation should go.
\n",
"components": [
{
"anchor": "component-routing_anti_patterns_registry",
"component_id": "routing_anti_patterns_registry",
"public_label": "Routing Anti Patterns Registry"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 9
},
"link_count": 21,
"links": [
{
"id": "routing_anti_patterns_registry",
"kind": "organ",
"kind_label": "component",
"label": "Routing Anti Patterns Registry",
"relation": "explains",
"url": "component-routing-anti-patterns-registry.html"
},
{
"id": "mechanism.routing_anti_patterns_registry.validates_public_routing_anti_patterns_registry",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public routing anti patterns registry",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-routing-anti-patterns-registry-validates-public-routing-anti-patterns-registry"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[2:paper_module.agent_route_observability_runtime].depends_on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[1:paper_module.navigation_hologram_route_plane].depends_on",
"url": "paper-module-navigation-hologram-route-plane.html"
}
],
"paper_module_id": "paper_module.routing_anti_patterns_registry",
"reader_packet": {
"abstract": "Routing Anti-Patterns Registry validates public anti-pattern registry rows without becoming route source authority or mutating routes.",
"authority_ceiling": "Declared public routing anti-pattern registry fixture and copied source body evidence only; no route source authority, no route mutation, no private routing-note export, no external model access, no launch-scope decision, no publishing-scope decision, and no whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.routing_anti_patterns_registry.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-routing-anti-patterns-registry-diagram-1",
"html_url": "docs/paper-module-routing-anti-patterns-registry.html#paper-module-routing-anti-patterns-registry-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Run[\"run / run-bundle\"] Build[\"_build_result\"] Rows[\"validate_copied_macro_registry_rows shape + derived repair state\"] Neg[\"_negative_findings six named rejections\"] Src[\"_source_module_manifest_result sha256 + anchors\"] Scan[\"scan_paths secret exclusion\"] Status[\"status = pass or blocked\"] Out[\"_write_receipts / result_card metadata-only records\"] Run --> Build Build --> Rows Build --> Neg Build --> Src Build --> Scan Rows --> Status Neg --> Status Src --> Status Scan --> Status Status --> Out",
"source_format": "mermaid",
"source_hash": "sha256:1b2d0673d50724dc5242bf2e2834e67c06743806ec8ad41cb5845798da0e2d99",
"source_line_end": 77,
"source_line_start": 56
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "routing_anti_patterns_registry",
"kind": "organ",
"kind_label": "component",
"label": "Routing Anti Patterns Registry",
"relation": "explains",
"url": "component-routing-anti-patterns-registry.html"
},
{
"id": "mechanism.routing_anti_patterns_registry.validates_public_routing_anti_patterns_registry",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public routing anti patterns registry",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-routing-anti-patterns-registry-validates-public-routing-anti-patterns-registry"
}
],
"relation": "explains"
},
{
"count": 10,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.routing_anti_patterns_registry",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 128,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/routing_anti_patterns_registry.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Hillside patterns library",
"section_id": "from-pattern-catalogs-to-cli-usability",
"url": "https://hillside.net/patterns/"
},
{
"kind": "external_reference",
"label": "Command Line Interface Guidelines",
"section_id": "from-pattern-catalogs-to-cli-usability",
"url": "https://clig.dev/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "It checks a public list of repeated navigation mistakes, each a typed row with a stable id and a plain explanation. The system keeps a public list of navigation mistakes that agents repeat: reaching for grep before asking the kernel for a route, sending work to a bridge before the scope is chosen, or switching execution mode in a chat reply without updating the disk contract. Each mistake is one typed row with a stable id and a plain explanation. This component checks that list.",
"level": 4,
"order": 1,
"section_id": "it-checks-a-public-list-of-repeated-navigation-mistakes",
"source_line_end": 6,
"source_line_start": 3,
"title": "It checks a public list of repeated navigation mistakes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It confirms the rows hold their shape and the exported copy matches the source byte for byte, emitting counts, ids, and digests without copying the body or editing a route. It reads a copy of the routing registry, confirms the rows hold their declared shape, and confirms that the exported copy still matches the source byte for byte. It emits a result record carrying counts, ids, digests, and verdicts. It never copies the registry body text into that record, and it never edits a route.",
"level": 4,
"order": 2,
"section_id": "it-confirms-the-rows-hold-their-shape",
"source_line_end": 10,
"source_line_start": 7,
"title": "It confirms the rows hold their shape"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It is a checker, not a router, and rejects any row that declares route or source authority even while describing a public anti-pattern. routingantipatterns_registry is a checker, not a router. A page that lists routing failures is easy to mistake for the thing that decides routes. This component refuses that role in code. A row may describe a public anti-pattern, but a row that declares route or source authority is rejected.",
"level": 4,
"order": 3,
"section_id": "fixture-checker-not-a-live-router",
"source_line_end": 14,
"source_line_start": 11,
"title": "Fixture checker, not a live router"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Navigation can fail silently, as when an agent greps a wide tree instead of taking a kernel route, so naming the failures as typed rows lets a reader recognise them. A navigation system can fail quietly. An agent greps a wide tree when one kernel route would have narrowed it first, and nothing complains until the work is already off course. Naming these failures as typed rows lets a reader recognise them instead of rediscovering them.",
"level": 4,
"order": 5,
"section_id": "navigation-fails-quietly-until-named",
"source_line_end": 20,
"source_line_start": 17,
"title": "Navigation fails quietly until named"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Duplicate ids, missing explanations, self-graded labels, or claimed route authority would break the list, so each failure mode becomes a named rejection. A loose list is not enough. If rows can carry duplicate ids, skip their explanations, grade themselves with a pre-written label, or quietly claim route authority, the list stops being trustworthy evidence. The component turns each of those failure modes into a named rejection, so a passing run means something specific.",
"level": 4,
"order": 6,
"section_id": "a-loose-list-is-not-trustworthy-evidence",
"source_line_end": 24,
"source_line_start": 21,
"title": "A loose list is not trustworthy evidence"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "run checks the fixture with its negative cases and run-bundle checks the exported bundle with a source-copy manifest, both calling buildresult in a fixed order. Two entrypoints share one validation core. run checks the first-wave fixture, including its negative cases. runroutingantipatternsbundle (the run-bundle command) checks the exported bundle, where a source-copy manifest is required and the negative cases are not present. Both call buildresult, which runs the checks in a fixed order.",
"level": 4,
"order": 8,
"section_id": "fixture-and-bundle-entrypoints-over-one-core",
"source_line_end": 39,
"source_line_start": 27,
"title": "Fixture and bundle entrypoints over one core"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "payloadfindings requires the routingantipatterns kind, uniquely-id'd rows with text and the three named anchors, and rejects authority overclaims, authority roles, and private fields. The row check is validatecopiedmacroregistryrows, which wraps payloadfindings. payloadfindings requires kind: routingantipatterns, a positive integer version, and at least one antipatterns row. Each row must carry a stable unique id and explanatory text. When the positive fixture is checked, the three named anchors kernelbeforegrep, bridgebeforescope, and modeinchatonly must all appear. It then walks every nested object and rejects three things: authority overclaims such as releaseauthorized, providercallsauth\u2026",
"level": 4,
"order": 9,
"section_id": "row-shape-anchors-and-forbidden-fields",
"source_line_end": 43,
"source_line_start": 40,
"title": "Row shape, anchors, and forbidden fields"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "routerepairstateforrow derives each row's repair state from its own id and text rather than a baked label, keeping bakedexpectedlabelssufficient false. One design choice sits in routerepairstateforrow. Rather than trust a label written into the row, the checker derives each row's repair state from the row's own id and text. kernelbeforegrep only earns kernelfirstnavigation if its text actually mentions grep, kernel, and route. A row that ships a pre-written label is recorded, and the field bakedexpectedlabels_sufficient stays false. The meaning stays grounded in text a reader can see, not in a self-asserted grade.",
"level": 4,
"order": 10,
"section_id": "repair-state-derived-from-text-not-labels",
"source_line_end": 47,
"source_line_start": 44,
"title": "Repair state derived from text, not labels"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourcemodulemanifestresult streams a sha256 over each copied module, requires all declared digests and anchors to match, and blocks the run on any mismatch. The exported-bundle path adds source-copy accountability through sourcemodulemanifestresult. It requires sourcemodulemanifest.json, checks that the import is classified copiednonsecretmacrobody, and confirms the manifest keeps body text out of the result record. For each listed module it streams a sha256 over the copied target with sha256, requires the declared sha256, sourcesha256, and targetsha256 to all match that digest, checks that the declared anchors appear in the copied body under source_modules/, and re-runs the row check over\u2026",
"level": 4,
"order": 11,
"section_id": "source-copy-digest-and-anchor-accountability",
"source_line_end": 51,
"source_line_start": 48,
"title": "Source-copy digest and anchor accountability"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "buildresult passes only when rows, negative cases, the secret scan, and the source-copy check all clear, then writes metadata-only records and a compact card. buildresult also runs a secret-exclusion scan over the input paths. It reports a pass only when the positive rows have no findings, no expected negative case is missing, the scan has no blocking hit, and the source-copy check is either clean or absent. writereceipts then writes the result, board, and validation records, and resultcard projects a compact card whose omission list names the full-payload keys it left out.",
"level": 4,
"order": 12,
"section_id": "secret-scan-and-body-free-receipts",
"source_line_end": 78,
"source_line_start": 52,
"title": "Secret scan and metadata-only result records"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 80,
"source_line_start": 79,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The first-wave fixture ships six perturbations. Each must be rejected with its own error code: The first-wave fixture ships six perturbations. Each must be rejected with its own error code: missingkind rejected with ROUTINGANTIPATTERNKIND_REQUIRED duplicateid rejected with ROUTINGANTIPATTERNDUPLICATE_ID missingtext rejected with ROUTINGANTIPATTERNTEXT_REQUIRED authorityoverclaim rejected with ROUTINGANTIPATTERNAUTHORITY_OVERCLAIM sourceauthoritymasquerade rejected with ROUTINGANTIPATTERNSOURCEAUTHORITY_FORBIDDEN privatesourceleakage rejected with ROUTINGANTIPATTERNPRIVATESOURCE_FORBIDDEN",
"level": 4,
"order": 14,
"section_id": "the-first-wave-fixture-ships-six-perturbations",
"source_line_end": 92,
"source_line_start": 81,
"title": "The first-wave fixture ships six perturbations"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A run passes only when every expected case appears with its code and missingnegativecases is empty, making the negatives proof obligations rather than illustrations. A run passes only when every expected case appears with its expected code and missingnegativecases is empty. That makes the negative cases proof obligations, not illustrations.",
"level": 4,
"order": 15,
"section_id": "negative-cases-as-proof-obligations",
"source_line_end": 96,
"source_line_start": 93,
"title": "Negative cases as proof obligations"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 98,
"source_line_start": 97,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The registry follows the pattern and anti-pattern catalog tradition and CLI usability practice, requiring stable ids and explanatory text while staying below route-source authority. This registry follows the pattern and anti-pattern catalog tradition: name recurring failure shapes so future operators can recognise and avoid them. The Hillside patterns library is the positive pattern-language ancestor, and the software anti-pattern literature supplies the inverse move of documenting practices that look useful but produce bad outcomes. The routing-specific presentation borrows from CLI usability practice: the Command Line Interface Guidelines press for discoverability and clear errors, and th\u2026",
"level": 4,
"order": 17,
"section_id": "from-pattern-catalogs-to-cli-usability",
"source_line_end": 102,
"source_line_start": 99,
"title": "From pattern catalogs to CLI usability"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate, check the public routing-registry diagnostic without writing tracked records:",
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 113,
"source_line_start": 103,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass proves the public fixture, the six negative cases, and the copied-body digest floor, and does not make the registry a route source. A pass proves the public fixture, the six negative cases, and the copied-body digest floor. It does not make the registry a route source.",
"level": 4,
"order": 19,
"section_id": "what-a-pass-proves-and-does-not",
"source_line_end": 117,
"source_line_start": 114,
"title": "What a pass proves and does not"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 119,
"source_line_start": 118,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The evidence supports one claim: the rows hold their shape, carry stable ids, keep the three anchors, derive repair state from text, and the exported copy matches the source byte for byte. The evidence here supports one claim: the public anti-pattern rows hold their declared shape, carry stable ids and explanatory text, keep the three named anchors, derive each repair state from row text rather than a baked label, and the exported copy matches the source registry byte for byte with no secret leak. The validation result records that back that claim are the records listed under Validation Result record Path.",
"level": 4,
"order": 21,
"section_id": "the-one-supported-claim",
"source_line_end": 123,
"source_line_start": 120,
"title": "The one supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The boundary stops there: the component decides and mutates no routes, exposes no private notes, and excludes external model access or launch, staying projection-only. The proof boundary stops there. This component is a checker, not route source authority. It does not decide routes, mutate routes, prove route freshness, expose private routing notes, authorize external model access, include launch operations or public sharing, prove maturity or hosted-public posture, or prove whole-system correctness. Its scope limit is projection-only: it can describe how navigation goes wrong without becoming the thing that decides how navigation should go.",
"level": 4,
"order": 22,
"section_id": "projection-only-never-route-authority",
"source_line_end": 127,
"source_line_start": 124,
"title": "Projection-only, never route authority"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/routing_anti_patterns_registry.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Routing Anti-Patterns Registry validates public anti-pattern registry rows without becoming route source authority or mutating routes."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "routing_anti_patterns_registry.py",
"path": "src/microcosm_core/organs/routing_anti_patterns_registry.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for routing anti-pattern registry validation, source-module digest checks, private-leak rejection, negative cases, and scope limits.",
"symbols": [
"run",
"run_routing_anti_patterns_bundle",
"_payload_findings",
"_source_module_manifest_result",
"_source_open_body_import_summary",
"_build_result",
"_write_receipts",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/routing_anti_patterns_registry.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/routing_anti_patterns_registry.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/routing_anti_patterns_registry.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Routing Anti-Patterns Registry validates the public routing anti-pattern registry contract: kind/version, unique row ids, text fields, required navigation anchors, source-module digest evidence, negative cases, and scope limits. It treats the registry as a checked public artifact, not route source authority, route mutation authority, private routing-note export, provider authority, launch, public sharing, or whole-system correctness.",
"summary_line": "Routing Anti-Patterns Registry validates public anti-pattern registry rows without becoming route source authority or mutating routes.",
"title": "Routing Anti-Patterns Registry"
},
{
"anchor": "paper-module-batch8-audio-level-rms-port",
"authority_ceiling": "Deterministic RMS parity evidence over public fixture inputs and copied source refs only; no macOS audio-session evidence, microphone permission authority, device capture, recorded audio, UI readiness, source-file changes, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
The Swift AudioLevelMonitor drives a live microphone level meter in a recording app. Almost all of that file is platform machinery: opening a capture session, picking a device, reading sample buffers off a callback. The Swift AudioLevelMonitor drives a live microphone level meter in a recording app. Almost all of that file is platform machinery: opening a capture session, picking a device, reading sample buffers off a callback. One small pure function inside it, normalizedLevel, turns a block of audio samples into a single number between zero and one. That number is the meter reading, and it is the only part that can be checked without a microphone.
\n
This component ports that one function to Python and runs it over synthetic sample arrays that ship in the public clone. This component ports that one function to Python and runs it over synthetic sample arrays that ship in the public clone. It answers one question: does the Python re-implementation produce the same level value as the Swift original, on inputs anyone can inspect. It writes a result record with the per-case comparison and a pass or blocked verdict.
\n
Everything device-specific stays on the Swift side. The port does not open an AVCaptureSession, request microphone permission, read recorded audio, or capture a device. Everything device-specific stays on the Swift side. The port does not open an AVCaptureSession, request microphone permission, read recorded audio, or capture a device.
\n
Purpose
\n
A live level meter is hard to test because it depends on real audio hardware and OS permissions that cannot live in a public fixture. A live level meter is hard to test because it depends on real audio hardware and OS permissions that cannot live in a public fixture. If you ported the whole monitor you would inherit all of that and prove nothing checkable. The narrow choice here is to isolate the amplitude arithmetic and hold everything else out.
\n
What crosses into Python is the calculation alone. The claim is deliberately small: numeric parity for one function over sample arrays we can publish. What crosses into Python is the calculation alone. The claim is deliberately small: numeric parity for one function over sample arrays we can publish. That is why the interesting content is what the port refuses to include, not what it copies.
\n
How it works
\n
The math lives in normalized_level(samples, sample_format). It accepts two format tags, float32 and int16. The math lives in normalized_level(samples, sample_format). It accepts two format tags, float32 and int16. Any other tag raises ValueError, which is how the unsupported-format case is exercised. An empty buffer returns 0.0 before any arithmetic runs.
\n
For a non-empty buffer it accumulates the square of each sample. Float samples are used as-is. For a non-empty buffer it accumulates the square of each sample. Float samples are used as-is. Int16 samples are first divided by 32767.0, the Swift Int16.max, to map the integer range onto roughly minus-one to one. It then takes the root mean square, sqrt(total / count), which summarises the block's energy as one amplitude. That value is multiplied by 8.0 and clamped with min(max(rms * 8.0, 0.0), 1.0). The gain of eight is a display choice carried over verbatim from the Swift source: quiet speech sits low on a zero-to-one meter without it, so the level is scaled up and then capped so loud input cannot overshoot one. The int16 divisor and the rms * 8 clamp are the two lines the exported bundle requires to match the copied Swift text.
\n
The runtime is _audio_evaluator. It loads the probe manifest and runs three evaluation passes, then joins their findings. The runtime is _audio_evaluator. It loads the probe manifest and runs three evaluation passes, then joins their findings.
\n
_evaluate_reference_cases reads the synthetic cases in the manifest. Each case carries samples, a format, an expected level, and a tolerance._evaluate_reference_cases reads the synthetic cases in the manifest. Each case carries samples, a format, an expected level, and a tolerance. For each one it calls normalized_level and compares the observed value against the expected value within the tolerance. A case that is malformed, that raises, or that lands outside tolerance becomes a finding. It also checks that all three named cases in EXPECTED_CASES (float32_reference_buffer, int16_reference_buffer, clamp_over_one_buffer) are present, and records a finding for any that are missing.
\n
_evaluate_byte_reference_cases covers the optional WAV cases. For each one _decode_int16_wav opens the file, and rejects anything that is not mono, 16-bit, uncompressed PCM, or whose byte length does not match its header._evaluate_byte_reference_cases covers the optional WAV cases. For each one _decode_int16_wav opens the file, and rejects anything that is not mono, 16-bit, uncompressed PCM, or whose byte length does not match its header. It unpacks the frames to int16 samples, recomputes the level from the raw bytes, and compares against the manifest expectation. A missing file, a decode failure, or an unsupported format each becomes a distinct finding rather than a silent skip.
\n
_evaluate_negative_exercises runs the three refusals directly. It confirms an empty buffer reads zero, that [1.0, -1.0, 0.5] scales past one and clamps to one, and that format pcm24 is refused with a ValueError._evaluate_negative_exercises runs the three refusals directly. It confirms an empty buffer reads zero, that [1.0, -1.0, 0.5] scales past one and clamps to one, and that format pcm24 is refused with a ValueError. Each expectation is cross-checked against a small JSON fixture and turns into a finding if it does not hold. The declared codes for these cases live in EXPECTED_NEGATIVE_CASES.
\n
Any finding from any pass flips the overall status from pass to blocked. run wraps the evaluator for fixture input and run_batch8_audio_level_rms_bundle wraps it for the exported source bundle; result_card projects the verdict, the reference-case counts, and the authority floor into a compact public card. Any finding from any pass flips the overall status from pass to blocked. run wraps the evaluator for fixture input and run_batch8_audio_level_rms_bundle wraps it for the exported source bundle; result_card projects the verdict, the reference-case counts, and the authority floor into a compact public card.
\nDiagram of the mechanism (10 steps).Diagram source & refs
The three named negative cases are the reason the port is worth running. audio_level_empty_buffer_zero requires an empty buffer to read 0.0. The three named negative cases are the reason the port is worth running. audio_level_empty_buffer_zero requires an empty buffer to read 0.0. audio_level_clamps_over_one requires an over-one RMS-scaled buffer to clamp to 1.0. audio_level_unknown_format_refused requires a format the port does not know to raise ValueError. If any of these behaviors changed, the corresponding code (BATCH8_AUDIO_LEVEL_EMPTY_BUFFER_ZERO, BATCH8_AUDIO_LEVEL_CLAMP_REQUIRED, BATCH8_AUDIO_LEVEL_UNKNOWN_FORMAT_REFUSED) is recorded and the run blocks.
\n
Prior Art Grounding
\n
The math is standard digital-audio metering. Root mean square amplitude is a common way to summarise signal energy for a level display, and the practice of keeping OS capture APIs outside a pure numeric test is ordinary engineering hygiene. The math is standard digital-audio metering. Root mean square amplitude is a common way to summarise signal energy for a level display, and the practice of keeping OS capture APIs outside a pure numeric test is ordinary engineering hygiene. Two honest anchors for the surrounding platform work are Apple's AVFoundation media framework, which is the capture layer the Swift original uses, and the FFmpeg documentation, where audio streams and levels are handled as explicit inputs and transforms. This component borrows only the RMS-level calculation shape and binds it to fixture-based Python parity checks.
\n
Validation Result record Path
\n
Run these from the microcosm-substrate/ public root:
\n
The first command writes the parity result and sign-off JSON. The second validates the copied Swift source module, its digest anchors, the negative exercises, and the body-exclusion scan. The first command writes the parity result and sign-off JSON. The second validates the copied Swift source module, its digest anchors, the negative exercises, and the body-exclusion scan. The test checks the Python port, bundle validation, result record body scan, and scope limit. A pass means the Python port reproduced every expected level within tolerance and refused every case it should refuse.
\n
Scope boundary
\n
Scope limit
\n
Numeric parity with the copied Swift calculation
\n
The strongest claim this supports is a deterministic Python port of the audio-level RMS calculation that matches the copied Swift text on published sample arrays, with a generated diagram view and navigation links available from the same source row. The strongest claim this supports is a deterministic Python port of the audio-level RMS calculation that matches the copied Swift text on published sample arrays, with a generated diagram view and navigation links available from the same source row. The proof boundary is numeric parity for normalized_level over public fixture inputs, nothing more.
\n
The scope limit stops there. This is not macOS audio-session evidence, not microphone permission authority, not device capture, not UI readiness, not source-file changes, and not public sharing or launch-scope decision. The scope limit stops there. This is not macOS audio-session evidence, not microphone permission authority, not device capture, not UI readiness, not source-file changes, and not public sharing or launch-scope decision. The result records carry refs, digests, sample counts, and parity verdicts only; copied source bodies and audio samples stay out of them. Any claim past parity would need new evidence this component does not produce.
\n",
"components": [
{
"anchor": "component-batch8_audio_level_rms_port",
"component_id": "batch8_audio_level_rms_port",
"public_label": "Audio Level RMS Port"
}
],
"doctrine_feed": {
"counts": {
"axiom": 5,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 7
},
"link_count": 18,
"links": [
{
"id": "batch8_audio_level_rms_port",
"kind": "organ",
"kind_label": "component",
"label": "Audio Level RMS Port",
"relation": "explains",
"url": "component-batch8-audio-level-rms-port.html"
},
{
"id": "mechanism.batch8_audio_level_rms_port.validates_public_audio_level_rms_port",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public audio level rms port",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch8-audio-level-rms-port-validates-public-audio-level-rms-port"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.batch7_demo_take_console_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Demo Take Console Bundle",
"relation": "depends on",
"url": "paper-module-batch7-demo-take-console-capsule.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.public_reveal_walkthrough",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Public Reveal Walkthrough",
"relation": "depends on",
"url": "paper-module-public-reveal-walkthrough.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.batch8_audio_level_rms_port",
"reader_packet": {
"abstract": "Set 8 Audio Level RMS Port validates deterministic RMS math parity over public synthetic samples without audio capture, microphone permission, or UI readiness authority.",
"authority_ceiling": "Deterministic RMS parity evidence over public fixture inputs and copied source refs only; no macOS audio-session evidence, microphone permission authority, device capture, recorded audio, UI readiness, source-file changes, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch8_audio_level_rms_port.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch8-audio-level-rms-port-diagram-1",
"html_url": "docs/paper-module-batch8-audio-level-rms-port.html#paper-module-batch8-audio-level-rms-port-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD manifest[\"Probe manifest synthetic arrays + WAV byte cases expected level per case\"] evaluator[\"_audio_evaluator\"] refcases[\"_evaluate_reference_cases float32, int16, clamp-over-one\"] bytecases[\"_evaluate_byte_reference_cases decode mono 16-bit PCM WAV\"] negatives[\"_evaluate_negative_exercises empty, clamp, unknown-format\"] level[\"normalized_level(samples, format)\"] compare[\"compare observed vs expected within tolerance\"] verdict{\"any finding?\"} blocked[\"status: blocked\"] passed[\"status: pass\"] manifest --> evaluator evaluator --> refcases evaluator --> bytecases evaluator --> negatives refcases --> level bytecases --> level negatives --> level level --> compare compare --> verdict verdict -->|\"yes\"| blocked verdict -->|\"no\"| passed",
"source_format": "mermaid",
"source_hash": "sha256:25606d1a333cb2f7cfd7134eb689a0a5b67ae7aae9037c467fe3a6a2f7af5ba0",
"source_line_end": 91,
"source_line_start": 67
}
],
"grouped_relation_feeds": [
{
"count": 5,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.batch7_demo_take_console_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Demo Take Console Bundle",
"relation": "depends on",
"url": "paper-module-batch7-demo-take-console-capsule.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.public_reveal_walkthrough",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Public Reveal Walkthrough",
"relation": "depends on",
"url": "paper-module-public-reveal-walkthrough.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch8_audio_level_rms_port",
"kind": "organ",
"kind_label": "component",
"label": "Audio Level RMS Port",
"relation": "explains",
"url": "component-batch8-audio-level-rms-port.html"
},
{
"id": "mechanism.batch8_audio_level_rms_port.validates_public_audio_level_rms_port",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public audio level rms port",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch8-audio-level-rms-port-validates-public-audio-level-rms-port"
}
],
"relation": "explains"
},
{
"count": 8,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch8_audio_level_rms_port",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 141,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_audio_level_rms_port.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "AVFoundation",
"section_id": "the-math-is-standard-digital-audio-metering",
"url": "https://developer.apple.com/av-foundation/"
},
{
"kind": "external_reference",
"label": "FFmpeg documentation",
"section_id": "the-math-is-standard-digital-audio-metering",
"url": "https://www.ffmpeg.org/documentation.html"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "The Swift AudioLevelMonitor drives a live microphone level meter in a recording app. Almost all of that file is platform machinery: opening a capture session, picking a device, reading sample buffers off a callback. The Swift AudioLevelMonitor drives a live microphone level meter in a recording app. Almost all of that file is platform machinery: opening a capture session, picking a device, reading sample buffers off a callback. One small pure function inside it, normalizedLevel, turns a block of audio samples into a single number between zero and one. That number is the meter reading, and it is the only part that can be checked without a microphone.",
"level": 4,
"order": 1,
"section_id": "the-swift-audiolevelmonitor-drives-a-live-microphone-level-meter-in-a-recording-app",
"source_line_end": 7,
"source_line_start": 3,
"title": "The Swift `AudioLevelMonitor` drives a live microphone level meter in a recording app"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "This component ports that one function to Python and runs it over synthetic sample arrays that ship in the public clone. This component ports that one function to Python and runs it over synthetic sample arrays that ship in the public clone. It answers one question: does the Python re-implementation produce the same level value as the Swift original, on inputs anyone can inspect. It writes a result record with the per-case comparison and a pass or blocked verdict.",
"level": 4,
"order": 2,
"section_id": "this-component-ports-that-one-function-to-python",
"source_line_end": 12,
"source_line_start": 8,
"title": "This component ports that one function to Python"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Everything device-specific stays on the Swift side. The port does not open an AVCaptureSession, request microphone permission, read recorded audio, or capture a device. Everything device-specific stays on the Swift side. The port does not open an AVCaptureSession, request microphone permission, read recorded audio, or capture a device.",
"level": 4,
"order": 3,
"section_id": "everything-device-specific-stays-on-the-swift-side",
"source_line_end": 17,
"source_line_start": 13,
"title": "Everything device-specific stays on the Swift side"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 19,
"source_line_start": 18,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A live level meter is hard to test because it depends on real audio hardware and OS permissions that cannot live in a public fixture. A live level meter is hard to test because it depends on real audio hardware and OS permissions that cannot live in a public fixture. If you ported the whole monitor you would inherit all of that and prove nothing checkable. The narrow choice here is to isolate the amplitude arithmetic and hold everything else out.",
"level": 4,
"order": 5,
"section_id": "a-live-level-meter-is-hard-to-test",
"source_line_end": 24,
"source_line_start": 20,
"title": "A live level meter is hard to test"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "What crosses into Python is the calculation alone. The claim is deliberately small: numeric parity for one function over sample arrays we can publish. What crosses into Python is the calculation alone. The claim is deliberately small: numeric parity for one function over sample arrays we can publish. That is why the interesting content is what the port refuses to include, not what it copies.",
"level": 4,
"order": 6,
"section_id": "what-crosses-into-python-is-the-calculation-alone",
"source_line_end": 29,
"source_line_start": 25,
"title": "What crosses into Python is the calculation alone"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 31,
"source_line_start": 30,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The math lives in normalizedlevel(samples, sampleformat). It accepts two format tags, float32 and int16. The math lives in normalizedlevel(samples, sampleformat). It accepts two format tags, float32 and int16. Any other tag raises ValueError, which is how the unsupported-format case is exercised. An empty buffer returns 0.0 before any arithmetic runs.",
"level": 4,
"order": 8,
"section_id": "the-math-lives-in-normalized-level-samples-sample-format",
"source_line_end": 36,
"source_line_start": 32,
"title": "The math lives in `normalized_level(samples, sample_format)`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "For a non-empty buffer it accumulates the square of each sample. Float samples are used as-is. For a non-empty buffer it accumulates the square of each sample. Float samples are used as-is. Int16 samples are first divided by 32767.0, the Swift Int16.max, to map the integer range onto roughly minus-one to one. It then takes the root mean square, sqrt(total / count), which summarises the block's energy as one amplitude. That value is multiplied by 8.0 and clamped with min(max(rms 8.0, 0.0), 1.0). The gain of eight is a display choice carried over verbatim from the Swift source: quiet speech sits low on a zero-to-one meter without it, so the level is scaled up and then capped so loud input can\u2026",
"level": 4,
"order": 9,
"section_id": "for-a-non-empty-buffer-it-accumulates-the-square-of-each-sample",
"source_line_end": 41,
"source_line_start": 37,
"title": "For a non-empty buffer it accumulates the square of each sample"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The runtime is audioevaluator. It loads the probe manifest and runs three evaluation passes, then joins their findings. The runtime is audioevaluator. It loads the probe manifest and runs three evaluation passes, then joins their findings.",
"level": 4,
"order": 10,
"section_id": "the-runtime-is-audio-evaluator",
"source_line_end": 46,
"source_line_start": 42,
"title": "The runtime is `_audio_evaluator`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluatereferencecases reads the synthetic cases in the manifest. Each case carries samples, a format, an expected level, and a tolerance. evaluatereferencecases reads the synthetic cases in the manifest. Each case carries samples, a format, an expected level, and a tolerance. For each one it calls normalizedlevel and compares the observed value against the expected value within the tolerance. A case that is malformed, that raises, or that lands outside tolerance becomes a finding. It also checks that all three named cases in EXPECTEDCASES (float32referencebuffer, int16referencebuffer, clampoveronebuffer) are present, and records a finding for any that are missing.",
"level": 4,
"order": 11,
"section_id": "evaluate-reference-cases-reads-the-synthetic-cases-in-the-manifest",
"source_line_end": 51,
"source_line_start": 47,
"title": "`_evaluate_reference_cases` reads the synthetic cases in the manifest"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluatebytereferencecases covers the optional WAV cases. For each one decodeint16wav opens the file, and rejects anything that is not mono, 16-bit, uncompressed PCM, or whose byte length does not match its header. evaluatebytereferencecases covers the optional WAV cases. For each one decodeint16_wav opens the file, and rejects anything that is not mono, 16-bit, uncompressed PCM, or whose byte length does not match its header. It unpacks the frames to int16 samples, recomputes the level from the raw bytes, and compares against the manifest expectation. A missing file, a decode failure, or an unsupported format each becomes a distinct finding rather than a silent skip.",
"level": 4,
"order": 12,
"section_id": "evaluate-byte-reference-cases-covers-the-optional-wav-cases",
"source_line_end": 56,
"source_line_start": 52,
"title": "`_evaluate_byte_reference_cases` covers the optional WAV cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluatenegativeexercises runs the three refusals directly. It confirms an empty buffer reads zero, that [1.0, -1.0, 0.5] scales past one and clamps to one, and that format pcm24 is refused with a ValueError. evaluatenegativeexercises runs the three refusals directly. It confirms an empty buffer reads zero, that [1.0, -1.0, 0.5] scales past one and clamps to one, and that format pcm24 is refused with a ValueError. Each expectation is cross-checked against a small JSON fixture and turns into a finding if it does not hold. The declared codes for these cases live in EXPECTEDNEGATIVE_CASES.",
"level": 4,
"order": 13,
"section_id": "evaluate-negative-exercises-runs-the-three-refusals-directly",
"source_line_end": 61,
"source_line_start": 57,
"title": "`_evaluate_negative_exercises` runs the three refusals directly"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "Any finding from any pass flips the overall status from pass to blocked. run wraps the evaluator for fixture input and runbatch8audiolevelrmsbundle wraps it for the exported source bundle; resultcard projects the verdict, the reference-case counts, and the authority floor into a compact public card. Any finding from any pass flips the overall status from pass to blocked. run wraps the evaluator for fixture input and runbatch8audiolevelrmsbundle wraps it for the exported source bundle; resultcard projects the verdict, the reference-case counts, and the authority floor into a compact public card.",
"level": 4,
"order": 14,
"section_id": "any-finding-from-any-pass-flips-the-overall-status-from-pass-to-blocked",
"source_line_end": 92,
"source_line_start": 62,
"title": "Any finding from any pass flips the overall status from `pass` to `blocked`"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "negative-cases",
"source_line_end": 94,
"source_line_start": 93,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The three named negative cases are the reason the port is worth running. audiolevelemptybufferzero requires an empty buffer to read 0.0. The three named negative cases are the reason the port is worth running. audiolevelemptybufferzero requires an empty buffer to read 0.0. audiolevelclampsoverone requires an over-one RMS-scaled buffer to clamp to 1.0. audiolevelunknownformatrefused requires a format the port does not know to raise ValueError. If any of these behaviors changed, the corresponding code (BATCH8AUDIOLEVELEMPTYBUFFERZERO, BATCH8AUDIOLEVELCLAMPREQUIRED, BATCH8AUDIOLEVELUNKNOWNFORMATREFUSED) is recorded and the run blocks.",
"level": 4,
"order": 16,
"section_id": "the-three-named-negative-cases-are-the-reason-the-port-is-worth-running",
"source_line_end": 99,
"source_line_start": 95,
"title": "The three named negative cases are the reason the port is worth running"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 101,
"source_line_start": 100,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The math is standard digital-audio metering. Root mean square amplitude is a common way to summarise signal energy for a level display, and the practice of keeping OS capture APIs outside a pure numeric test is ordinary engineering hygiene. The math is standard digital-audio metering. Root mean square amplitude is a common way to summarise signal energy for a level display, and the practice of keeping OS capture APIs outside a pure numeric test is ordinary engineering hygiene. Two honest anchors for the surrounding platform work are Apple's AVFoundation media framework, which is the capture layer the Swift original uses, and the FFmpeg documentation, where audio streams and levels are handl\u2026",
"level": 4,
"order": 18,
"section_id": "the-math-is-standard-digital-audio-metering",
"source_line_end": 106,
"source_line_start": 102,
"title": "The math is standard digital-audio metering"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run these from the microcosm-substrate/ public root:",
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 123,
"source_line_start": 107,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The first command writes the parity result and sign-off JSON. The second validates the copied Swift source module, its digest anchors, the negative exercises, and the body-exclusion scan. The first command writes the parity result and sign-off JSON. The second validates the copied Swift source module, its digest anchors, the negative exercises, and the body-exclusion scan. The test checks the Python port, bundle validation, result record body scan, and scope limit. A pass means the Python port reproduced every expected level within tolerance and refused every case it should refuse.",
"level": 4,
"order": 20,
"section_id": "the-first-command-writes-the-parity-result-and-acceptance-json",
"source_line_end": 128,
"source_line_start": 124,
"title": "The first command writes the parity result and sign-off JSON"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 130,
"source_line_start": 129,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim this supports is a deterministic Python port of the audio-level RMS calculation that matches the copied Swift text on published sample arrays, with a generated diagram view and navigation links available from the same source row. The strongest claim this supports is a deterministic Python port of the audio-level RMS calculation that matches the copied Swift text on published sample arrays, with a generated diagram view and navigation links available from the same source row. The proof boundary is numeric parity for normalized_level over public fixture inputs, nothing more.",
"level": 4,
"order": 22,
"section_id": "numeric-parity-with-the-copied-swift-calculation",
"source_line_end": 135,
"source_line_start": 131,
"title": "Numeric parity with the copied Swift calculation"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The scope limit stops there. This is not macOS audio-session evidence, not microphone permission authority, not device capture, not UI readiness, not source-file changes, and not public sharing or launch-scope decision. The scope limit stops there. This is not macOS audio-session evidence, not microphone permission authority, not device capture, not UI readiness, not source-file changes, and not public sharing or launch-scope decision. The result records carry refs, digests, sample counts, and parity verdicts only; copied source bodies and audio samples stay out of them. Any claim past parity would need new evidence this component does not produce.",
"level": 4,
"order": 23,
"section_id": "the-authority-ceiling-stops-there",
"source_line_end": 140,
"source_line_start": 136,
"title": "The scope limit stops there"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch8_audio_level_rms_port.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 8 Audio Level RMS Port validates deterministic RMS math parity over public synthetic samples without audio capture, microphone permission, or UI readiness authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch8_audio_level_rms_port.py",
"path": "src/microcosm_core/organs/batch8_audio_level_rms_port.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for deterministic RMS parity checks, public sample fixtures, negative exercises, and scope limits.",
"symbols": [
"normalized_level",
"_evaluate_reference_cases",
"_evaluate_negative_exercises",
"_audio_evaluator",
"run",
"run_batch8_audio_level_rms_bundle",
"result_card",
"EXPECTED_NEGATIVE_CASES"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch8_audio_level_rms_port.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_audio_level_rms_port.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_audio_level_rms_port.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 8 Audio Level RMS Port ports the Swift AudioLevelMonitor normalizedLevel RMS calculation into a bounded Python component and checks float, int16, clamp, empty-buffer, and unsupported-format cases over public synthetic sample arrays. It carries source-module refs, digests, anchors, sample counts, parity verdicts, negative cases, and an scope limit while excluding AVCaptureSession startup, microphone permission, recorded audio, device state, UI readiness, source-file changes, launch, public sharing, and whole-system correctness.",
"summary_line": "Set 8 Audio Level RMS Port validates deterministic RMS math parity over public synthetic samples without audio capture, microphone permission, or UI readiness authority.",
"title": "Set 8 Audio Level RMS Port"
},
{
"anchor": "paper-module-batch8-compliance-pipeline-capsule",
"authority_ceiling": "Declared public compliance and pipeline fixture evidence plus copied source-module digest checks only; no full compliance-ledger refresh, bridge/external model access authority, source note mutation, repository mutation, complete compliance proof, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
batch8_compliance_pipeline_capsule takes two internal subsystems, copies their source bodies into the public clone, and runs them on synthetic inputs.batch8_compliance_pipeline_capsule takes two internal subsystems, copies their source bodies into the public clone, and runs them on synthetic inputs. The first subsystem is a compliance scanner registry and its bounded ledger builder. The second is a six-stage pipeline that turns a long block of raw text into a short observe plan. The component runs six checks over the copied bodies and writes result records that carry no source text.
\n
Most bundle components in this group are shape linters. They read the copied source for expected tokens and pass when the tokens are present. Most bundle components in this group are shape linters. They read the copied source for expected tokens and pass when the tokens are present. This one goes further for the parts it can. Several of its checks import the copied bodies and call the real functions on made-up inputs, so a pass records observed behaviour, not just presence.
\n
The question it answers is narrow. When these two subsystems are copied into the public clone and driven on a sample, do they still behave the way their internal contracts say, without touching a live ledger or dispatching any work? The question it answers is narrow. When these two subsystems are copied into the public clone and driven on a sample, do they still behave the way their internal contracts say, without touching a live ledger or dispatching any work?
\n
Purpose
\n
A loose check would grep the copied files for a function name and stop there. That proves the text was copied, not that it runs. A loose check would grep the copied files for a function name and stop there. That proves the text was copied, not that it runs. The one behaviour worth showing directly is digest preservation. The pipeline compresses a long input down to a short digest before it decides what to inspect next. If that compression silently drops an instruction, the downstream agent never sees it. So the component feeds the real digest_raw_seed an eighty-line block of low-signal text with one directive line buried inside, then checks the directive survives. The paired negative case removes the marker that recognises directive lines and confirms the directive is then lost. That pairing is the point: a compression step asserted to keep the one line that matters, with a test that fails when it does not.
\n
The standing limit is deliberate. The bounded compliance check runs the ledger builder in check-and-report mode, which reads and reports but never writes. The standing limit is deliberate. The bounded compliance check runs the ledger builder in check-and-report mode, which reads and reports but never writes. The pipeline checks stop before any dispatch. The result is evidence that the copied mechanics work on a sample, not a claim that the live compliance ledger is fresh or that every branch is covered.
\n
How it works
\n
run and run_batch8_compliance_pipeline_bundle both call run_crown_jewel_organ with _evaluate as the evaluator and evaluate_negative_case as the negative-case evaluator.run and run_batch8_compliance_pipeline_bundle both call run_crown_jewel_organ with _evaluate as the evaluator and evaluate_negative_case as the negative-case evaluator. _evaluate first decides its mode. If the input directory holds a source_module_manifest.json and a source_modules directory, it is an exported bundle and every check runs in standalone mode against the copied token contracts only. A plain fixture input lets the checks import live modules from the parent tree where present. _evaluate then runs six checks in order and passes only if all six pass and the manifest records at least eleven copied source modules, which is the count in SOURCE_REQUIRED_ANCHORS.
\n
The six checks:
\n
Function
What it does
_compliance_registry_runtime_witness
Reads the copied registry, coverage adapter, and ledger builder and confirms each carries its contract tokens. In live mode it also imports the registry and reads adapter, domain, and baseline counts as a shape witness.
_compliance_coverage_bounded_check
Runs the ledger builder as a subprocess with --check --report for two standards. Passes only if the run reports no write, refreshed both standard ids, and returned a truthful check status with a next-step command.
_baseline_companion_scanner_contract
Calls scan_standard_baseline on a sample standard and checks the returned row admits it is a baseline-inventory row with no domain adapter.
_pipeline_digest_and_shard_normalization
Imports the copied extract stage and exercises digest_raw_seed, _normalize_shards_payload, and _pick_diverse_shards.
_pipeline_observe_compile_helpers
Runs the compile-stage helpers that pull file mentions from free text, order follow-up files, and lift probe questions from a plan.
_pipeline_dispatch_process_boundary_contract
Calls _receipt_selection_meta and reads the execute, process, select, and emit stage bodies to confirm the dispatch boundary markers are present.
\n
Two of these are worth reading closely. _compliance_coverage_bounded_check builds the command build_compliance_ledger.py --check --report --standard-id std_compliance_coverage --standard-id std_microcosm, runs it through _run_public_witness, and parses the summary. Two of these are worth reading closely. _compliance_coverage_bounded_check builds the command build_compliance_ledger.py --check --report --standard-id std_compliance_coverage --standard-id std_microcosm, runs it through _run_public_witness, and parses the summary. The pass condition is strict. The run must report wrote_ledger false, both standard ids refreshed, a partial projection, a ratchet status of ready_next_row or closed_batch, and a truthful check result, where truthful means one of ok with no error findings, failed caused by error findings, or failed caused by a missing projection artifact. When the parent repo is not present the check falls back to _standalone_compliance_coverage_contract, which reads the copied builder for its bounded-check tokens instead of running it.
\n
_pipeline_digest_and_shard_normalization builds a source note of a heading, one narrative line, one directive line, and eighty low-signal lines, then digests it to 240 characters._pipeline_digest_and_shard_normalization builds a source note of a heading, one narrative line, one directive line, and eighty low-signal lines, then digests it to 240 characters. It checks directive_preserved. It also feeds _normalize_shards_payload a shard with status weird and checks the status is rewritten to pending while the original value is kept as status_variant. It calls _pick_diverse_shards and checks one concept group contributes at most two shards.
\n
The six negative cases each copy the bundle to a scratch directory, remove one load-bearing token from one copied body, and re-run the matching check. The six negative cases each copy the bundle to a scratch directory, remove one load-bearing token from one copied body, and re-run the matching check. _mutated_source_negative carries this out and reports blocked only when the check both blocks and reports the expected condition, so a green run means each contract was exercised rather than skipped.
\nDiagram of the mechanism (5 steps).Diagram source & refs
flowchart TD bundle["Copied source bundle 11 bodies body_in_receipt: false"] compliance["Compliance checks registry witness, bounded check, baseline scanner"] pipeline["Pipeline checks digest and shards, compile helpers, dispatch boundary"] negatives["6 negative cases remove one token per body; matching check must block"] records["Result records result, board, validation"] bundle --> compliance bundle --> pipeline bundle --> negatives compliance --> records pipeline --> records negatives --> records
\n
Negative cases
\n
EXPECTED_NEGATIVE_CASES names six. missing_compliance_registry removes _LazyComplianceAdapters from the copied registry and expects code BATCH8_COMPLIANCE_REGISTRY_REQUIRED.EXPECTED_NEGATIVE_CASES names six. missing_compliance_registry removes _LazyComplianceAdapters from the copied registry and expects code BATCH8_COMPLIANCE_REGISTRY_REQUIRED. bounded_check_failed removes _bounded_check_command from the builder. baseline_missing_standard flips baseline_inventory_only in the baseline adapter. digest_loses_directive removes the directive marker from the extract stage so the buried line drops. compile_helper_empty_mentions removes a compile helper name. pipeline_boundary_missing removes observe_dispatch_skipped from the execute stage. Each case passes only when the matching check turns from pass to blocked.
\n
Prior Art Grounding
\n
This component borrows from control-assessment, policy-as-code, provenance, and observability practice without adopting any one framework. This component borrows from control-assessment, policy-as-code, provenance, and observability practice without adopting any one framework. Useful anchors are NIST SP 800-53 Rev. 5 as a control-catalog pattern for naming and reporting control posture, Open Policy Agent as a pattern for evaluating structured inputs without embedding every rule in the caller, SLSA provenance for treating artifact origin as an explicit attestation, and OpenTelemetry for staged-pipeline instrumentation. The component borrows the scanner, policy, and pipeline-stage shapes and validates bounded no-write behaviour and pure helper mechanics only.
\n
Validation Result record Path
\n
Reader-verifiable commands, run from the microcosm-substrate/ public root:
\n
The fixture command writes the exercise result and sign-off JSON. The bundle command validates the copied source modules, manifest digests, observed negative cases, and body scans. The fixture command writes the exercise result and sign-off JSON. The bundle command validates the copied source modules, manifest digests, observed negative cases, and body scans. A pass means all six checks passed, the eleven source modules were present, and the six negative cases each blocked. These validation result records are the evidence a reader recomputes; they do not refresh the live ledger or dispatch any work.
\n
Scope boundary
\n
Scope limit
\n
The six-check fixture proof boundary
\n
The strongest claim the evidence supports is this. The copied compliance registry, bounded ledger builder, baseline scanner, and the extract, compile, execute, and process pipeline stages behave as their internal contracts say on a sample, the bounded check reads and reports without writing, and six perturbations each block the matching check. The strongest claim the evidence supports is this. The copied compliance registry, bounded ledger builder, baseline scanner, and the extract, compile, execute, and process pipeline stages behave as their internal contracts say on a sample, the bounded check reads and reports without writing, and six perturbations each block the matching check. That is the proof boundary.
\n
Claims outside the sample run
\n
It refuses more. It does not refresh the full compliance ledger, so stale unselected rows stay outside the claim. It refuses more. It does not refresh the full compliance ledger, so stale unselected rows stay outside the claim. It does not dispatch any provider or bridge work, change source records, or certify every compliance and pipeline branch. It is not a public sharing or launch-scope decision and not a whole-system correctness proof. That is the scope limit for this module.
\n",
"components": [
{
"anchor": "component-batch8_compliance_pipeline_capsule",
"component_id": "batch8_compliance_pipeline_capsule",
"public_label": "Compliance Pipeline Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 5,
"principle": 8
},
"link_count": 22,
"links": [
{
"id": "batch8_compliance_pipeline_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Compliance Pipeline Bundle",
"relation": "explains",
"url": "component-batch8-compliance-pipeline-capsule.html"
},
{
"id": "mechanism.batch8_compliance_pipeline_capsule.validates_public_compliance_pipeline_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public compliance pipeline bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch8-compliance-pipeline-capsule-validates-public-compliance-pipeline-capsule"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.engine_room_egress_self_compliance_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Egress Self-Compliance Gate",
"relation": "depends on",
"url": "paper-module-engine-room-egress-self-compliance-gate.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.batch11_saturation_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 11 Saturation Engines Bundle",
"relation": "depends on",
"url": "paper-module-batch11-saturation-engines-capsule.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch11_saturation_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 11 Saturation Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[73:paper_module.batch11_saturation_engines_capsule].depends_on",
"url": "paper-module-batch11-saturation-engines-capsule.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[19:paper_module.macro_projection_import_protocol].depends_on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"paper_module_id": "paper_module.batch8_compliance_pipeline_capsule",
"reader_packet": {
"abstract": "Set 8 Compliance Pipeline Bundle validates copied compliance scanner and observe-pipeline mechanics without refreshing the full ledger or dispatching bridge/provider work.",
"authority_ceiling": "Declared public compliance and pipeline fixture evidence plus copied source-module digest checks only; no full compliance-ledger refresh, bridge/external model access authority, source note mutation, repository mutation, complete compliance proof, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch8_compliance_pipeline_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch8-compliance-pipeline-capsule-diagram-1",
"html_url": "docs/paper-module-batch8-compliance-pipeline-capsule.html#paper-module-batch8-compliance-pipeline-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD bundle[\"Copied source bundle 11 bodies body_in_receipt: false\"] compliance[\"Compliance checks registry witness, bounded check, baseline scanner\"] pipeline[\"Pipeline checks digest and shards, compile helpers, dispatch boundary\"] negatives[\"6 negative cases remove one token per body; matching check must block\"] records[\"Result records result, board, validation\"] bundle --> compliance bundle --> pipeline bundle --> negatives compliance --> records pipeline --> records negatives --> records",
"source_format": "mermaid",
"source_hash": "sha256:2828497274e38f346101f8ab8237256495a8c188b1558bd94dcaaf4e597a9c2e",
"source_line_end": 76,
"source_line_start": 63
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
}
],
"relation": "abides by"
},
{
"count": 5,
"links": [
{
"id": "paper_module.engine_room_egress_self_compliance_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Egress Self-Compliance Gate",
"relation": "depends on",
"url": "paper-module-engine-room-egress-self-compliance-gate.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.batch11_saturation_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 11 Saturation Engines Bundle",
"relation": "depends on",
"url": "paper-module-batch11-saturation-engines-capsule.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch8_compliance_pipeline_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Compliance Pipeline Bundle",
"relation": "explains",
"url": "component-batch8-compliance-pipeline-capsule.html"
},
{
"id": "mechanism.batch8_compliance_pipeline_capsule.validates_public_compliance_pipeline_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public compliance pipeline bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch8-compliance-pipeline-capsule-validates-public-compliance-pipeline-capsule"
}
],
"relation": "explains"
},
{
"count": 9,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch8_compliance_pipeline_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 126,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_compliance_pipeline_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "SP 800-53 Rev. 5",
"section_id": "this-component-borrows-from-control-assessment",
"url": "https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final"
},
{
"kind": "external_reference",
"label": "Open Policy Agent",
"section_id": "this-component-borrows-from-control-assessment",
"url": "https://www.openpolicyagent.org/docs/latest"
},
{
"kind": "external_reference",
"label": "SLSA provenance",
"section_id": "this-component-borrows-from-control-assessment",
"url": "https://slsa.dev/spec/v1.2/provenance"
},
{
"kind": "external_reference",
"label": "OpenTelemetry",
"section_id": "this-component-borrows-from-control-assessment",
"url": "https://opentelemetry.io/docs/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "batch8compliancepipelinecapsule takes two internal subsystems, copies their source bodies into the public clone, and runs them on synthetic inputs.** batch8compliancepipeline_capsule takes two internal subsystems, copies their source bodies into the public clone, and runs them on synthetic inputs. The first subsystem is a compliance scanner registry and its bounded ledger builder. The second is a six-stage pipeline that turns a long block of raw text into a short observe plan. The component runs six checks over the copied bodies and writes result records that carry no source text.",
"level": 4,
"order": 1,
"section_id": "batch8-compliance-pipeline-capsule-takes-two-internal-subsystems",
"source_line_end": 7,
"source_line_start": 3,
"title": "`batch8_compliance_pipeline_capsule` takes two internal subsystems"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Most bundle components in this group are shape linters. They read the copied source for expected tokens and pass when the tokens are present. Most bundle components in this group are shape linters. They read the copied source for expected tokens and pass when the tokens are present. This one goes further for the parts it can. Several of its checks import the copied bodies and call the real functions on made-up inputs, so a pass records observed behaviour, not just presence.",
"level": 4,
"order": 2,
"section_id": "most-capsule-components-in-this-group-are-shape-linters",
"source_line_end": 12,
"source_line_start": 8,
"title": "Most bundle components in this group are shape linters"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The question it answers is narrow. When these two subsystems are copied into the public clone and driven on a sample, do they still behave the way their internal contracts say, without touching a live ledger or dispatching any work? The question it answers is narrow. When these two subsystems are copied into the public clone and driven on a sample, do they still behave the way their internal contracts say, without touching a live ledger or dispatching any work?",
"level": 4,
"order": 3,
"section_id": "the-question-it-answers-is-narrow",
"source_line_end": 17,
"source_line_start": 13,
"title": "The question it answers is narrow"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 19,
"source_line_start": 18,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A loose check would grep the copied files for a function name and stop there. That proves the text was copied, not that it runs. A loose check would grep the copied files for a function name and stop there. That proves the text was copied, not that it runs. The one behaviour worth showing directly is digest preservation. The pipeline compresses a long input down to a short digest before it decides what to inspect next. If that compression silently drops an instruction, the downstream agent never sees it. So the component feeds the real digestrawseed an eighty-line block of low-signal text with one directive line buried inside, then checks the directive survives. The paired negative case rem\u2026",
"level": 4,
"order": 5,
"section_id": "a-loose-check-would-grep-the-copied-files-for-a-function-name-and-stop-there",
"source_line_end": 24,
"source_line_start": 20,
"title": "A loose check would grep the copied files for a function name and stop there"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The standing limit is deliberate. The bounded compliance check runs the ledger builder in check-and-report mode, which reads and reports but never writes. The standing limit is deliberate. The bounded compliance check runs the ledger builder in check-and-report mode, which reads and reports but never writes. The pipeline checks stop before any dispatch. The result is evidence that the copied mechanics work on a sample, not a claim that the live compliance ledger is fresh or that every branch is covered.",
"level": 4,
"order": 6,
"section_id": "the-standing-limit-is-deliberate",
"source_line_end": 29,
"source_line_start": 25,
"title": "The standing limit is deliberate"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 31,
"source_line_start": 30,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "run and runbatch8compliancepipelinebundle both call runcrownjewelorgan with evaluate as the evaluator and evaluatenegativecase as the negative-case evaluator. run and runbatch8compliancepipelinebundle both call runcrownjewelorgan with evaluate as the evaluator and evaluatenegativecase as the negative-case evaluator. evaluate first decides its mode. If the input directory holds a sourcemodulemanifest.json and a sourcemodules directory, it is an exported bundle and every check runs in standalone mode against the copied token contracts only. A plain fixture input lets the checks import live modules from the parent tree where present. evaluate then runs six checks in order and passes only if al\u2026",
"level": 4,
"order": 8,
"section_id": "run-and-run-batch8-compliance-pipeline-bundle-both-call-run-crown-jewel-organ-with-evaluate-as-the-evaluator",
"source_line_end": 47,
"source_line_start": 32,
"title": "`run` and `run_batch8_compliance_pipeline_bundle` both call `run_crown_jewel_organ` with `_evaluate` as the evaluator"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Two of these are worth reading closely. compliancecoverageboundedcheck builds the command buildcomplianceledger.py --check --report --standard-id stdcompliancecoverage --standard-id stdmicrocosm, runs it through runpublicwitness, and parses the summary. Two of these are worth reading closely. compliancecoverageboundedcheck builds the command buildcomplianceledger.py --check --report --standard-id stdcompliancecoverage --standard-id stdmicrocosm, runs it through runpublicwitness, and parses the summary. The pass condition is strict. The run must report wroteledger false, both standard ids refreshed, a partial projection, a ratchet status of readynextrow or closedbatch, and a truthful check r\u2026",
"level": 4,
"order": 9,
"section_id": "two-of-these-are-worth-reading-closely",
"source_line_end": 52,
"source_line_start": 48,
"title": "Two of these are worth reading closely"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "pipelinedigestandshardnormalization builds a source note of a heading, one narrative line, one directive line, and eighty low-signal lines, then digests it to 240 characters. pipelinedigestandshardnormalization builds a source note of a heading, one narrative line, one directive line, and eighty low-signal lines, then digests it to 240 characters. It checks directivepreserved. It also feeds normalizeshardspayload a shard with status weird and checks the status is rewritten to pending while the original value is kept as statusvariant. It calls pickdiverse_shards and checks one concept group contributes at most two shards.",
"level": 4,
"order": 10,
"section_id": "pipeline-digest-and-shard-normalization-builds-a-raw-seed-of-a-heading",
"source_line_end": 57,
"source_line_start": 53,
"title": "`_pipeline_digest_and_shard_normalization` builds a source note of a heading"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "The six negative cases each copy the bundle to a scratch directory, remove one load-bearing token from one copied body, and re-run the matching check. The six negative cases each copy the bundle to a scratch directory, remove one load-bearing token from one copied body, and re-run the matching check. mutatedsource_negative carries this out and reports blocked only when the check both blocks and reports the expected condition, so a green run means each contract was exercised rather than skipped.",
"level": 4,
"order": 11,
"section_id": "the-six-negative-cases-each-copy-the-bundle-to-a-scratch-directory",
"source_line_end": 77,
"source_line_start": 58,
"title": "The six negative cases each copy the bundle to a scratch directory"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 79,
"source_line_start": 78,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVECASES names six. missingcomplianceregistry removes LazyComplianceAdapters from the copied registry and expects code BATCH8COMPLIANCEREGISTRYREQUIRED. EXPECTEDNEGATIVECASES names six. missingcomplianceregistry removes LazyComplianceAdapters from the copied registry and expects code BATCH8COMPLIANCEREGISTRYREQUIRED. boundedcheckfailed removes boundedcheckcommand from the builder. baselinemissingstandard flips baselineinventoryonly in the baseline adapter. digestlosesdirective removes the directive marker from the extract stage so the buried line drops. compilehelperemptymentions removes a compile helper name. pipelineboundarymissing removes observedispatchskipped from the exec\u2026",
"level": 4,
"order": 13,
"section_id": "expected-negative-cases-names-six",
"source_line_end": 84,
"source_line_start": 80,
"title": "`EXPECTED_NEGATIVE_CASES` names six"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "prior-art-grounding",
"source_line_end": 86,
"source_line_start": 85,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "This component borrows from control-assessment, policy-as-code, provenance, and observability practice without adopting any one framework. This component borrows from control-assessment, policy-as-code, provenance, and observability practice without adopting any one framework. Useful anchors are NIST SP 800-53 Rev. 5 as a control-catalog pattern for naming and reporting control posture, Open Policy Agent as a pattern for evaluating structured inputs without embedding every rule in the caller, SLSA provenance for treating artifact origin as an explicit attestation, and OpenTelemetry for staged-pipeline instrumentation. The component borrows the scanner, policy, and pipeline-stage shapes and\u2026",
"level": 4,
"order": 15,
"section_id": "this-component-borrows-from-control-assessment",
"source_line_end": 91,
"source_line_start": 87,
"title": "This component borrows from control-assessment"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Reader-verifiable commands, run from the microcosm-substrate/ public root:",
"level": 2,
"order": 16,
"section_id": "validation-receipt-path",
"source_line_end": 108,
"source_line_start": 92,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The fixture command writes the exercise result and sign-off JSON. The bundle command validates the copied source modules, manifest digests, observed negative cases, and body scans. The fixture command writes the exercise result and sign-off JSON. The bundle command validates the copied source modules, manifest digests, observed negative cases, and body scans. A pass means all six checks passed, the eleven source modules were present, and the six negative cases each blocked. These validation result records are the evidence a reader recomputes; they do not refresh the live ledger or dispatch any work.",
"level": 4,
"order": 17,
"section_id": "the-fixture-command-writes-the-exercise-result-and-acceptance-json",
"source_line_end": 113,
"source_line_start": 109,
"title": "The fixture command writes the exercise result and sign-off JSON"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "claim-ceiling",
"source_line_end": 115,
"source_line_start": 114,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim the evidence supports is this. The copied compliance registry, bounded ledger builder, baseline scanner, and the extract, compile, execute, and process pipeline stages behave as their internal contracts say on a sample, the bounded check reads and reports without writing, and six perturbations each block the matching check. The strongest claim the evidence supports is this. The copied compliance registry, bounded ledger builder, baseline scanner, and the extract, compile, execute, and process pipeline stages behave as their internal contracts say on a sample, the bounded check reads and reports without writing, and six perturbations each block the matching check. That is\u2026",
"level": 4,
"order": 19,
"section_id": "the-six-check-fixture-proof-boundary",
"source_line_end": 120,
"source_line_start": 116,
"title": "The six-check fixture proof boundary"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It refuses more. It does not refresh the full compliance ledger, so stale unselected rows stay outside the claim. It refuses more. It does not refresh the full compliance ledger, so stale unselected rows stay outside the claim. It does not dispatch any provider or bridge work, change source records, or certify every compliance and pipeline branch. It is not a public sharing or launch-scope decision and not a whole-system correctness proof. That is the scope limit for this module.",
"level": 4,
"order": 20,
"section_id": "claims-outside-the-sample-run",
"source_line_end": 125,
"source_line_start": 121,
"title": "Claims outside the sample run"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch8_compliance_pipeline_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 8 Compliance Pipeline Bundle validates copied compliance scanner and observe-pipeline mechanics without refreshing the full ledger or dispatching bridge/provider work."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch8_compliance_pipeline_capsule.py",
"path": "src/microcosm_core/organs/batch8_compliance_pipeline_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for bounded compliance scanner, pipeline helper, source digest, negative-case, and scope limit checks.",
"symbols": [
"_compliance_registry_runtime_witness",
"_compliance_coverage_bounded_check",
"_baseline_companion_scanner_contract",
"_pipeline_digest_and_shard_normalization",
"_pipeline_observe_compile_helpers",
"_pipeline_dispatch_process_boundary_contract",
"_evaluate",
"run"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch8_compliance_pipeline_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_compliance_pipeline_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_compliance_pipeline_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 8 Compliance Pipeline Bundle imports compliance scanner registry, bounded compliance ledger builder, and observe-loop pipeline helper bodies into a runnable component. It checks registry shape, bounded no-write compliance checks, baseline scanner truth accounting, pure pipeline helper behavior, source-module digests, negative cases, and scope limits while excluding full compliance-ledger refresh, bridge/external model access, source note mutation, repository mutation, launch, public sharing, and complete compliance proof.",
"summary_line": "Set 8 Compliance Pipeline Bundle validates copied compliance scanner and observe-pipeline mechanics without refreshing the full ledger or dispatching bridge/provider work.",
"title": "Set 8 Compliance Pipeline Bundle"
},
{
"anchor": "paper-module-batch8-policy-engines-capsule",
"authority_ceiling": "Deterministic public policy-engine fixture evidence and copied source refs only; no live campaign execution, external model access, repository mutation, private artifact export, market validation, publishing-scope decision, launch-scope decision, or whole-system correctness.",
"body_html": "
This component takes three copied public policy-engine bodies and runs each one against a small fixed fixture to confirm it still makes the same deterministic decision. This component takes three copied public policy-engine bodies and runs each one against a small fixed fixture to confirm it still makes the same deterministic decision. The three engines are a lab contract audit red/green gate, a market-fusion readiness gate that fails closed, and a campaign dispatch transition adjudicator.
\n
It is a checker, not an engine host. It does not run live campaigns, use external model services or models, mutate any repository, validate a market, or approve public sharing or launch. It is a checker, not an engine host. It does not run live campaigns, use external model services or models, mutate any repository, validate a market, or approve public sharing or launch. What it emits is a metadata-only result record: which engines were found, what each exercise decided, and the three stable negative-case codes that fired.
\n
Purpose
\n
Each of the three copied bodies shares one shape: a machine-checkable gate that runs before any judgement or downstream action and refuses by default when evidence is absent. Each of the three copied bodies shares one shape: a machine-checkable gate that runs before any judgement or downstream action and refuses by default when evidence is absent. Copying a gate into a public bundle is cheap. Confirming the copy still behaves like the original is the part that needs a test. This component answers one question for a cold reader: do these copied gate bodies make the same red, refuse, and reject decisions when driven by public fixture inputs?
\n
A loose check would import the modules and assert they load. That passes even if a gate silently started accepting what it used to reject. A loose check would import the modules and assert they load. That passes even if a gate silently started accepting what it used to reject. So each exercise drives a concrete decision with a known answer and marks itself blocked the moment the decision changes.
\n
How it works
\n
run and run_batch8_policy_engines_bundle both forward torun_crown_jewel_organ with the shared SPEC, one for fixture input and one for the exported bundle.run and run_batch8_policy_engines_bundle both forward to run_crown_jewel_organ with the shared SPEC, one for fixture input and one for the exported bundle. The work happens in _capsule_evaluator. It loads the probe manifest, reads the declared engine ids, and checks that set equals the three ids in EXPECTED_ENGINES. A mismatch records BATCH8_POLICY_ENGINE_SET_MISMATCH. It then reads the source-module manifest through _source_rows and checks the three required module ids are present; a gap records BATCH8_POLICY_ENGINE_SOURCE_MODULE_MISSING. Only when all three modules are present does it run the exercises. Any exercise that returns a status other than pass records BATCH8_POLICY_ENGINE_RUNTIME_BLOCKED. The bundle is pass only when the findings list is empty.
\n
The copied bodies are loaded by_load_copied_module, which resolves each module's target path from the manifest and executes it under a temporary sys.path, so a body import failure surfaces as a blocked exercise rather than a crash. The copied bodies are loaded by _load_copied_module, which resolves each module's target path from the manifest and executes it under a temporary sys.path, so a body import failure surfaces as a blocked exercise rather than a crash.
\n
_exercise_lab_contract_audit writes a green set of lab node artifacts, runs the copied compute_lab_contract_audit, and expects status green._exercise_lab_contract_audit writes a green set of lab node artifacts, runs the copied compute_lab_contract_audit, and expects status green. It then rewrites one compute-node output to carry a banned ? and runs the audit again, expecting status red with QUESTION_MARK_OUTPUT in the hard-fail list. Both halves must hold or the exercise is blocked.
\n
_exercise_market_fusion calls the copied preflight_candidate_situation three ways. A registered situation must refuse with named reasons that do not include candidate_situation_gate_missing._exercise_market_fusion calls the copied preflight_candidate_situation three ways. A registered situation must refuse with named reasons that do not include candidate_situation_gate_missing. An unregistered situation must refuse with candidate_situation_gate_missing. A malformed payload sent through preflight_consumer_claims must also refuse with that same missing-gate reason. The distinction is the point: a refusal the gate reasoned about stays legible as different from a fail-closed default on something it never saw.
\n
_exercise_campaign_transition calls the copied validate_dispatch_transition. A candidate to blocked move must return legal_transition._exercise_campaign_transition calls the copied validate_dispatch_transition. A candidate to blocked move must return legal_transition. A completed to completed move must return already_target. A completed to running move must raise CampaignTransitionError with the word terminal in its message, because completed is a terminal state.
\n
evaluate_negative_case re-runs the matching exercise for each of the three declared cases and reports the stable error code when the refusal holds.evaluate_negative_case re-runs the matching exercise for each of the three declared cases and reports the stable error code when the refusal holds. result_card projects a run into a compact public card carrying engine ids, the authority floor from AUTHORITY_CEILING, and a metadata-only flag set, with no source bodies or scan scope in the card.
\nDiagram of the mechanism (6 steps).Diagram source & refs
flowchart TD manifest["Probe manifest + source manifest"] evaluator["_capsule_evaluator engine set == EXPECTED_ENGINES three source modules present"] lab["_exercise_lab_contract_audit green then red on banned ?"] market["_exercise_market_fusion refuse; fail-closed on missing gate"] campaign["_exercise_campaign_transition legal / already_target / terminal"] result["metadata-only result record pass only when findings empty"] manifest --> evaluator evaluator --> lab evaluator --> market evaluator --> campaign lab --> result market --> result campaign --> result
\n
Negative cases
\n
The fixture declares three negative cases, one per engine, and EXPECTED_NEGATIVE_CASES maps each to its stable code. The fixture declares three negative cases, one per engine, and EXPECTED_NEGATIVE_CASES maps each to its stable code. lab_contract_question_mark_red_gate fires BATCH8_LAB_CONTRACT_QUESTION_MARK_RED_GATE when the injected ? output forces a red audit. market_fusion_missing_gate_refused fires BATCH8_MARKET_FUSION_MISSING_GATE_REFUSED when an unregistered situation refuses with the missing-gate reason. campaign_completed_to_running_refused fires BATCH8_CAMPAIGN_COMPLETED_TO_RUNNING_REFUSED when the terminal-state move is rejected. A green run with none of these firing would mean a gate stopped refusing, which the bundle reports as blocked.
\n
Prior Art Grounding
\n
Policy-as-code evaluation without policy authority
\n
The design borrows the policy-as-code idea of evaluating policy as a separate engine over structured input, as in Open Policy Agent. The design borrows the policy-as-code idea of evaluating policy as a separate engine over structured input, as in Open Policy Agent. The explicit risk-gate and red/green decision posture echoes NIST's AI Risk Management Framework. The market-claim language stays inside the boundary described by the CFTC's prediction markets explainer. This component keeps the deterministic-gate and fail-closed-preflight shape and holds it to fixture inputs and copied public source.
\n
Validation Result record Path
\n
Run from the microcosm-substrate/ public root:
\n
The first command writes the fixture result record and sign-off JSON. The second validates the copied source bundle, its manifest digests, negative cases, and the metadata-only posture. The first command writes the fixture result record and sign-off JSON. The second validates the copied source bundle, its manifest digests, negative cases, and the metadata-only posture. The test checks the red/green gate, the fail-closed refusals, bundle validation, and the no-launch ceiling. A pass means the copied gates still decide the fixtures the same way. It is bounded checker evidence, nothing more.
\n
Scope boundary
\n
Scope limit
\n
Three copied policy engines under fixture inputs
\n
The strongest claim this component supports: three copied public policy-engine bodies still make their expected deterministic decisions over synthetic fixture inputs, and the result record carries no source bodies or private state. The strongest claim this component supports: three copied public policy-engine bodies still make their expected deterministic decisions over synthetic fixture inputs, and the result record carries no source bodies or private state. That is the whole proof boundary.
\n
Claims excluded past the fixture boundary
\n
It refuses everything past that line. It is not lab correctness, not live campaign execution, not market validation, not provider or model dispatch, not repository or source-file changes, not whole-system safety, and not public sharing or launch-scope decision. It refuses everything past that line. It is not lab correctness, not live campaign execution, not market validation, not provider or model dispatch, not repository or source-file changes, not whole-system safety, and not public sharing or launch-scope decision. Those would need changed bundle authority, not a green run here. The scope limit is a public-system bundle, and the validation result records above are the only evidence it stands on.
\n",
"components": [
{
"anchor": "component-batch8_policy_engines_capsule",
"component_id": "batch8_policy_engines_capsule",
"public_label": "Policy Engines Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 8
},
"link_count": 20,
"links": [
{
"id": "batch8_policy_engines_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Policy Engines Bundle",
"relation": "explains",
"url": "component-batch8-policy-engines-capsule.html"
},
{
"id": "mechanism.batch8_policy_engines_capsule.validates_public_policy_engines_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public policy engines bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch8-policy-engines-capsule-validates-public-policy-engines-capsule"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-4",
"kind": "principle",
"kind_label": "principle",
"label": "Possession is not permission",
"relation": "governed by",
"url": "doctrine.html#dcard-p-4"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-3",
"kind": "axiom",
"kind_label": "axiom",
"label": "Authority by derivation, not possession",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-3"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.batch7_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Source Engines Bundle",
"relation": "depends on",
"url": "paper-module-batch7-macro-engines-capsule.html"
},
{
"id": "paper_module.proof_derived_governed_mutation_authorization",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof-Derived Governed Mutation Authorization",
"relation": "depends on",
"url": "paper-module-proof-derived-governed-mutation-authorization.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.batch8_policy_engines_capsule",
"reader_packet": {
"abstract": "Set 8 Policy Engines Bundle validates three deterministic public policy-engine exercises without running campaigns, providers, markets, or repository mutations.",
"authority_ceiling": "Deterministic public policy-engine fixture evidence and copied source refs only; no live campaign execution, external model access, repository mutation, private artifact export, market validation, publishing-scope decision, launch-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch8_policy_engines_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch8-policy-engines-capsule-diagram-1",
"html_url": "docs/paper-module-batch8-policy-engines-capsule.html#paper-module-batch8-policy-engines-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD manifest[\"Probe manifest + source manifest\"] evaluator[\"_capsule_evaluator engine set == EXPECTED_ENGINES three source modules present\"] lab[\"_exercise_lab_contract_audit green then red on banned ?\"] market[\"_exercise_market_fusion refuse; fail-closed on missing gate\"] campaign[\"_exercise_campaign_transition legal / already_target / terminal\"] result[\"metadata-only result record pass only when findings empty\"] manifest --> evaluator evaluator --> lab evaluator --> market evaluator --> campaign lab --> result market --> result campaign --> result",
"source_format": "mermaid",
"source_hash": "sha256:68f47d3a48cca33ef1e24e0fe79e5c8b75b112017c6ad3db69ae93e66cf544ea",
"source_line_end": 72,
"source_line_start": 57
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-3",
"kind": "axiom",
"kind_label": "axiom",
"label": "Authority by derivation, not possession",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-3"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.batch7_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Source Engines Bundle",
"relation": "depends on",
"url": "paper-module-batch7-macro-engines-capsule.html"
},
{
"id": "paper_module.proof_derived_governed_mutation_authorization",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof-Derived Governed Mutation Authorization",
"relation": "depends on",
"url": "paper-module-proof-derived-governed-mutation-authorization.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch8_policy_engines_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Policy Engines Bundle",
"relation": "explains",
"url": "component-batch8-policy-engines-capsule.html"
},
{
"id": "mechanism.batch8_policy_engines_capsule.validates_public_policy_engines_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public policy engines bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch8-policy-engines-capsule-validates-public-policy-engines-capsule"
}
],
"relation": "explains"
},
{
"count": 9,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-4",
"kind": "principle",
"kind_label": "principle",
"label": "Possession is not permission",
"relation": "governed by",
"url": "doctrine.html#dcard-p-4"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch8_policy_engines_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 122,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_policy_engines_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Open Policy Agent",
"section_id": "policy-as-code-evaluation-without-policy-authority",
"url": "https://www.openpolicyagent.org/docs/latest"
},
{
"kind": "external_reference",
"label": "AI Risk Management Framework",
"section_id": "policy-as-code-evaluation-without-policy-authority",
"url": "https://www.nist.gov/itl/ai-risk-management-framework"
},
{
"kind": "external_reference",
"label": "prediction markets",
"section_id": "policy-as-code-evaluation-without-policy-authority",
"url": "https://www.cftc.gov/LearnandProtect/PredictionMarkets"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "This component takes three copied public policy-engine bodies and runs each one against a small fixed fixture to confirm it still makes the same deterministic decision. This component takes three copied public policy-engine bodies and runs each one against a small fixed fixture to confirm it still makes the same deterministic decision. The three engines are a lab contract audit red/green gate, a market-fusion readiness gate that fails closed, and a campaign dispatch transition adjudicator.",
"level": 4,
"order": 1,
"section_id": "this-component-takes-three-copied-public-policy-engine-bodies",
"source_line_end": 7,
"source_line_start": 3,
"title": "This component takes three copied public policy-engine bodies"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It is a checker, not an engine host. It does not run live campaigns, use external model services or models, mutate any repository, validate a market, or approve public sharing or launch. It is a checker, not an engine host. It does not run live campaigns, use external model services or models, mutate any repository, validate a market, or approve public sharing or launch. What it emits is a metadata-only result record: which engines were found, what each exercise decided, and the three stable negative-case codes that fired.",
"level": 4,
"order": 2,
"section_id": "it-is-a-checker-not-an-engine-host",
"source_line_end": 12,
"source_line_start": 8,
"title": "It is a checker, not an engine host"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 3,
"section_id": "purpose",
"source_line_end": 14,
"source_line_start": 13,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Each of the three copied bodies shares one shape: a machine-checkable gate that runs before any judgement or downstream action and refuses by default when evidence is absent. Each of the three copied bodies shares one shape: a machine-checkable gate that runs before any judgement or downstream action and refuses by default when evidence is absent. Copying a gate into a public bundle is cheap. Confirming the copy still behaves like the original is the part that needs a test. This component answers one question for a cold reader: do these copied gate bodies make the same red, refuse, and reject decisions when driven by public fixture inputs?",
"level": 4,
"order": 4,
"section_id": "each-of-the-three-copied-bodies-shares-one-shape",
"source_line_end": 19,
"source_line_start": 15,
"title": "Each of the three copied bodies shares one shape"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A loose check would import the modules and assert they load. That passes even if a gate silently started accepting what it used to reject. A loose check would import the modules and assert they load. That passes even if a gate silently started accepting what it used to reject. So each exercise drives a concrete decision with a known answer and marks itself blocked the moment the decision changes.",
"level": 4,
"order": 5,
"section_id": "a-loose-check-would-import-the-modules-and-assert-they-load",
"source_line_end": 24,
"source_line_start": 20,
"title": "A loose check would import the modules and assert they load"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 6,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run and runbatch8policyenginesbundle both forward to runcrownjewelorgan with the shared SPEC, one for fixture input and one for the exported bundle.** run and runbatch8policyenginesbundle both forward to runcrownjewelorgan with the shared SPEC, one for fixture input and one for the exported bundle. The work happens in capsuleevaluator. It loads the probe manifest, reads the declared engine ids, and checks that set equals the three ids in EXPECTEDENGINES. A mismatch records BATCH8POLICYENGINESETMISMATCH. It then reads the source-module manifest through sourcerows and checks the three required module ids are present; a gap records BATCH8POLICYENGINESOURCEMODULEMISSING. Only when all three mod\u2026",
"level": 4,
"order": 7,
"section_id": "run-and-run-batch8-policy-engines-bundle-both-forward-to-run-crown-jewel-organ-with-the-shared-spec-one-for-fixture-input",
"source_line_end": 31,
"source_line_start": 27,
"title": "`run` and `run_batch8_policy_engines_bundle` both forward to `run_crown_jewel_organ` with the shared `SPEC`, one for fixture input"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The copied bodies are loaded by loadcopiedmodule, which resolves each module's target path from the manifest and executes it under a temporary sys.path, so a body import failure surfaces as a blocked exercise rather than a crash.** The copied bodies are loaded by loadcopied_module, which resolves each module's target path from the manifest and executes it under a temporary sys.path, so a body import failure surfaces as a blocked exercise rather than a crash.",
"level": 4,
"order": 8,
"section_id": "the-copied-bodies-are-loaded-by-load-copied-module",
"source_line_end": 36,
"source_line_start": 32,
"title": "The copied bodies are loaded by `_load_copied_module`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "exerciselabcontractaudit writes a green set of lab node artifacts, runs the copied computelabcontractaudit, and expects status green. exerciselabcontractaudit writes a green set of lab node artifacts, runs the copied computelabcontractaudit, and expects status green. It then rewrites one compute-node output to carry a banned ? and runs the audit again, expecting status red with QUESTIONMARK_OUTPUT in the hard-fail list. Both halves must hold or the exercise is blocked.",
"level": 4,
"order": 9,
"section_id": "exercise-lab-contract-audit-writes-a-green-set-of-lab-node-artifacts",
"source_line_end": 41,
"source_line_start": 37,
"title": "`_exercise_lab_contract_audit` writes a green set of lab node artifacts"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "exercisemarketfusion calls the copied preflightcandidatesituation three ways. A registered situation must refuse with named reasons that do not include candidatesituationgatemissing.** exercisemarketfusion calls the copied preflightcandidatesituation three ways. A registered situation must refuse with named reasons that do not include candidatesituationgatemissing. An unregistered situation must refuse with candidatesituationgatemissing. A malformed payload sent through preflightconsumer_claims must also refuse with that same missing-gate reason. The distinction is the point: a refusal the gate reasoned about stays legible as different from a fail-closed default on something it never saw.",
"level": 4,
"order": 10,
"section_id": "exercise-market-fusion-calls-the-copied-preflight-candidate-situation-three-ways",
"source_line_end": 46,
"source_line_start": 42,
"title": "`_exercise_market_fusion` calls the copied `preflight_candidate_situation` three ways"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "exercisecampaigntransition calls the copied validatedispatchtransition. A candidate to blocked move must return legaltransition.** exercisecampaigntransition calls the copied validatedispatchtransition. A candidate to blocked move must return legaltransition. A completed to completed move must return already_target. A completed to running move must raise CampaignTransitionError with the word terminal in its message, because completed is a terminal state.",
"level": 4,
"order": 11,
"section_id": "exercise-campaign-transition-calls-the-copied-validate-dispatch-transition",
"source_line_end": 51,
"source_line_start": 47,
"title": "`_exercise_campaign_transition` calls the copied `validate_dispatch_transition`"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "evaluatenegativecase re-runs the matching exercise for each of the three declared cases and reports the stable error code when the refusal holds. evaluatenegativecase re-runs the matching exercise for each of the three declared cases and reports the stable error code when the refusal holds. resultcard projects a run into a compact public card carrying engine ids, the authority floor from AUTHORITYCEILING, and a metadata-only flag set, with no source bodies or scan scope in the card.",
"level": 4,
"order": 12,
"section_id": "evaluate-negative-case-re-runs-the-matching-exercise-for-each-of-the-three-declared-cases",
"source_line_end": 73,
"source_line_start": 52,
"title": "`evaluate_negative_case` re-runs the matching exercise for each of the three declared cases"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 75,
"source_line_start": 74,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The fixture declares three negative cases, one per engine, and EXPECTEDNEGATIVECASES maps each to its stable code. The fixture declares three negative cases, one per engine, and EXPECTEDNEGATIVECASES maps each to its stable code. labcontractquestionmarkredgate fires BATCH8LABCONTRACTQUESTIONMARKREDGATE when the injected ? output forces a red audit. marketfusionmissinggaterefused fires BATCH8MARKETFUSIONMISSINGGATEREFUSED when an unregistered situation refuses with the missing-gate reason. campaigncompletedtorunningrefused fires BATCH8CAMPAIGNCOMPLETEDTORUNNING_REFUSED when the terminal-state move is rejected. A green run with none of these firing would mean a gate stopped refusing, which th\u2026",
"level": 4,
"order": 14,
"section_id": "the-fixture-declares-three-negative-cases",
"source_line_end": 80,
"source_line_start": 76,
"title": "The fixture declares three negative cases"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 82,
"source_line_start": 81,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The design borrows the policy-as-code idea of evaluating policy as a separate engine over structured input, as in Open Policy Agent. The design borrows the policy-as-code idea of evaluating policy as a separate engine over structured input, as in Open Policy Agent. The explicit risk-gate and red/green decision posture echoes NIST's AI Risk Management Framework. The market-claim language stays inside the boundary described by the CFTC's prediction markets explainer. This component keeps the deterministic-gate and fail-closed-preflight shape and holds it to fixture inputs and copied public source.",
"level": 4,
"order": 16,
"section_id": "policy-as-code-evaluation-without-policy-authority",
"source_line_end": 87,
"source_line_start": 83,
"title": "Policy-as-code evaluation without policy authority"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from the microcosm-substrate/ public root:",
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 104,
"source_line_start": 88,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The first command writes the fixture result record and sign-off JSON. The second validates the copied source bundle, its manifest digests, negative cases, and the metadata-only posture. The first command writes the fixture result record and sign-off JSON. The second validates the copied source bundle, its manifest digests, negative cases, and the metadata-only posture. The test checks the red/green gate, the fail-closed refusals, bundle validation, and the no-launch ceiling. A pass means the copied gates still decide the fixtures the same way. It is bounded checker evidence, nothing more.",
"level": 4,
"order": 18,
"section_id": "the-first-command-writes-the-fixture-receipt-and-acceptance-json",
"source_line_end": 109,
"source_line_start": 105,
"title": "The first command writes the fixture result record and sign-off JSON"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "claim-ceiling",
"source_line_end": 111,
"source_line_start": 110,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim this component supports: three copied public policy-engine bodies still make their expected deterministic decisions over synthetic fixture inputs, and the result record carries no source bodies or private state. The strongest claim this component supports: three copied public policy-engine bodies still make their expected deterministic decisions over synthetic fixture inputs, and the result record carries no source bodies or private state. That is the whole proof boundary.",
"level": 4,
"order": 20,
"section_id": "three-copied-policy-engines-under-fixture-inputs",
"source_line_end": 116,
"source_line_start": 112,
"title": "Three copied policy engines under fixture inputs"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It refuses everything past that line. It is not lab correctness, not live campaign execution, not market validation, not provider or model dispatch, not repository or source-file changes, not whole-system safety, and not public sharing or launch-scope decision. It refuses everything past that line. It is not lab correctness, not live campaign execution, not market validation, not provider or model dispatch, not repository or source-file changes, not whole-system safety, and not public sharing or launch-scope decision. Those would need changed bundle authority, not a green run here. The scope limit is a public-system bundle, and the validation result records above are the only evidence it st\u2026",
"level": 4,
"order": 21,
"section_id": "claims-excluded-past-the-fixture-boundary",
"source_line_end": 121,
"source_line_start": 117,
"title": "Claims excluded past the fixture boundary"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch8_policy_engines_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 8 Policy Engines Bundle validates three deterministic public policy-engine exercises without running campaigns, providers, markets, or repository mutations."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch8_policy_engines_capsule.py",
"path": "src/microcosm_core/organs/batch8_policy_engines_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for lab-contract, market-fusion, campaign-transition, source-manifest, negative-case, and scope limit checks.",
"symbols": [
"_exercise_lab_contract_audit",
"_exercise_market_fusion",
"_exercise_campaign_transition",
"_capsule_evaluator",
"run",
"run_batch8_policy_engines_bundle",
"result_card",
"EXPECTED_NEGATIVE_CASES"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch8_policy_engines_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_policy_engines_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_policy_engines_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 8 Policy Engines Bundle imports Lab contract audit red/green gating, market-fusion fail-closed claim preflight, and campaign dispatch transition adjudication as exact copied source bodies with bounded public exercises. It checks source-module manifests, stable negative cases, exercise outcomes, source digests, and scope limits while excluding live campaigns, external model access, repository mutation, private artifact export, market-level conclusions, launch, public sharing, and whole-system safety.",
"summary_line": "Set 8 Policy Engines Bundle validates three deterministic public policy-engine exercises without running campaigns, providers, markets, or repository mutations.",
"title": "Set 8 Policy Engines Bundle"
},
{
"anchor": "paper-module-batch8-structural-theses-capsule",
"authority_ceiling": "Deterministic fixture evidence over public synthetic thesis rows and copied source refs only; no financial decisions, investment recommendation, live-market validation, external model access, portfolio authority, publishing-scope decision, launch-scope decision, or whole-system correctness.",
"body_html": "
Structural theses run through the copied evaluator
\n
This bundle imports the finance module tools/finance/structural_theses.py as a copied body and runs it over public synthetic thesis rows rather than building a new evaluator. This component imports one finance module, tools/finance/structural_theses.py, as an exact copied source body and runs it over public synthetic thesis rows. It does not build a new evaluator. It drives the copied source's own family builder and validator and records what they return.
\n
Losers remain visible beside surviving families
\n
The copied source asks which reasoning families survive once every claim is resolved forward with failures kept in view, freezing "obvious" at commitment time so post-hoc mutations are rejected. The question the copied source answers is narrow. Given claims that looked structurally obvious at the time they were written, which reasoning families still survive once every claim is resolved forward and the failures are kept in view? "Obvious" is frozen at commitment time and is never a label applied after the outcome is known. A thesis whose meaning shifts once the result is in is a post-hoc mutation, and the validator rejects it.
\n
Review candidates are not tradable winners
\n
The output is a metadata-only record of verdicts, counts, codes, and refs; source text, market data, and payloads stay out, and the strongest label is review_candidate. The component emits a metadata-only result record: verdicts, counts, error codes, and source refs. Copied source text, market data, and model-output data stay out. The strongest label a surviving pattern can carry is review_candidate, a flag for human review. There is no tradable winner in the output vocabulary.
\n
Purpose
\n
The survivorship failure mode
\n
Collecting only patterns that worked and calling the list a method just measures hindsight, because winners look like signal once the losers are dropped before counting. A loose version of this idea is easy and wrong. Collect patterns that worked, call the list a method, and you have measured hindsight. The failure mode is survivorship: winners alone look like signal because the losers were dropped before anyone counted.
\n
Losers and controls as required evidence
\n
The copied source requires refuted theses and negative controls to flow through the same pipeline as confirmed ones, and the bundle proves this discipline fires on public fixtures. The discipline the copied source enforces is that losers and negative controls are required evidence, not noise. A refuted thesis must flow through the same pipeline as a confirmed one and stay legible as valid evidence. A negative control must be present and must not resolve into a confirmed claim. The bundle's job is to prove that this discipline actually fires on public fixtures, and that it refuses the inputs designed to break it.
\n
How it works
\n
run and its two callbacks
\n
run hands the fixed SPEC to run_crown_jewel_organ with a clean-path and a negative-case callback; the real mechanism lives in the evaluator helpers that follow. The runner is run, which hands the fixed SPEC to the shared run_crown_jewel_organ helper with two callbacks: _structural_evaluator for the clean path and evaluate_negative_case for single negative cases. run_batch8_structural_theses_bundle is the same entry in bundle-input mode. The mechanism lives in the evaluator and its helpers, which execute in this order.
\n
Isolated import in _load_copied_structural_module
\n
_load_copied_structural_module imports the copied source under a private name, isolating sys.path and namespaces so its system and tools packages load from the bundle, not the host._load_copied_structural_module imports the copied structural_theses.py from the exported bundle's source_modules tree. It uses importlib with a private module name, and isolates the import inside _temporary_sys_path and _temporary_import_namespaces so the copied system and tools packages load from the bundle rather than the host. It raises ImportError if the spec or loader is missing.
\n
Probe inputs from _family_inputs
\n
_family_inputs reads the probe manifest and returns the thesis rows, the realized_returns map, and the sampling_frame, filtering out non-mapping rows._family_inputs reads the probe manifest and returns three values: the list of thesis rows, the realized_returns map, and the sampling_frame. Non-mapping rows are filtered out.
\n
Family build and validation in _build_family
\n
_build_family calls the copied build_structural_thesis_family under a purged_holdout split, driving real CP1 admission, CP2 resolution, and replay before validate_structural_thesis_family runs._build_family calls the copied source's build_structural_thesis_family with those inputs, a benchmark, a split policy of purged_holdout, and an embargo. Under the hood the copied source maps each thesis onto the existing forecast-claim shape and drives the real CP1 admission, CP2 resolution, proper-scoring replay, and purged walk-forward replay against the fixture prices. The evaluator then calls the copied validate_structural_thesis_family on the result.
\n
Fixed-id checks in _structural_evaluator
\n
_structural_evaluator checks by fixed id that the winner confirms, the loser refutes as valid evidence, and the control stays a control, blocking on any mismatch._structural_evaluator checks the clean family by fixed id. The winner sob_2015 must resolve claim_confirmed_forward. The loser rmd_2018_loser must resolve claim_refuted_forward with loser_is_valid_evidence true. The control ctrl_2020 must stay a control and must not become claim_confirmed_forward. The surviving pattern must land in family memory as a candidate_set, and authority_boundary.investment_recommendation_authorized must be false. Any mismatch becomes a BATCH8_STRUCTURAL_THESES_REFERENCE_CASE_MISMATCH finding and the status flips to blocked.
\n
The three perturbations in _evaluate_negative_exercises
\n
_evaluate_negative_exercises runs the three perturbations below and records a finding for any refusal that failed to fire; evaluate_negative_case runs them one at a time._evaluate_negative_exercises runs the three perturbations described below. It returns the observed codes plus one component-level finding per refusal that failed to fire. The single-case form evaluate_negative_case runs the same three checks one at a time for the runner's negative-case pass.
\n
Public card from result_card
\n
result_card projects a compact public card carrying an authority_floor with every *_authorized flag false and a body_floor reporting no source bodies.result_card projects the result into a compact public card. It copies the authority_floor from the ceiling (every *_authorized flag stays false) and a body_floor that reports no source bodies in the result record or card.
\n
Function
Role
run / run_batch8_structural_theses_bundle
Entry points; fixture mode and bundle mode
_load_copied_structural_module
Imports the copied source in an isolated namespace
_family_inputs
Reads theses, realized returns, sampling frame from the probe
The three refusals in EXPECTED_NEGATIVE_CASES are the substance of the proof, each forcing one discipline in the copied validator to fire. The three refusals in EXPECTED_NEGATIVE_CASES are the substance of the proof. Each forces one discipline in the copied validator to fire.
\n
Survivor-only family must be rejected
\n
A family built from winner rows alone must trip NO_LOSER_FLOWED_THROUGH, NO_NEGATIVE_CONTROL, and SURVIVORSHIP_SAMPLE, or the check records a survivor-only failure. Survivor-only (structural_theses_survivor_only_rejected). _evaluate_negative_exercises builds a family from the winner rows alone, dropping every failed thesis. The copied validator must raise NO_LOSER_FLOWED_THROUGH, NO_NEGATIVE_CONTROL, and SURVIVORSHIP_SAMPLE. If any of the three is missing, the check records BATCH8_STRUCTURAL_THESES_SURVIVOR_ONLY_REJECTED.
\n
Forward-gate breach must be rejected
\n
Appending a refuted pattern as a review_candidate must raise FORWARD_GATE_BREACH, since only a pattern that survived at-time replay may earn that label. Forward-gate breach (structural_theses_forward_gate_breach_rejected). A refuted pattern is appended to the clean family's forward_lens.candidates with a review_candidate research state. Only a pattern that survived at-time replay may earn a review_candidate, so the copied validator must raise FORWARD_GATE_BREACH, tracked as BATCH8_STRUCTURAL_THESES_FORWARD_GATE_BREACH_REJECTED.
\n
Control leak must be rejected
\n
Mutating every control row to claim_confirmed_forward must raise CONTROL_LEAK, so a negative control can never resolve into a confirmed claim. Control leak (structural_theses_control_leak_rejected). Every control row in a copy of the clean family is mutated to claim_confirmed_forward. The copied validator must raise CONTROL_LEAK, tracked as BATCH8_STRUCTURAL_THESES_CONTROL_LEAK_REJECTED.
\n
A missed refusal blocks the entire run
\n
Any expected refusal that does not fire becomes a blocking finding rather than being counted as a pass. If a refusal does not fire, the run records a blocked finding rather than a pass.
\n
Prior Art Grounding
\n
Factor-family, event-study, and survivorship lineage
\n
The shape borrows from empirical-finance validation: Fama-French factor families, MacKinlay event studies, and survivorship-bias control, applied only to public synthetic rows and copied source. The copied source borrows shape from empirical-finance validation. Fama and French's common risk factors work is the precedent for decomposing structural market claims into named factor families and testable rows. MacKinlay's event-study methodology is the prior pattern for separating an event window, an expected baseline, and abnormal-return evidence. Brown, Goetzmann, Ibbotson, and Ross on survivorship bias motivates the required loser and control cases instead of winner-only replay. The bundle keeps the factor-family, event-window, and bias-control shape and applies it only to public synthetic rows and copied source.
\n
Validation Result record Path
\n
Reader-verifiable commands, run from the microcosm-substrate/ public root:
\n
What the three commands verify
\n
The fixture command writes the family result, the bundle command validates source refs and ceiling fields, and the test checks replay and refusals; a pass says nothing about live markets. The first command writes the bounded family result and sign-off JSON. The bundle command validates the copied source refs, digest anchors, public winner/loser/control cases, and the scope limit fields. The test checks family replay, the three refusals, body exclusion, and the claim boundary. A pass means the copied source ran the public fixture and every refusal fired; it means nothing about live markets.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The evidence supports only that the copied source deterministically resolves the winner, loser, and control family and rejects the three perturbations on public synthetic rows. The strongest claim the evidence supports: the copied structural_theses.py runs the public synthetic winner, loser, and control family deterministically, resolves sob_2015, rmd_2018_loser, and ctrl_2020 with the expected at-time semantics, and rejects the survivor-only, forward-gate-breach, and control-leak perturbations. The proof boundary is deterministic fixture validation over public synthetic rows and copied source refs only. The scope limit holds every *_authorized flag false. This is not financial decisions, an investment recommendation, live-market validation, external model access, portfolio authority, public sharing, or launch. Generated Mermaid and Atlas links are navigation aids and do not raise this ceiling.
\n",
"components": [
{
"anchor": "component-batch8_structural_theses_capsule",
"component_id": "batch8_structural_theses_capsule",
"public_label": "Structural Theses Finance Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 7
},
"link_count": 19,
"links": [
{
"id": "batch8_structural_theses_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Structural Theses Finance Bundle",
"relation": "explains",
"url": "component-batch8-structural-theses-capsule.html"
},
{
"id": "mechanism.batch8_structural_theses_capsule.validates_public_structural_theses_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public structural theses bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch8-structural-theses-capsule-validates-public-structural-theses-capsule"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.prediction_oracle_reconciliation",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Prediction Oracle Reconciliation",
"relation": "depends on",
"url": "paper-module-prediction-oracle-reconciliation.html"
},
{
"id": "paper_module.batch12_prediction_market_board_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 12 Prediction Market Board Bundle",
"relation": "depends on",
"url": "paper-module-batch12-prediction-market-board-capsule.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch12_prediction_market_board_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 12 Prediction Market Board Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[65:paper_module.batch12_prediction_market_board_capsule].depends_on",
"url": "paper-module-batch12-prediction-market-board-capsule.html"
}
],
"paper_module_id": "paper_module.batch8_structural_theses_capsule",
"reader_packet": {
"abstract": "Set 8 Structural Theses Bundle validates public synthetic thesis-family replay without financial decisions, live-market validation, external model access, or portfolio authority.",
"authority_ceiling": "Deterministic fixture evidence over public synthetic thesis rows and copied source refs only; no financial decisions, investment recommendation, live-market validation, external model access, portfolio authority, publishing-scope decision, launch-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch8_structural_theses_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch8-structural-theses-capsule-diagram-1",
"html_url": "docs/paper-module-batch8-structural-theses-capsule.html#paper-module-batch8-structural-theses-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Probe[\"Public probe manifest theses, realized returns, sampling frame\"] --> Load[\"_load_copied_structural_module isolated import of copied source\"] Load --> Build[\"_build_family build_structural_thesis_family CP1 admit, CP2 resolve, proper-scoring + purged replay\"] Build --> Clean[\"_structural_evaluator validate_structural_thesis_family\"] Clean --> Check{\"sob_2015 confirmed, rmd_2018_loser refuted + valid, ctrl_2020 not confirmed?\"} Build --> Neg[\"_evaluate_negative_exercises\"] Neg --> S[\"Survivor-only family\"] Neg --> F[\"Refuted pattern smuggled forward\"] Neg --> C[\"Control mutated to confirmed\"] Check -->|yes| Pass[\"metadata-only pass record\"] Check -->|no| Block[\"Blocked finding\"] S -->|refusal fires| Pass F -->|refusal fires| Pass C -->|refusal fires| Pass",
"source_format": "mermaid",
"source_hash": "sha256:11c79c2828efe240fb311a41def193e52f2179a8c0750d3fe7a6f43df90537a5",
"source_line_end": 80,
"source_line_start": 65
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.prediction_oracle_reconciliation",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Prediction Oracle Reconciliation",
"relation": "depends on",
"url": "paper-module-prediction-oracle-reconciliation.html"
},
{
"id": "paper_module.batch12_prediction_market_board_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 12 Prediction Market Board Bundle",
"relation": "depends on",
"url": "paper-module-batch12-prediction-market-board-capsule.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch8_structural_theses_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Structural Theses Finance Bundle",
"relation": "explains",
"url": "component-batch8-structural-theses-capsule.html"
},
{
"id": "mechanism.batch8_structural_theses_capsule.validates_public_structural_theses_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public structural theses bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch8-structural-theses-capsule-validates-public-structural-theses-capsule"
}
],
"relation": "explains"
},
{
"count": 8,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch8_structural_theses_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 138,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_structural_theses_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "common risk factors",
"section_id": "factor-family-event-study-and-survivorship-lineage",
"url": "https://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html"
},
{
"kind": "external_reference",
"label": "event-study methodology",
"section_id": "factor-family-event-study-and-survivorship-lineage",
"url": "https://www.bu.edu/econ/files/2011/01/MacKinlay-1996-Event-Studies-in-Economics-and-Finance.pdf"
},
{
"kind": "external_reference",
"label": "survivorship bias",
"section_id": "factor-family-event-study-and-survivorship-lineage",
"url": "https://terpconnect.umd.edu/~wermers/ftpsite/FAME/Brown_Goetzmann_Ibbotson_Ross.pdf"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "This bundle imports the finance module tools/finance/structuraltheses.py as a copied body and runs it over public synthetic thesis rows rather than building a new evaluator. This component imports one finance module, tools/finance/structuraltheses.py, as an exact copied source body and runs it over public synthetic thesis rows. It does not build a new evaluator. It drives the copied source's own family builder and validator and records what they return.",
"level": 4,
"order": 1,
"section_id": "structural-theses-run-through-the-copied-evaluator",
"source_line_end": 6,
"source_line_start": 3,
"title": "Structural theses run through the copied evaluator"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The copied source asks which reasoning families survive once every claim is resolved forward with failures kept in view, freezing \"obvious\" at commitment time so post-hoc mutations are rejected. The question the copied source answers is narrow. Given claims that looked structurally obvious at the time they were written, which reasoning families still survive once every claim is resolved forward and the failures are kept in view? \"Obvious\" is frozen at commitment time and is never a label applied after the outcome is known. A thesis whose meaning shifts once the result is in is a post-hoc mutation, and the validator rejects it.",
"level": 4,
"order": 2,
"section_id": "losers-remain-visible-beside-surviving-families",
"source_line_end": 10,
"source_line_start": 7,
"title": "Losers remain visible beside surviving families"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The output is a metadata-only record of verdicts, counts, codes, and refs; source text, market data, and payloads stay out, and the strongest label is reviewcandidate. The component emits a metadata-only result record: verdicts, counts, error codes, and source refs. Copied source text, market data, and model-output data stay out. The strongest label a surviving pattern can carry is reviewcandidate, a flag for human review. There is no tradable winner in the output vocabulary.",
"level": 4,
"order": 3,
"section_id": "review-candidates-are-not-tradable-winners",
"source_line_end": 14,
"source_line_start": 11,
"title": "Review candidates are not tradable winners"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Collecting only patterns that worked and calling the list a method just measures hindsight, because winners look like signal once the losers are dropped before counting. A loose version of this idea is easy and wrong. Collect patterns that worked, call the list a method, and you have measured hindsight. The failure mode is survivorship: winners alone look like signal because the losers were dropped before anyone counted.",
"level": 4,
"order": 5,
"section_id": "the-survivorship-failure-mode",
"source_line_end": 20,
"source_line_start": 17,
"title": "The survivorship failure mode"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The copied source requires refuted theses and negative controls to flow through the same pipeline as confirmed ones, and the bundle proves this discipline fires on public fixtures. The discipline the copied source enforces is that losers and negative controls are required evidence, not noise. A refuted thesis must flow through the same pipeline as a confirmed one and stay legible as valid evidence. A negative control must be present and must not resolve into a confirmed claim. The bundle's job is to prove that this discipline actually fires on public fixtures, and that it refuses the inputs designed to break it.",
"level": 4,
"order": 6,
"section_id": "losers-and-controls-as-required-evidence",
"source_line_end": 24,
"source_line_start": 21,
"title": "Losers and controls as required evidence"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run hands the fixed SPEC to runcrownjewelorgan with a clean-path and a negative-case callback; the real mechanism lives in the evaluator helpers that follow. The runner is run, which hands the fixed SPEC to the shared runcrownjewelorgan helper with two callbacks: structuralevaluator for the clean path and evaluatenegativecase for single negative cases. runbatch8structuralthesesbundle is the same entry in bundle-input mode. The mechanism lives in the evaluator and its helpers, which execute in this order.",
"level": 4,
"order": 8,
"section_id": "run-and-its-two-callbacks",
"source_line_end": 30,
"source_line_start": 27,
"title": "`run` and its two callbacks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "loadcopiedstructuralmodule imports the copied source under a private name, isolating sys.path and namespaces so its system and tools packages load from the bundle, not the host. loadcopiedstructuralmodule imports the copied structuraltheses.py from the exported bundle's sourcemodules tree. It uses importlib with a private module name, and isolates the import inside temporarysyspath and temporaryimport_namespaces so the copied system and tools packages load from the bundle rather than the host. It raises ImportError if the spec or loader is missing.",
"level": 4,
"order": 9,
"section_id": "isolated-import-in-load-copied-structural-module",
"source_line_end": 34,
"source_line_start": 31,
"title": "Isolated import in `_load_copied_structural_module`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "familyinputs reads the probe manifest and returns the thesis rows, the realizedreturns map, and the samplingframe, filtering out non-mapping rows. familyinputs reads the probe manifest and returns three values: the list of thesis rows, the realizedreturns map, and the sampling_frame. Non-mapping rows are filtered out.",
"level": 4,
"order": 10,
"section_id": "probe-inputs-from-family-inputs",
"source_line_end": 38,
"source_line_start": 35,
"title": "Probe inputs from `_family_inputs`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildfamily calls the copied buildstructuralthesisfamily under a purgedholdout split, driving real CP1 admission, CP2 resolution, and replay before validatestructuralthesisfamily runs. buildfamily calls the copied source's buildstructuralthesisfamily with those inputs, a benchmark, a split policy of purgedholdout, and an embargo. Under the hood the copied source maps each thesis onto the existing forecast-claim shape and drives the real CP1 admission, CP2 resolution, proper-scoring replay, and purged walk-forward replay against the fixture prices. The evaluator then calls the copied validatestructuralthesis_family on the result.",
"level": 4,
"order": 11,
"section_id": "family-build-and-validation-in-build-family",
"source_line_end": 42,
"source_line_start": 39,
"title": "Family build and validation in `_build_family`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "structuralevaluator checks by fixed id that the winner confirms, the loser refutes as valid evidence, and the control stays a control, blocking on any mismatch. structuralevaluator checks the clean family by fixed id. The winner sob2015 must resolve claimconfirmedforward. The loser rmd2018loser must resolve claimrefutedforward with loserisvalidevidence true. The control ctrl2020 must stay a control and must not become claimconfirmedforward. The surviving pattern must land in family memory as a candidateset, and authorityboundary.investmentrecommendationauthorized must be false. Any mismatch becomes a BATCH8STRUCTURALTHESESREFERENCECASE_MISMATCH finding and the status flips to blocked.",
"level": 4,
"order": 12,
"section_id": "fixed-id-checks-in-structural-evaluator",
"source_line_end": 46,
"source_line_start": 43,
"title": "Fixed-id checks in `_structural_evaluator`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluatenegativeexercises runs the three perturbations below and records a finding for any refusal that failed to fire; evaluatenegativecase runs them one at a time. evaluatenegativeexercises runs the three perturbations described below. It returns the observed codes plus one component-level finding per refusal that failed to fire. The single-case form evaluatenegative_case runs the same three checks one at a time for the runner's negative-case pass.",
"level": 4,
"order": 13,
"section_id": "the-three-perturbations-in-evaluate-negative-exercises",
"source_line_end": 50,
"source_line_start": 47,
"title": "The three perturbations in `_evaluate_negative_exercises`"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "resultcard projects a compact public card carrying an authorityfloor with every authorized flag false and a bodyfloor reporting no source bodies. resultcard projects the result into a compact public card. It copies the authorityfloor from the ceiling (every authorized flag stays false) and a bodyfloor that reports no source bodies in the result record or card.",
"level": 4,
"order": 14,
"section_id": "public-card-from-result-card",
"source_line_end": 81,
"source_line_start": 51,
"title": "Public card from `result_card`"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "negative-cases",
"source_line_end": 83,
"source_line_start": 82,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The three refusals in EXPECTEDNEGATIVECASES are the substance of the proof, each forcing one discipline in the copied validator to fire. The three refusals in EXPECTEDNEGATIVECASES are the substance of the proof. Each forces one discipline in the copied validator to fire.",
"level": 4,
"order": 16,
"section_id": "three-refusals-carry-the-proof",
"source_line_end": 87,
"source_line_start": 84,
"title": "Three refusals carry the proof"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A family built from winner rows alone must trip NOLOSERFLOWEDTHROUGH, NONEGATIVECONTROL, and SURVIVORSHIPSAMPLE, or the check records a survivor-only failure. Survivor-only (structuralthesessurvivoronlyrejected). evaluatenegativeexercises builds a family from the winner rows alone, dropping every failed thesis. The copied validator must raise NOLOSERFLOWEDTHROUGH, NONEGATIVECONTROL, and SURVIVORSHIPSAMPLE. If any of the three is missing, the check records BATCH8STRUCTURALTHESESSURVIVORONLYREJECTED.",
"level": 4,
"order": 17,
"section_id": "survivor-only-family-must-be-rejected",
"source_line_end": 91,
"source_line_start": 88,
"title": "Survivor-only family must be rejected"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Appending a refuted pattern as a reviewcandidate must raise FORWARDGATEBREACH, since only a pattern that survived at-time replay may earn that label. Forward-gate breach (structuralthesesforwardgatebreachrejected). A refuted pattern is appended to the clean family's forwardlens.candidates with a reviewcandidate research state. Only a pattern that survived at-time replay may earn a reviewcandidate, so the copied validator must raise FORWARDGATEBREACH, tracked as BATCH8STRUCTURALTHESESFORWARDGATEBREACH_REJECTED.",
"level": 4,
"order": 18,
"section_id": "forward-gate-breach-must-be-rejected",
"source_line_end": 95,
"source_line_start": 92,
"title": "Forward-gate breach must be rejected"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Mutating every control row to claimconfirmedforward must raise CONTROLLEAK, so a negative control can never resolve into a confirmed claim. Control leak (structuralthesescontrolleakrejected). Every control row in a copy of the clean family is mutated to claimconfirmedforward. The copied validator must raise CONTROLLEAK, tracked as BATCH8STRUCTURALTHESESCONTROLLEAK_REJECTED.",
"level": 4,
"order": 19,
"section_id": "control-leak-must-be-rejected",
"source_line_end": 99,
"source_line_start": 96,
"title": "Control leak must be rejected"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Any expected refusal that does not fire becomes a blocking finding rather than being counted as a pass. If a refusal does not fire, the run records a blocked finding rather than a pass.",
"level": 4,
"order": 20,
"section_id": "a-missed-refusal-blocks-the-entire-run",
"source_line_end": 104,
"source_line_start": 100,
"title": "A missed refusal blocks the entire run"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "prior-art-grounding",
"source_line_end": 106,
"source_line_start": 105,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The shape borrows from empirical-finance validation: Fama-French factor families, MacKinlay event studies, and survivorship-bias control, applied only to public synthetic rows and copied source. The copied source borrows shape from empirical-finance validation. Fama and French's common risk factors work is the precedent for decomposing structural market claims into named factor families and testable rows. MacKinlay's event-study methodology is the prior pattern for separating an event window, an expected baseline, and abnormal-return evidence. Brown, Goetzmann, Ibbotson, and Ross on survivorship bias motivates the required loser and control cases instead of winner-only replay. The bundle ke\u2026",
"level": 4,
"order": 22,
"section_id": "factor-family-event-study-and-survivorship-lineage",
"source_line_end": 110,
"source_line_start": 107,
"title": "Factor-family, event-study, and survivorship lineage"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Reader-verifiable commands, run from the microcosm-substrate/ public root:",
"level": 2,
"order": 23,
"section_id": "validation-receipt-path",
"source_line_end": 127,
"source_line_start": 111,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The fixture command writes the family result, the bundle command validates source refs and ceiling fields, and the test checks replay and refusals; a pass says nothing about live markets. The first command writes the bounded family result and sign-off JSON. The bundle command validates the copied source refs, digest anchors, public winner/loser/control cases, and the scope limit fields. The test checks family replay, the three refusals, body exclusion, and the claim boundary. A pass means the copied source ran the public fixture and every refusal fired; it means nothing about live markets.",
"level": 4,
"order": 24,
"section_id": "what-the-three-commands-verify",
"source_line_end": 131,
"source_line_start": 128,
"title": "What the three commands verify"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 25,
"section_id": "claim-ceiling",
"source_line_end": 133,
"source_line_start": 132,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The evidence supports only that the copied source deterministically resolves the winner, loser, and control family and rejects the three perturbations on public synthetic rows. The strongest claim the evidence supports: the copied structuraltheses.py runs the public synthetic winner, loser, and control family deterministically, resolves sob2015, rmd2018loser, and ctrl2020 with the expected at-time semantics, and rejects the survivor-only, forward-gate-breach, and control-leak perturbations. The proof boundary is deterministic fixture validation over public synthetic rows and copied source refs only. The scope limit holds every authorized flag false. This is not financial decisions, an inves\u2026",
"level": 4,
"order": 26,
"section_id": "the-strongest-supported-claim",
"source_line_end": 137,
"source_line_start": 134,
"title": "The strongest supported claim"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch8_structural_theses_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 8 Structural Theses Bundle validates public synthetic thesis-family replay without financial decisions, live-market validation, external model access, or portfolio authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch8_structural_theses_capsule.py",
"path": "src/microcosm_core/organs/batch8_structural_theses_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for structural-thesis family fixtures, source digest checks, negative exercises, and scope limits.",
"symbols": [
"_load_copied_structural_module",
"_family_inputs",
"_build_family",
"_evaluate_negative_exercises",
"_structural_evaluator",
"run",
"run_batch8_structural_theses_bundle",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch8_structural_theses_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_structural_theses_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_structural_theses_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 8 Structural Theses Bundle imports tools/finance/structural_theses.py as exact copied source and exercises CP1/CP2 thesis-family validation over public synthetic winner, loser, and control rows. It checks source digest parity, anchors, public family evidence, loser evidence, negative controls, survivor-only rejection, forward-gate-breach rejection, control-leak rejection, runtime verdicts, and scope limits while excluding financial decisions, investment recommendations, live market data, external model access, portfolio action, launch, public sharing, and whole-system correctness.",
"summary_line": "Set 8 Structural Theses Bundle validates public synthetic thesis-family replay without financial decisions, live-market validation, external model access, or portfolio authority.",
"title": "Set 8 Structural Theses Bundle"
},
{
"anchor": "paper-module-batch8-tools-tail-primitives-capsule",
"authority_ceiling": "Deterministic public tools-tail fixture evidence and copied source refs only; no oracle truth, prediction correctness, semantic edit correctness, Lab execution authority, live bridge authority, repository mutation authority, external model access, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
Four copied tools run against fixed expectations
\n
This bundle copies four small tools into the public system as their real bodies and runs each on a fixed input against a hard-coded expected value. This component copies four small tools into the public system as their real source bodies, then runs each one on a fixed input and checks the output against a hard-coded expected value. The four are an observer set diff, a JSON-patch interpreter, a ledger-id hash, and a shadow envelope parser.
\n
Execution distinguishes working copies from stubs
\n
It answers one question about copied code: does the copy still behave like the original, or has it quietly become a stub that only looks right? It exists to answer one question about copied code: does the copy still behave the way the original did, or has it quietly become a stub that only looks right? The check runs the copied body in process and compares the result field by field. If the copy were hollow, the comparison fails instead of passing with a green tick.
\n
metadata-only records retain values and refusals
\n
It emits a metadata-only record of which exercises passed, their small computed values, the rejected negative cases, and a fixed scope block, never the copied source text. It emits a metadata-only result record: which exercises passed, the small computed values they produced, the negative cases they rejected, and a fixed scope block. It does not copy the source text of the tools into that record.
\n
Purpose
\n
A digest match does not establish execution
\n
A digest match proves bytes were copied but not that they still run, so a loose hash check would report success even when an import broke. A digest match proves that bytes were copied. It does not establish the copied bytes still run. When a tool moves from the private system into the public clone, a loose check would hash the file, see the hash match, and report success even if an import broke or a function returned the wrong shape. That is the failure this component is built to avoid.
\n
Behavioural evidence, not a correctness claim
\n
Rather than re-implementing the tools, it loads each copied module and calls the real functions for behavioural evidence, leaving open whether the answers mean anything. So it does not re-describe or re-implement the four tools. It loads each copied module from the manifest and calls the real functions, then asserts exact results. The evidence is behavioural: the code executed. It is not a claim that the answers mean anything about the world. Diffing two sets of observer rows is set arithmetic. Applying a JSON patch is interpreting an edit script. The gap between "the mechanism runs as copied" and "the answer is correct" stays open on purpose.
\n
How it works
\n
evaluate entry point and _load_copied_module
\n
evaluate reads the manifest, checks mechanism rows, runs the four live exercises, and passes only when findings is empty; _load_copied_module imports each copied body by spec.evaluate is the entry point. It reads the probe manifest, checks each declared mechanism row through _mechanism_status, runs the four live exercises in RUNTIME_EXERCISES, and folds any missing mechanism, failed exercise, or blocked anchor into a findings list. The result is pass only when that list is empty. _load_copied_module does the loading: it resolves the target path from the source manifest, imports the file by spec, and runs it under a temporary sys.path so the copied module can find its own imports.
\n
Function
What it runs
_exercise_observer_diff
diff_evidence, diff_predictions on keyed rows
_exercise_version_committer
_apply_op edit ops plus a refused scalar traversal
_exercise_ledger_id
generate_ledger_id, normalize_lane on lane aliases
_exercise_shadow
the copied run over a written envelope directory
evaluate_negative_case
one rejection path per tool
\n
Observer diff in _exercise_observer_diff
\n
_exercise_observer_diff checks the overlap, missed, and extra id sets, drops a malformed row without crashing, and confirms skipped never leaks into the serialised result._exercise_observer_diff feeds three evidence rows (E1 lab, a row with no ledger_id, E3 lab) against two oracle rows (E1, E2). It checks overlap_ledger_ids == ["E1"], missed_ledger_ids == ["E2"], and extra_ledger_ids == ["E3"]. The malformed row is dropped rather than crashing the diff, and the exercise asserts the string skipped never appears in the serialised result, so a parse gap cannot leak through as silent data. Predictions split into matching == ["A"], divergent == ["B"], and missing_targets == ["C"].
\n
Patch ops and refusal in _exercise_version_committer
\n
_exercise_version_committer applies four _apply_op edits to a known document, then confirms traversing into a scalar raises VersionCommitterError rather than making a partial write._exercise_version_committer applies four ops with _apply_op (set, merge, append, set) and checks the final document equals {"profile": {"name": "Microcosm", "meta": {"kind": "fixture"}}, "events": [{"id": "E1"}]}. It then tries to traverse into a scalar, set on /profile/name where profile is the string "scalar", and confirms this raises VersionCommitterError. The interesting property is the refusal: a malformed path is a controlled error, not a partial write.
\n
Identity stability in _exercise_ledger_id
\n
_exercise_ledger_id confirms a lane alias and its canonical name hash to the same P_ id, an unknown lane falls back to X_, and a missing identity field raises ValueError._exercise_ledger_id checks identity stability. normalize_lane("poly") resolves to POLYMARKET, so generate_ledger_id("poly", record) and generate_ledger_id("POLYMARKET", record) return the same id, which starts with P_. An unknown lane falls back to an X_ prefix. A record missing the identity field its lane requires raises ValueError instead of hashing a blank.
\n
Envelope parsing in _exercise_shadow
\n
_exercise_shadow runs the copied parser over a written envelope directory and checks that well-formed rows parse while the malformed tuple is counted once as a comma_arity gap._exercise_shadow writes a small envelope directory into a temporary path: miner tuples with one malformed entry, a spine line, and a prediction row. It calls the copied run and reads parse_stats. It checks hard_failure is false, that the well-formed miner, spine, and prediction rows parsed, and that the malformed tuple was counted once under comma_arity. The parser reports its own coverage hole rather than swallowing it.
\nDiagram of the mechanism (6 steps).Diagram source & refs
flowchart TD manifest["Probe manifest + source module manifest"] load["_load_copied_module import copied body by spec"] exercises["Four live exercises diff | patch | ledger-id | shadow"] negatives["evaluate_negative_case one rejection per tool"] evaluate["evaluate collect findings"] record["metadata-only result record pass only if findings empty"] manifest --> load load --> exercises load --> negatives exercises --> evaluate negatives --> evaluate evaluate --> record
\n
Negative cases
\n
One rejecting path per tool
\n
For each tool the component runs a rejecting path from EXPECTED_NEGATIVE_CASES, passing only when the tool refuses the bad input with the expected code. For each tool the component also runs the rejecting path, declared in EXPECTED_NEGATIVE_CASES and evaluated by evaluate_negative_case. A case passes only when the tool refuses the bad input with the expected code.
\n
Malformed observer row dropped
\n
A row with no ledger_id is dropped and the good oracle id lands in missed_ledger_ids, raising BATCH8_OBSERVER_DIFF_MALFORMED_ROW_SKIPPED.observer_diff_malformed_key_skipped: a row with no ledger_id is dropped, the good oracle id lands in missed_ledger_ids, code BATCH8_OBSERVER_DIFF_MALFORMED_ROW_SKIPPED.
\n
Scalar traversal refused
\n
Setting a path inside a scalar value raises VersionCommitterError rather than writing, recorded as BATCH8_VERSION_COMMITTER_SCALAR_TRAVERSAL_REFUSED.version_committer_scalar_traversal_refused: setting a path inside a scalar raises VersionCommitterError, code BATCH8_VERSION_COMMITTER_SCALAR_TRAVERSAL_REFUSED.
\n
Missing identity field refused
\n
A record with no identity field raises ValueError instead of hashing a blank, recorded as BATCH8_LEDGER_ID_MISSING_IDENTITY_REFUSED.ledger_id_missing_identity_refused: a record with no identity field raises ValueError, code BATCH8_LEDGER_ID_MISSING_IDENTITY_REFUSED.
\n
Malformed tuple counted as gap
\n
The malformed tuple is counted as a comma_arity coverage gap rather than swallowed, recorded as BATCH8_SHADOW_DSL_MALFORMED_TUPLE_COVERAGE_GAP.shadow_dsl_malformed_tuple_coverage_gap: the malformed tuple is counted as a comma_arity gap, code BATCH8_SHADOW_DSL_MALFORMED_TUPLE_COVERAGE_GAP.
\n
Prior Art Grounding
\n
The four tools follow standard patterns. Useful anchors:
\n
IETF RFC 6902, which defines JSON Patch operations such as add, remove, replace, move, copy, and test.
IETF RFC 9162, where Certificate Transparency uses an append-only Merkle tree as an auditable log pattern.
W3C PROV, for representing the provenance of derived artifacts and their generating activities.
\n
Borrowed shapes, no correctness claim
\n
The patch-op, identity-hash, append-only-log, and provenance shapes are borrowed, but the component runs only deterministic fixture exercises and claims nothing about correctness. The patch-operation, identity-hash, append-only-log, and provenance shapes are borrowed. This component keeps to deterministic fixture exercises over those shapes and claims nothing about correctness of the edits, ids, or predictions.
\n
Validation Result record Path
\n
Reader-verifiable commands, run from the microcosm-substrate/ public root:
\n
What the three commands verify
\n
The fixture command writes the result, the bundle command validates copied sources and digests, and the test checks mechanics and negative cases; a pass means the copied bodies produced exact values. The fixture command writes the bounded result and sign-off JSON. The bundle command validates the copied sources, manifest digests, the four exercises, the body-exclusion posture, and the scope fields. The focused test checks fixture mechanics, bundle validation, and the negative cases. A pass means the copied bodies ran and produced the exact expected values on the public fixture.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The evidence supports only that the four copied tools import, run, produce exact expected values, and reject the four bad inputs, with no body text in the record. The strongest claim the evidence supports: the four copied tools import and run in the public clone, produce the exact expected values on a fixed input, and reject the four declared bad inputs. The result record carries source refs, digests, counts, and outcomes, never copied body text or private state.
\n
What the scope limit refuses
\n
Running the mechanisms is not a claim they are correct: AUTHORITY_CEILING refuses oracle truth, prediction correctness, live execution, mutation, and launch. The proof boundary stops there. Running set arithmetic is not a claim that the observer rows are correct. Applying an edit script is not a claim that the edit is right. A stable hash is not oracle truth. The scope limit refuses oracle truth, prediction correctness, semantic edit correctness, live bridge or Lab execution, external model access, repository or source-file changes, publishing-scope decision, launch-scope decision, and whole-system correctness. These values are fixed in AUTHORITY_CEILING and copied into every result card by result_card.
\n",
"components": [
{
"anchor": "component-batch8_tools_tail_primitives_capsule",
"component_id": "batch8_tools_tail_primitives_capsule",
"public_label": "Tools-Tail Primitives Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 7
},
"link_count": 19,
"links": [
{
"id": "batch8_tools_tail_primitives_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Tools-Tail Primitives Bundle",
"relation": "explains",
"url": "component-batch8-tools-tail-primitives-capsule.html"
},
{
"id": "mechanism.batch8_tools_tail_primitives_capsule.validates_public_tools_tail_primitives_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public tools tail primitives bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch8-tools-tail-primitives-capsule-validates-public-tools-tail-primitives-capsule"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.batch7_oracle_sibling_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Oracle Sibling Bundle",
"relation": "depends on",
"url": "paper-module-batch7-oracle-sibling-capsule.html"
},
{
"id": "paper_module.batch7_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Source Engines Bundle",
"relation": "depends on",
"url": "paper-module-batch7-macro-engines-capsule.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.batch8_tools_tail_primitives_capsule",
"reader_packet": {
"abstract": "Set 8 Tools-Tail Primitives Bundle validates four public tools-tail primitive exercises without Oracle truth, external model access, live bridge work, or repository mutation authority.",
"authority_ceiling": "Deterministic public tools-tail fixture evidence and copied source refs only; no oracle truth, prediction correctness, semantic edit correctness, Lab execution authority, live bridge authority, repository mutation authority, external model access, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch8_tools_tail_primitives_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch8-tools-tail-primitives-capsule-diagram-1",
"html_url": "docs/paper-module-batch8-tools-tail-primitives-capsule.html#paper-module-batch8-tools-tail-primitives-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD manifest[\"Probe manifest + source module manifest\"] load[\"_load_copied_module import copied body by spec\"] exercises[\"Four live exercises diff | patch | ledger-id | shadow\"] negatives[\"evaluate_negative_case one rejection per tool\"] evaluate[\"evaluate collect findings\"] record[\"metadata-only result record pass only if findings empty\"] manifest --> load load --> exercises load --> negatives exercises --> evaluate negatives --> evaluate evaluate --> record",
"source_format": "mermaid",
"source_hash": "sha256:47603dd835f7d44382aab74de5a671213918451ec15d90b6e0bba49beb23e413",
"source_line_end": 135,
"source_line_start": 121
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-9",
"kind": "axiom",
"kind_label": "axiom",
"label": "Compensable transactional effects",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-9"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.batch7_oracle_sibling_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Oracle Sibling Bundle",
"relation": "depends on",
"url": "paper-module-batch7-oracle-sibling-capsule.html"
},
{
"id": "paper_module.batch7_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Source Engines Bundle",
"relation": "depends on",
"url": "paper-module-batch7-macro-engines-capsule.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch8_tools_tail_primitives_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Tools-Tail Primitives Bundle",
"relation": "explains",
"url": "component-batch8-tools-tail-primitives-capsule.html"
},
{
"id": "mechanism.batch8_tools_tail_primitives_capsule.validates_public_tools_tail_primitives_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public tools tail primitives bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch8-tools-tail-primitives-capsule-validates-public-tools-tail-primitives-capsule"
}
],
"relation": "explains"
},
{
"count": 8,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch8_tools_tail_primitives_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 245,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_tools_tail_primitives_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "RFC 6902",
"section_id": "prior-art-grounding",
"url": "https://datatracker.ietf.org/doc/html/rfc6902"
},
{
"kind": "external_reference",
"label": "RFC 9162",
"section_id": "prior-art-grounding",
"url": "https://www.rfc-editor.org/rfc/rfc9162"
},
{
"kind": "external_reference",
"label": "PROV",
"section_id": "prior-art-grounding",
"url": "https://www.w3.org/TR/prov-overview/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This bundle copies four small tools into the public system as their real bodies and runs each on a fixed input against a hard-coded expected value.** This component copies four small tools into the public system as their real source bodies, then runs each one on a fixed input and checks the output against a hard-coded expected value. The four are an observer set diff, a JSON-patch interpreter, a ledger-id hash, and a shadow envelope parser.",
"level": 4,
"order": 1,
"section_id": "four-copied-tools-run-against-fixed-expectations",
"source_line_end": 11,
"source_line_start": 3,
"title": "Four copied tools run against fixed expectations"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It answers one question about copied code: does the copy still behave like the original, or has it quietly become a stub that only looks right?** It exists to answer one question about copied code: does the copy still behave the way the original did, or has it quietly become a stub that only looks right? The check runs the copied body in process and compares the result field by field. If the copy were hollow, the comparison fails instead of passing with a green tick.",
"level": 4,
"order": 2,
"section_id": "execution-distinguishes-working-copies-from-stubs",
"source_line_end": 20,
"source_line_start": 12,
"title": "Execution distinguishes working copies from stubs"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It emits a metadata-only record of which exercises passed, their small computed values, the rejected negative cases, and a fixed scope block, never the copied source text.** It emits a metadata-only result record: which exercises passed, the small computed values they produced, the negative cases they rejected, and a fixed scope block. It does not copy the source text of the tools into that record.",
"level": 4,
"order": 3,
"section_id": "body-free-records-retain-values-and-refusals",
"source_line_end": 28,
"source_line_start": 21,
"title": "metadata-only records retain values and refusals"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 30,
"source_line_start": 29,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A digest match proves bytes were copied but not that they still run, so a loose hash check would report success even when an import broke.** A digest match proves that bytes were copied. It does not establish the copied bytes still run. When a tool moves from the private system into the public clone, a loose check would hash the file, see the hash match, and report success even if an import broke or a function returned the wrong shape. That is the failure this component is built to avoid.",
"level": 4,
"order": 5,
"section_id": "a-digest-match-does-not-prove-execution",
"source_line_end": 39,
"source_line_start": 31,
"title": "A digest match does not establish execution"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Rather than re-implementing the tools, it loads each copied module and calls the real functions for behavioural evidence, leaving open whether the answers mean anything.** So it does not re-describe or re-implement the four tools. It loads each copied module from the manifest and calls the real functions, then asserts exact results. The evidence is behavioural: the code executed. It is not a claim that the answers mean anything about the world. Diffing two sets of observer rows is set arithmetic. Applying a JSON patch is interpreting an edit script. The gap between \"the mechanism runs as copied\" and \"the answer is correct\" stays open on purpose.",
"level": 4,
"order": 6,
"section_id": "behavioural-evidence-not-a-correctness-claim",
"source_line_end": 51,
"source_line_start": 40,
"title": "Behavioural evidence, not a correctness claim"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 53,
"source_line_start": 52,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "**evaluate reads the manifest, checks mechanism rows, runs the four live exercises, and passes only when findings is empty; loadcopied_module imports each copied body by spec.** evaluate is the entry point. It reads the probe manifest, checks each declared mechanism row through mechanismstatus, runs the four live exercises in RUNTIME_EXERCISES, and folds any missing mechanism, failed exercise, or blocked anchor into a findings list. The result is pass only when that list is empty. loadcopied_module does the loading: it resolves the target path from the source manifest, imports the file by spec, and runs it under a temporary sys.path so the copied module can find its own imports.",
"level": 4,
"order": 8,
"section_id": "evaluate-entry-point-and-load-copied-module",
"source_line_end": 73,
"source_line_start": 54,
"title": "`evaluate` entry point and `_load_copied_module`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "exerciseobserver_diff checks the overlap, missed, and extra id sets, drops a malformed row without crashing, and confirms skipped never leaks into the serialised result. exerciseobserverdiff feeds three evidence rows (E1 lab, a row with no ledger_id, E3 lab) against two oracle rows (E1, E2). It checks overlapledgerids == [\"E1\"], missedledgerids == [\"E2\"], and extraledgerids == [\"E3\"]. The malformed row is dropped rather than crashing the diff, and the exercise asserts the string skipped never appears in the serialised result, so a parse gap cannot leak through as silent data. Predictions split into matching == [\"A\"], divergent == [\"B\"], and missing_targets == [\"C\"].",
"level": 4,
"order": 9,
"section_id": "observer-diff-in-exercise-observer-diff",
"source_line_end": 86,
"source_line_start": 74,
"title": "Observer diff in `_exercise_observer_diff`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "exerciseversioncommitter applies four apply_op edits to a known document, then confirms traversing into a scalar raises VersionCommitterError rather than making a partial write. exerciseversioncommitter applies four ops with applyop (set, merge, append, set) and checks the final document equals `{\"profile\": {\"name\": \"Microcosm\", \"meta\": {\"kind\": \"fixture\"}}, \"events\": [{\"id\": \"E1\"}]}. It then tries to traverse into a scalar, set` on /profile/name where profile is the string \"scalar\", and confirms this raises VersionCommitterError. The interesting property is the refusal: a malformed path is a controlled error, not a partial write.",
"level": 4,
"order": 10,
"section_id": "patch-ops-and-refusal-in-exercise-version-committer",
"source_line_end": 98,
"source_line_start": 87,
"title": "Patch ops and refusal in `_exercise_version_committer`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "exerciseledger_id confirms a lane alias and its canonical name hash to the same P id, an unknown lane falls back to X, and a missing identity field raises ValueError. exerciseledgerid checks identity stability. normalize_lane(\"poly\") resolves to POLYMARKET, so generateledgerid(\"poly\", record) and generateledgerid(\"POLYMARKET\", record) return the same id, which starts with P. An unknown lane falls back to an X prefix. A record missing the identity field its lane requires raises ValueError instead of hashing a blank.",
"level": 4,
"order": 11,
"section_id": "identity-stability-in-exercise-ledger-id",
"source_line_end": 109,
"source_line_start": 99,
"title": "Identity stability in `_exercise_ledger_id`"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "exerciseshadow runs the copied parser over a written envelope directory and checks that well-formed rows parse while the malformed tuple is counted once as a commaarity gap. exerciseshadow writes a small envelope directory into a temporary path: miner tuples with one malformed entry, a spine line, and a prediction row. It calls the copied run and reads parsestats. It checks hardfailure is false, that the well-formed miner, spine, and prediction rows parsed, and that the malformed tuple was counted once under comma_arity. The parser reports its own coverage hole rather than swallowing it.",
"level": 4,
"order": 12,
"section_id": "envelope-parsing-in-exercise-shadow",
"source_line_end": 136,
"source_line_start": 110,
"title": "Envelope parsing in `_exercise_shadow`"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 138,
"source_line_start": 137,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**For each tool the component runs a rejecting path from EXPECTEDNEGATIVECASES, passing only when the tool refuses the bad input with the expected code.** For each tool the component also runs the rejecting path, declared in EXPECTEDNEGATIVECASES and evaluated by evaluatenegativecase. A case passes only when the tool refuses the bad input with the expected code.",
"level": 4,
"order": 14,
"section_id": "one-rejecting-path-per-tool",
"source_line_end": 146,
"source_line_start": 139,
"title": "One rejecting path per tool"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A row with no ledgerid is dropped and the good oracle id lands in missedledgerids, raising BATCH8OBSERVERDIFFMALFORMEDROWSKIPPED. observerdiffmalformedkeyskipped: a row with no ledger_id is dropped, the good oracle id lands in missedledgerids, code BATCH8OBSERVERDIFFMALFORMEDROW_SKIPPED.",
"level": 4,
"order": 15,
"section_id": "malformed-observer-row-dropped",
"source_line_end": 153,
"source_line_start": 147,
"title": "Malformed observer row dropped"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Setting a path inside a scalar value raises VersionCommitterError rather than writing, recorded as BATCH8VERSIONCOMMITTERSCALARTRAVERSALREFUSED. versioncommitterscalartraversalrefused: setting a path inside a scalar raises VersionCommitterError, code BATCH8VERSIONCOMMITTERSCALARTRAVERSAL_REFUSED.",
"level": 4,
"order": 16,
"section_id": "scalar-traversal-refused",
"source_line_end": 160,
"source_line_start": 154,
"title": "Scalar traversal refused"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A record with no identity field raises ValueError instead of hashing a blank, recorded as BATCH8LEDGERIDMISSINGIDENTITYREFUSED. ledgeridmissingidentityrefused: a record with no identity field raises ValueError, code BATCH8LEDGERIDMISSINGIDENTITY_REFUSED.",
"level": 4,
"order": 17,
"section_id": "missing-identity-field-refused",
"source_line_end": 166,
"source_line_start": 161,
"title": "Missing identity field refused"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The malformed tuple is counted as a commaarity coverage gap rather than swallowed, recorded as BATCH8SHADOWDSLMALFORMEDTUPLECOVERAGEGAP. shadowdslmalformedtuplecoveragegap: the malformed tuple is counted as a commaarity gap, code BATCH8SHADOWDSLMALFORMEDTUPLECOVERAGE_GAP.",
"level": 4,
"order": 18,
"section_id": "malformed-tuple-counted-as-gap",
"source_line_end": 172,
"source_line_start": 167,
"title": "Malformed tuple counted as gap"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The four tools follow standard patterns. Useful anchors: IETF RFC 6902, which defines JSON Patch operations such as add, remove, replace, move, copy, and test. IETF RFC 9162, where Certificate Transparency uses an append-only Merkle tree as an auditable log pattern. W3C PROV, for representing the provenance of derived artifacts and their generating activities.",
"level": 2,
"order": 19,
"section_id": "prior-art-grounding",
"source_line_end": 184,
"source_line_start": 173,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The patch-op, identity-hash, append-only-log, and provenance shapes are borrowed, but the component runs only deterministic fixture exercises and claims nothing about correctness.** The patch-operation, identity-hash, append-only-log, and provenance shapes are borrowed. This component keeps to deterministic fixture exercises over those shapes and claims nothing about correctness of the edits, ids, or predictions.",
"level": 4,
"order": 20,
"section_id": "borrowed-shapes-no-correctness-claim",
"source_line_end": 192,
"source_line_start": 185,
"title": "Borrowed shapes, no correctness claim"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Reader-verifiable commands, run from the microcosm-substrate/ public root:",
"level": 2,
"order": 21,
"section_id": "validation-receipt-path",
"source_line_end": 209,
"source_line_start": 193,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The fixture command writes the result, the bundle command validates copied sources and digests, and the test checks mechanics and negative cases; a pass means the copied bodies produced exact values.** The fixture command writes the bounded result and sign-off JSON. The bundle command validates the copied sources, manifest digests, the four exercises, the body-exclusion posture, and the scope fields. The focused test checks fixture mechanics, bundle validation, and the negative cases. A pass means the copied bodies ran and produced the exact expected values on the public fixture.",
"level": 4,
"order": 22,
"section_id": "what-the-three-commands-verify",
"source_line_end": 220,
"source_line_start": 210,
"title": "What the three commands verify"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 23,
"section_id": "claim-ceiling",
"source_line_end": 222,
"source_line_start": 221,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The evidence supports only that the four copied tools import, run, produce exact expected values, and reject the four bad inputs, with no body text in the record.** The strongest claim the evidence supports: the four copied tools import and run in the public clone, produce the exact expected values on a fixed input, and reject the four declared bad inputs. The result record carries source refs, digests, counts, and outcomes, never copied body text or private state.",
"level": 4,
"order": 24,
"section_id": "the-strongest-supported-claim",
"source_line_end": 231,
"source_line_start": 223,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Running the mechanisms is not a claim they are correct: AUTHORITYCEILING refuses oracle truth, prediction correctness, live execution, mutation, and launch.** The proof boundary stops there. Running set arithmetic is not a claim that the observer rows are correct. Applying an edit script is not a claim that the edit is right. A stable hash is not oracle truth. The scope limit refuses oracle truth, prediction correctness, semantic edit correctness, live bridge or Lab execution, external model access, repository or source-file changes, public sharing decision, launch-scope decision, and whole-system correctness. These values are fixed in scope_limit and copied into every result card by result\u2026",
"level": 4,
"order": 25,
"section_id": "what-the-authority-ceiling-refuses",
"source_line_end": 244,
"source_line_start": 232,
"title": "What the scope limit refuses"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch8_tools_tail_primitives_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 8 Tools-Tail Primitives Bundle validates four public tools-tail primitive exercises without Oracle truth, external model access, live bridge work, or repository mutation authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch8_tools_tail_primitives_capsule.py",
"path": "src/microcosm_core/organs/batch8_tools_tail_primitives_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for observer-diff, JSON-patch VM, ledger-id, shadow-envelope, source-manifest, negative-case, and scope limit checks.",
"symbols": [
"_exercise_observer_diff",
"_exercise_version_committer",
"_exercise_ledger_id",
"_exercise_shadow",
"evaluate",
"run",
"run_batch8_bundle",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch8_tools_tail_primitives_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_tools_tail_primitives_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_tools_tail_primitives_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 8 Tools-Tail Primitives Bundle imports observer set diffing, JSON patch interpretation, ledger identity hashing, and shadow envelope parse coverage as exact copied source bodies with bounded public exercises. It checks source-module manifests, stable negative cases, exercise outcomes, mechanism rows, digests, and scope limits while excluding GodMode execution, external model access, live bridge work, repository mutation, private lab artifact export, oracle truth, launch, public sharing, and whole-system correctness.",
"summary_line": "Set 8 Tools-Tail Primitives Bundle validates four public tools-tail primitive exercises without Oracle truth, external model access, live bridge work, or repository mutation authority.",
"title": "Set 8 Tools-Tail Primitives Bundle"
},
{
"anchor": "paper-module-batch8-validator-checker-capsule",
"authority_ceiling": "Selected public checker-group fixture and copied validator-source evidence only; no launch-scope decision, hosted-public proof, source-file changes, complete validator-suite proof, external model access authority, publishing-scope decision, launch-scope decision, or whole-system correctness.",
"body_html": "
Six copied checker groups are executed
\n
This component runs a selection of a copied validator's checker functions instead of asserting from a distance that a body imported but never exercised is correct. This component takes one copied validator source file and actually runs a selection of its checker functions, instead of asserting from a distance that the file is correct. An earlier import copied the whole validator body into Microcosm but only ever called its status-judge entry point. Most of the functions sat there as text that nothing exercised.
\n
Observed behavior closes the import-only gap
\n
It closes that gap for six named checker groups by importing the copied validator, driving selected functions with small public inputs, and writing a metadata-only result record. The bundle closes that gap for six named checker groups. It imports the copied self-indexing-cognitive-system/src/idea_microcosm/validators.py, drives selected functions with small public inputs, and writes a metadata-only result record. A pass means those six groups behaved the way their names claim. It does not mean the whole validator suite is fresh or correct.
\n
Purpose
\n
Copying a file is not running it
\n
Copying a file and running it are different claims, so this component calls six checker groups to make a green run rest on observed behavior rather than file presence. Copying a file and running it are different claims. A copied body can contain a function called private_boundary_hits and still be wrong, because nobody ever fed it a string with a home path in it. Asserting "the file is present" is cheap and says almost nothing. This component picks six groups of checkers and calls them, so a green run rests on observed behavior rather than on the file existing.
\n
Six groups span the validator's judgements
\n
The six groups are chosen to span the validator's kinds of judgement, from source-shape anchoring to the no-write entry point, each reaching a different part of the imported body. The six groups are chosen to span the kinds of judgement the validator makes: source-shape anchoring, status-policy judging, private-boundary scanning, specimen checker groups, launch-gate checker groups, and the no-write validate(root, write_receipt=False) entry point. Each group reaches into a different part of the imported body.
\n
Two choices keep a green run honest
\n
Two design choices keep a green run honest: absent private state falls back to source anchors marked public_runtime_source_only, and negative cases are judged from engine outputs, not matched text. Two design choices keep a green run honest. When the private system state is not present on disk, the component does not pretend the runtime checkers passed. It falls back to reading the copied source for the named anchors and marks the rest of the engines public_runtime_source_only, recording that as a stated limit. And the negative cases are judged from the engine outputs, not from whether a fixture file happens to contain the right error string, so a check cannot pass by matching text.
\n
How it works
\n
Two entry points feed one runner
\n
The run and run_batch8_validator_checker_bundle entry points feed one shared runner that validates the source manifest, calls _evaluate, and writes result, board, validation, and sign-off artifacts. The entry points are run and run_batch8_validator_checker_bundle. Both hand the input directory to a shared crown-jewel runner, which validates the source manifest and then calls _evaluate. The runner writes four artifacts: a result, a board, a validation record, and an sign-off record.
\n
_evaluate picks live or source-only mode
\n
_evaluate picks the mode first: six live engines when private state is present, or the source-anchor engine plus source-only witnesses otherwise, turning any non-passing engine into a finding._evaluate picks the mode first. If the private system state is present and the input is a normal fixture, it runs six live engines. If the input is an exported bundle, or the private state is absent, it runs the source-anchor engine plus source-only witnesses for the other five. It then turns any non-passing engine into a finding, and adds a finding if the copied source manifest does not report exactly one module.
\n
The six engines, in the order _evaluate lists them:
\n
Function
What it does
_validator_source_anchor_matrix
Reads the copied body, counts def _..failures( definitions, and checks the required anchors are present. Passes only if the anchors and launch-gate anchors are present and at least 30 checker-failure functions exist.
_status_policy_judge_matrix
Imports the validator and calls judge_status_request on four transitions. A well-evidenced transition must return allow, a transition missing evidence must downgrade, a prohibited upgrade must block, and a poisoned policy must block. It also drives _validate_transition and requires at least one transition failure.
_private_boundary_scanner_matrix
Writes a temporary file with a planted home path and email address, calls private_boundary_hits, and passes only if the returned patterns include both private_home_path and private_email. It records body_in_receipt: false.
_specimen_checker_matrix
Runs six specimen checkers through _zero_failure_matrix, including _status_collapse_suite_failures and _source_shuttle_specimen_failures, and passes only if every one reports zero failures on the existing fixture.
_release_gate_checker_matrix
Runs six launch-gate checkers through _zero_failure_matrix, including _release_standards_axiom_gate_failures and _release_root_compiler_failures, again requiring zero failures.
_validate_entrypoint_witness
Calls validate(root, write_receipt=False) and passes if the status is ok with at least 20 checks. Without private state it checks only that the def validate(root: Path anchor is present in the copied source.
\n
Shared checker helper and the bundle-mode stub
\n
_zero_failure_matrix imports the validator and records a failure count per named checker, while _public_runtime_source_only_engine is the bundle-mode stub reporting a pass backed only by copied-source anchors._zero_failure_matrix is the shared helper the two checker-group engines use. It imports the validator, calls each named checker against the private root, and records the failure count per checker. _public_runtime_source_only_engine is the stub the bundle mode substitutes for a live engine: it reports a pass backed by copied-source anchors, with macro_state_available: false stated on the row.
\n
result_card carries authority and body floors
\n
result_card compresses the run into two floors: an authority_floor of fixed all-false ceiling fields, and a body_floor recording body_in_receipt: false so no copied body text ships.result_card compresses the run into a card with two floors. The authority_floor copies the fixed ceiling fields, all false: release_authorized, publication_authorized, provider_dispatch, model_dispatch, source_mutation_authorized, full_validator_suite_freshness_claim, public_clone_or_hosting_authority, and test_completeness_proof. The body_floor records body_in_receipt: false so the card never carries copied body text.
\nDiagram of the mechanism (8 steps).Diagram source & refs
flowchart TD A["Fixture input or exported bundle"] --> B["Validate source manifest"] B --> C{"Private state present?"} C -- "yes, fixture input" --> D["Six live checker engines"] C -- "no, or bundle input" --> E["Source-anchor engine plus source-only witnesses"] D --> F["Semantic negative-case checks"] E --> F F --> G["Result, board, validation, sign-off records"] G --> H["Result card: authority_floor and body_floor"]
\n
Negative cases
\n
Six negative cases judged from engine output
\n
The fixture declares six negative cases with stable error codes, and _observed_negative_case reads live engine output or the source-only witness to decide whether each guard actually held. The fixture declares six negative cases in EXPECTED_NEGATIVE_CASES, each paired with a stable error code. _observed_negative_case reads the live engine output (or the source-only witness) to decide whether the guard actually held, and evaluate_negative_case emits the error code when it did.
\n
The six cases and their error codes
\n
The six cases run from missing_validator_source to validate_entrypoint_bypassed, and because judgement reads runtime evidence, each moves only when the matching engine observed the protective behavior. The cases are missing_validator_source (BATCH8_VALIDATOR_SOURCE_REQUIRED), policy_allows_poisoning (BATCH8_VALIDATOR_POLICY_POISONING_BLOCK_REQUIRED), private_boundary_blind (BATCH8_VALIDATOR_PRIVATE_BOUNDARY_REQUIRED), specimen_checker_missing (BATCH8_VALIDATOR_SPECIMEN_CHECKERS_REQUIRED), release_gate_missing (BATCH8_VALIDATOR_RELEASE_GATES_REQUIRED), and validate_entrypoint_bypassed (BATCH8_VALIDATOR_ENTRYPOINT_REQUIRED). Because the judgement reads runtime evidence, a case moves only when the matching engine actually observed the protective behavior.
\n
Prior Art Grounding
\n
Three ordinary patterns, no new invention
\n
The component borrows three ordinary patterns, fixture-driven testing, schema validation, and policy/checker separation, adding nothing beyond bounded checker exercises over copied public source. The component borrows three ordinary engineering patterns: fixture-driven testing, where public inputs and expected negative cases are declared and replayed; schema validation, where a data shape is checked against a declared expectation; and policy/checker separation, where the code that decides is kept apart from the code that invokes it. Useful public references are JSON Schema, pytest fixtures, and Open Policy Agent. Microcosm keeps this component to bounded checker exercises over copied public source and adds nothing beyond that shape.
\n
Validation Result record Path
\n
Reader-verifiable commands, run from the microcosm-substrate/ public root:
\n
What passing all three commands proves
\n
A pass on all three commands means the six named groups behaved as claimed and the negative cases moved with runtime evidence, not a claim about the rest of the suite. The first command writes the bounded checker result and sign-off record. The second runs the exported bundle path, which checks copied source, manifest digests, and the source-only witnesses without importing the private validator. A pass on all three means the six named groups behaved as claimed and the negative cases moved with runtime evidence. It is not a claim about the rest of the suite.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim is narrow: six named checker groups run and behave as their names claim, and six negative cases are rejected on runtime evidence, not matched text. The strongest claim the evidence supports is narrow: for six named checker groups, the copied validator functions run and behave the way their names claim, and six declared negative cases are rejected on runtime evidence rather than on matched text. The copied body being large adds nothing to the claim. Only the named anchors, engines, digests, negative cases, and card fields count.
\n
What the bundle refuses to claim
\n
It refuses the rest: no launch-scope decision, no hosted-public proof, no source-file changes, and no complete-suite proof, since bundle mode never re-runs the private suite. It refuses the rest. This is not launch-scope decision, not hosted-public proof, not source-file changes, and not a complete validator-suite proof. Exported bundle mode validates copied anchors and source-only witnesses, it does not re-run the private suite. The proof boundary is selected checker-group exercises over copied public source. Any broader claim about the validator has to come from the source itself, not from this page.
\n",
"components": [
{
"anchor": "component-batch8_validator_checker_capsule",
"component_id": "batch8_validator_checker_capsule",
"public_label": "Validator Checker Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 7,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 8
},
"link_count": 21,
"links": [
{
"id": "batch8_validator_checker_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Validator Checker Bundle",
"relation": "explains",
"url": "component-batch8-validator-checker-capsule.html"
},
{
"id": "mechanism.batch8_validator_checker_capsule.validates_public_validator_checker_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public validator checker bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch8-validator-checker-capsule-validates-public-validator-checker-capsule"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.microcosm_axiom_substrate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Microcosm Axiom System",
"relation": "depends on",
"url": "paper-module-microcosm-axiom-substrate.html"
},
{
"id": "paper_module.engine_room_generated_projection_drift_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Generated Projection Drift Gate",
"relation": "depends on",
"url": "paper-module-engine-room-generated-projection-drift-gate.html"
},
{
"id": "paper_module.public_reveal_walkthrough",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Public Reveal Walkthrough",
"relation": "depends on",
"url": "paper-module-public-reveal-walkthrough.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch10_cold_eval_honesty_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 10 Cold Eval Honesty Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[67:paper_module.batch10_cold_eval_honesty_capsule].depends_on",
"url": "paper-module-batch10-cold-eval-honesty-capsule.html"
},
{
"id": "paper_module.batch12_release_claim_language_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 12 launch claim-Language Gate",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[66:paper_module.batch12_release_claim_language_gate].depends_on",
"url": "paper-module-batch12-release-claim-language-gate.html"
}
],
"paper_module_id": "paper_module.batch8_validator_checker_capsule",
"reader_packet": {
"abstract": "Set 8 Validator Checker Bundle validates selected public checker groups without becoming launch-scope decision or a complete validator-suite proof.",
"authority_ceiling": "Selected public checker-group fixture and copied validator-source evidence only; no launch-scope decision, hosted-public proof, source-file changes, complete validator-suite proof, external model access authority, publishing-scope decision, launch-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch8_validator_checker_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch8-validator-checker-capsule-diagram-1",
"html_url": "docs/paper-module-batch8-validator-checker-capsule.html#paper-module-batch8-validator-checker-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Fixture input or exported bundle\"] --> B[\"Validate source manifest\"] B --> C{\"Private state present?\"} C -- \"yes, fixture input\" --> D[\"Six live checker engines\"] C -- \"no, or bundle input\" --> E[\"Source-anchor engine plus source-only witnesses\"] D --> F[\"Semantic negative-case checks\"] E --> F F --> G[\"Result, board, validation, sign-off records\"] G --> H[\"Result card: authority_floor and body_floor\"]",
"source_format": "mermaid",
"source_hash": "sha256:4d283e01b83ea3b87da3f9fa568ed38fb9c9120fddcfde8f904af04e09e80537",
"source_line_end": 126,
"source_line_start": 116
}
],
"grouped_relation_feeds": [
{
"count": 7,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.microcosm_axiom_substrate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Microcosm Axiom System",
"relation": "depends on",
"url": "paper-module-microcosm-axiom-substrate.html"
},
{
"id": "paper_module.engine_room_generated_projection_drift_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Generated Projection Drift Gate",
"relation": "depends on",
"url": "paper-module-engine-room-generated-projection-drift-gate.html"
},
{
"id": "paper_module.public_reveal_walkthrough",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Public Reveal Walkthrough",
"relation": "depends on",
"url": "paper-module-public-reveal-walkthrough.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch8_validator_checker_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Validator Checker Bundle",
"relation": "explains",
"url": "component-batch8-validator-checker-capsule.html"
},
{
"id": "mechanism.batch8_validator_checker_capsule.validates_public_validator_checker_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public validator checker bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch8-validator-checker-capsule-validates-public-validator-checker-capsule"
}
],
"relation": "explains"
},
{
"count": 9,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch8_validator_checker_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 225,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_validator_checker_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "JSON Schema",
"section_id": "three-ordinary-patterns-no-new-invention",
"url": "https://json-schema.org/"
},
{
"kind": "external_reference",
"label": "pytest fixtures",
"section_id": "three-ordinary-patterns-no-new-invention",
"url": "https://docs.pytest.org/en/stable/reference/fixtures.html"
},
{
"kind": "external_reference",
"label": "Open Policy Agent",
"section_id": "three-ordinary-patterns-no-new-invention",
"url": "https://www.openpolicyagent.org/docs/latest"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component runs a selection of a copied validator's checker functions instead of asserting from a distance that a body imported but never exercised is correct.** This component takes one copied validator source file and actually runs a selection of its checker functions, instead of asserting from a distance that the file is correct. An earlier import copied the whole validator body into Microcosm but only ever called its status-judge entry point. Most of the functions sat there as text that nothing exercised.",
"level": 4,
"order": 1,
"section_id": "six-copied-checker-groups-are-executed",
"source_line_end": 12,
"source_line_start": 3,
"title": "Six copied checker groups are executed"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It closes that gap for six named checker groups by importing the copied validator, driving selected functions with small public inputs, and writing a metadata-only result record.** The bundle closes that gap for six named checker groups. It imports the copied self-indexing-cognitive-system/src/idea_microcosm/validators.py, drives selected functions with small public inputs, and writes a metadata-only result record. A pass means those six groups behaved the way their names claim. It does not mean the whole validator suite is fresh or correct.",
"level": 4,
"order": 2,
"section_id": "observed-behavior-closes-the-import-only-gap",
"source_line_end": 23,
"source_line_start": 13,
"title": "Observed behavior closes the import-only gap"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 3,
"section_id": "purpose",
"source_line_end": 25,
"source_line_start": 24,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Copying a file and running it are different claims, so this component calls six checker groups to make a green run rest on observed behavior rather than file presence.** Copying a file and running it are different claims. A copied body can contain a function called privateboundaryhits and still be wrong, because nobody ever fed it a string with a home path in it. Asserting \"the file is present\" is cheap and says almost nothing. This component picks six groups of checkers and calls them, so a green run rests on observed behavior rather than on the file existing.",
"level": 4,
"order": 4,
"section_id": "copying-a-file-is-not-running-it",
"source_line_end": 36,
"source_line_start": 26,
"title": "Copying a file is not running it"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The six groups are chosen to span the validator's kinds of judgement, from source-shape anchoring to the no-write entry point, each reaching a different part of the imported body.** The six groups are chosen to span the kinds of judgement the validator makes: source-shape anchoring, status-policy judging, private-boundary scanning, specimen checker groups, launch-gate checker groups, and the no-write validate(root, write_receipt=False) entry point. Each group reaches into a different part of the imported body.",
"level": 4,
"order": 5,
"section_id": "six-groups-span-the-validator-s-judgements",
"source_line_end": 46,
"source_line_start": 37,
"title": "Six groups span the validator's judgements"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Two design choices keep a green run honest: absent private state falls back to source anchors marked publicruntimesource_only, and negative cases are judged from engine outputs, not matched text.** Two design choices keep a green run honest. When the private system state is not present on disk, the component does not pretend the runtime checkers passed. It falls back to reading the copied source for the named anchors and marks the rest of the engines publicruntimesource_only, recording that as a stated limit. And the negative cases are judged from the engine outputs, not from whether a fixture file happens to contain the right error string, so a check cannot pass by matching text.",
"level": 4,
"order": 6,
"section_id": "two-choices-keep-a-green-run-honest",
"source_line_end": 58,
"source_line_start": 47,
"title": "Two choices keep a green run honest"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 60,
"source_line_start": 59,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The run and runbatch8validatorchecker_bundle entry points feed one shared runner that validates the source manifest, calls _evaluate, and writes result, board, validation, and sign-off artifacts.** The entry points are run and runbatch8validatorcheckerbundle. Both hand the input directory to a shared crown-jewel runner, which validates the source manifest and then calls _evaluate. The runner writes four artifacts: a result, a board, a validation record, and an sign-off record.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-feed-one-runner",
"source_line_end": 70,
"source_line_start": 61,
"title": "Two entry points feed one runner"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "**_evaluate picks the mode first: six live engines when private state is present, or the source-anchor engine plus source-only witnesses otherwise, turning any non-passing engine into a finding. evaluate picks the mode first. If the private system state is present and the input is a normal fixture, it runs six live engines. If the input is an exported bundle, or the private state is absent, it runs the source-anchor engine plus source-only witnesses for the other five. It then turns any non-passing engine into a finding, and adds a finding if the copied source manifest does not report exactly one module. The six engines, in the order _evaluate lists them:",
"level": 4,
"order": 9,
"section_id": "evaluate-picks-live-or-source-only-mode",
"source_line_end": 92,
"source_line_start": 71,
"title": "`_evaluate` picks live or source-only mode"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "zerofailure_matrix imports the validator and records a failure count per named checker, while publicruntimesourceonly_engine is the bundle-mode stub reporting a pass backed only by copied-source anchors. zerofailurematrix is the shared helper the two checker-group engines use. It imports the validator, calls each named checker against the private root, and records the failure count per checker. publicruntimesourceonly_engine is the stub the bundle mode substitutes for a live engine: it reports a pass backed by copied-source anchors, with macrostateavailable: false stated on the row.",
"level": 4,
"order": 10,
"section_id": "shared-checker-helper-and-the-bundle-mode-stub",
"source_line_end": 103,
"source_line_start": 93,
"title": "Shared checker helper and the bundle-mode stub"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "resultcard compresses the run into two floors: an authority_floor of fixed all-false ceiling fields, and a body_floor recording bodyinreceipt: false so no copied body text ships. resultcard compresses the run into a card with two floors. The authority_floor copies the fixed ceiling fields, all false: releaseauthorized, publicationauthorized, providerdispatch, modeldispatch, sourcemutationauthorized, fullvalidatorsuitefreshnessclaim, publiccloneorhostingauthority, and testcompletenessproof. The bodyfloor records bodyin_receipt: false so the card never carries copied body text.",
"level": 4,
"order": 11,
"section_id": "result-card-carries-authority-and-body-floors",
"source_line_end": 127,
"source_line_start": 104,
"title": "`result_card` carries authority and body floors"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 129,
"source_line_start": 128,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The fixture declares six negative cases with stable error codes, and observednegative_case reads live engine output or the source-only witness to decide whether each guard actually held.** The fixture declares six negative cases in EXPECTEDNEGATIVECASES, each paired with a stable error code. observednegative_case reads the live engine output (or the source-only witness) to decide whether the guard actually held, and evaluatenegativecase emits the error code when it did.",
"level": 4,
"order": 13,
"section_id": "six-negative-cases-judged-from-engine-output",
"source_line_end": 139,
"source_line_start": 130,
"title": "Six negative cases judged from engine output"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The six cases run from missingvalidator_source to validateentrypointbypassed, and because judgement reads runtime evidence, each moves only when the matching engine observed the protective behavior.** The cases are missingvalidatorsource (BATCH8VALIDATORSOURCEREQUIRED), policyallows_poisoning (BATCH8VALIDATORPOLICYPOISONINGBLOCKREQUIRED), privateboundary_blind (BATCH8VALIDATORPRIVATEBOUNDARYREQUIRED), specimencheckermissing (BATCH8VALIDATORSPECIMENCHECKERSREQUIRED), releasegatemissing (BATCH8VALIDATORRELEASEGATESREQUIRED), and validateentrypointbypassed (BATCH8VALIDATORENTRYPOINT_REQUIRED). Because the judgement reads runtime evidence, a case moves only when the matching engine actually obs\u2026",
"level": 4,
"order": 14,
"section_id": "the-six-cases-and-their-error-codes",
"source_line_end": 154,
"source_line_start": 140,
"title": "The six cases and their error codes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 156,
"source_line_start": 155,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component borrows three ordinary patterns, fixture-driven testing, schema validation, and policy/checker separation, adding nothing beyond bounded checker exercises over copied public source.** The component borrows three ordinary engineering patterns: fixture-driven testing, where public inputs and expected negative cases are declared and replayed; schema validation, where a data shape is checked against a declared expectation; and policy/checker separation, where the code that decides is kept apart from the code that invokes it. Useful public references are JSON Schema, pytest fixtures, and Open Policy Agent. Microcosm keeps this component to bounded checker exercises over copied pu\u2026",
"level": 4,
"order": 16,
"section_id": "three-ordinary-patterns-no-new-invention",
"source_line_end": 172,
"source_line_start": 157,
"title": "Three ordinary patterns, no new invention"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Reader-verifiable commands, run from the microcosm-substrate/ public root:",
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 189,
"source_line_start": 173,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass on all three commands means the six named groups behaved as claimed and the negative cases moved with runtime evidence, not a claim about the rest of the suite.** The first command writes the bounded checker result and sign-off record. The second runs the exported bundle path, which checks copied source, manifest digests, and the source-only witnesses without importing the private validator. A pass on all three means the six named groups behaved as claimed and the negative cases moved with runtime evidence. It is not a claim about the rest of the suite.",
"level": 4,
"order": 18,
"section_id": "what-passing-all-three-commands-proves",
"source_line_end": 200,
"source_line_start": 190,
"title": "What passing all three commands proves"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "claim-ceiling",
"source_line_end": 202,
"source_line_start": 201,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest supported claim is narrow: six named checker groups run and behave as their names claim, and six negative cases are rejected on runtime evidence, not matched text.** The strongest claim the evidence supports is narrow: for six named checker groups, the copied validator functions run and behave the way their names claim, and six declared negative cases are rejected on runtime evidence rather than on matched text. The copied body being large adds nothing to the claim. Only the named anchors, engines, digests, negative cases, and card fields count.",
"level": 4,
"order": 20,
"section_id": "the-strongest-supported-claim-is-narrow",
"source_line_end": 213,
"source_line_start": 203,
"title": "The strongest supported claim is narrow"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It refuses the rest: no launch-scope decision, no hosted-public proof, no source-file changes, and no complete-suite proof, since bundle mode never re-runs the private suite.** It refuses the rest. This is not launch-scope decision, not hosted-public proof, not source-file changes, and not a complete validator-suite proof. Exported bundle mode validates copied anchors and source-only witnesses, it does not re-run the private suite. The proof boundary is selected checker-group exercises over copied public source. Any broader claim about the validator has to come from the source itself, not from this page.",
"level": 4,
"order": 21,
"section_id": "what-the-capsule-refuses-to-claim",
"source_line_end": 224,
"source_line_start": 214,
"title": "What the bundle refuses to claim"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch8_validator_checker_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 8 Validator Checker Bundle validates selected public checker groups without becoming launch-scope decision or a complete validator-suite proof."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch8_validator_checker_capsule.py",
"path": "src/microcosm_core/organs/batch8_validator_checker_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for selected validator checker groups, source anchors, private-boundary checks, negative cases, and scope limits.",
"symbols": [
"_validator_source_anchor_matrix",
"_status_policy_judge_matrix",
"_private_boundary_scanner_matrix",
"_zero_failure_matrix",
"_specimen_checker_matrix",
"_release_gate_checker_matrix",
"_validate_entrypoint_witness",
"_public_runtime_source_only_engine"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch8_validator_checker_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_validator_checker_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch8_validator_checker_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 8 Validator Checker Bundle imports the real idea_microcosm validators body and exercises policy well-formedness, status transition judging, private-boundary scanning, zero-failure, specimen, launch-gate, source-bundle, source-shuttle, concurrency, native-guard, launch-root compiler, and no-write validate entrypoint groups. It carries source anchors, public runtime-only evidence, negative cases, and scope limits while excluding launch-scope decision, hosted-public proof, source-file changes, full validator-suite proof, external model access, public sharing, and whole-system correctness.",
"summary_line": "Set 8 Validator Checker Bundle validates selected public checker groups without becoming launch-scope decision or a complete validator-suite proof.",
"title": "Set 8 Validator Checker Bundle"
},
{
"anchor": "paper-module-batch12-market-dashboard-read-model-capsule",
"authority_ceiling": "Fixture-bound market-dashboard read-model evidence and copied source refs only; no launch-scope decision, external model access, whole-system equivalence, market-level conclusions, investment-related actions, publishing-scope decision, or whole-system correctness.",
"body_html": "
The dashboard read model is exercised as copied code
\n
A dashboard read model turns a market graph into screen shapes, and it is the easiest place for a stale feed or overclaim to slip in. A market dashboard is a presentation layer. It takes a generated graph of market situations and turns it into the shapes a screen consumes: a trust strip, a ranked queue, a detail index, a graph slice, facets, and drilldowns. That read model is also the easiest place for an overclaim to slip in. A stale feed can render green. An auto_apply_allowed flag can stay true. A copied label can read "strong buy" without anyone deciding it should.
\n
Three checks bind rows, filters, and claim limits
\n
It imports the copied, read-model source as runnable code and drives its checks over small synthetic fixtures. The question it answers is narrow: when the read model is handed a planted overclaim, does the source flag it, or does the surface stay quiet? A pass is fixture evidence that the copied checks behave. It is not market-level conclusions, investment-related actions, or a live feed.
\n
Dashboard evidence remains metadata-only and non-live
\n
Three copied checks run, and every result is metadata-only: refs, digests, counts, and verdicts only, never the source text. The component runs three copied checks and records what each one did. Every result is metadata-only: refs, digests, counts, and verdicts only, never the copied source text or any model-output data.
\n
Purpose
\n
The failure mode worth catching
\n
A loose harness that runs the copied source once on clean input proves nothing; the real risk is a read model silent on a planted overclaim. The copied source is system/lib/market_dashboard_read_model.py. It carries three functions this component cares about: validate_market_dashboard_read_model, _runtime_feed_freshness_overlay, and _related_situations, plus the constant TRADING_CLAIM_PATTERN. A loose harness would import that source, run it once on a clean input, watch it pass, and call it validated. That proves nothing about the boundary. The failure mode worth catching is a read model that passes its own structure checks and then stays silent on a planted overclaim.
\n
A missing rejection is the finding
\n
Each check is fed one input per failure shape, and a missing rejection, not a passing run, is what gets recorded as blocked. So this component feeds the checks inputs designed to trip them, one input per failure shape, and treats a missing rejection as the finding. If the validator does not emit the expected error for a dangling edge or an auto-apply overclaim, that is what gets recorded as blocked. The interesting evidence is what the checks refuse.
\n
How it works
\n
One sandboxed import feeds three checks
\n
The component imports the copied source in a sandbox, runs all three checks over fixture cases, and collects their findings. The component imports the copied source in a sandbox, then drives the three checks over fixture cases and collects findings.
\n
Stubbed sandbox import in _load_source_module
\n
The copied file is executed under throwaway system stubs in sys.modules, so it imports without pulling in any private host dependency._load_source_module resolves the copied file through the exported bundle manifest with _source_target and executes it under _market_dashboard_import_stubs. The stub context registers throwaway system, system.lib, and system.lib.market_situation_graph modules in sys.modules so the copied source imports without pulling in any private dependency, then restores the previous module table on exit. Nothing from the host system is loaded; the graph module is a stand-in with fixed filename and schema constants.
\n
Structure and overclaim gate over five negatives
\n
validate_market_dashboard_read_model first checks the read model is well formed, then enforces the claim boundary, and must emit the matching error for five negatives._compute_validator_cases drives validate_market_dashboard_read_model. That copied check is the structural and overclaim gate. It first checks the read model is well formed: the schema version matches, every queued situation resolves to a detail entry, every graph-slice edge points at a node that exists, and each drilldown route returns metadata only with no .. traversal. It then enforces the claim boundary: auto_apply_allowed must be false, review_gated must be true, no_advice_mode must be on, the silent-omission count must be zero, and copied text is scanned against TRADING_CLAIM_PATTERN for trading and action language. The component runs it against five negative fixtures and confirms the source emits the matching error fragment for each one.
\n
Three freshness states, no green from stale
\n
The overlay classifies each synthetic run as fresh, stale, or blocked, and a stale or absent run never reports green._compute_freshness_cases drives _runtime_feed_freshness_overlay. For each case it writes a synthetic readiness file with _write_readiness, then asks the overlay to classify the run into one of three states. fresh_green_feed needs the run ready, all targets met, no blockers, and same-day generation. stale_green_feed is artifact-backed but no longer same-day; the fixture ages one case 14 days. blocked_missing_artifact covers a run with no readiness file, a success shortfall, or a blocker present. A stale or absent run never reports green. Historical proof does not stand in for a live feed.
\n
No-overlap cohort stays empty
\n
_related_situations builds a capped see-also cohort from shared entities or type, and a situation with no overlap returns empty, not an invented link._compute_related_cases drives _related_situations, the "see also" cohort builder. It collects situations that share an entity or match the situation type, excludes the focus situation, and caps the list. The component checks one boundary in particular: a situation with no entity overlap and a different type returns an empty cohort, not an invented link.
\n
Clean payload, case sets, and the result card
\n
_evaluate runs a clean payload and the three case sets, emits one row per check, and result_card projects the counts and floors into a card._evaluate runs a clean payload first (a clean fixture that produces validator errors is itself a finding), then runs the three case sets, folds their findings together, and emits one mechanism row per check with a pass or blocked status and its negative-case rows. evaluate_negative_case is the per-case entry the runner uses for the ten declared negative cases; it re-derives the source result and matches it against the expected fragment or state. run and run_batch12_market_dashboard_read_model_bundle wrap _evaluate through run_crown_jewel_organ for fixture and exported-bundle inputs, and result_card projects the result into a compact card carrying the mechanism count, the computed negative-case count, and the authority and body floors.
\n
Function
Role
_load_source_module
Imports the copied source under stubbed system modules
_compute_validator_cases
Drives the structure and overclaim gate over five negatives
_compute_freshness_cases
Classifies each run into fresh, stale, or blocked
_compute_related_cases
Checks entity-overlap cohorts and the no-overlap case
_evaluate
Runs the clean payload and all cases, collects findings
evaluate_negative_case
Re-derives one negative case against its expected fragment
\nDiagram of the mechanism (12 steps).Diagram source & refs
Source refs
Validate market dashboard read model
validate_market_dashboard_read_model
Blocked missing artifact
blocked_missing_artifact
flowchart TD A["Synthetic dashboard, freshness, related fixtures"] --> B["_load_source_module (stubbed import of market_dashboard_read_model.py)"] B --> C["validate_market_dashboard_read_model"] C --> C1["Structure: schema, queue-to-detail, graph edges, drilldown route safety"] C --> C2["Ceiling: no auto-apply, review-gated, no-advice, no trading language, zero silent omissions"] B --> D["_runtime_feed_freshness_overlay"] D --> D1["fresh_green_feed"] D --> D2["stale_green_feed"] D --> D3["blocked_missing_artifact"] B --> E["_related_situations"] E --> E1["Entity or type match; self-excluded, capped; no overlap means empty"] C1 --> F["metadata-only result record (refs, digests, counts, verdicts)"] C2 --> F D1 --> F D2 --> F D3 --> F E1 --> F
\n
Negative cases
\n
Ten declared negative cases
\n
Ten cases live in EXPECTED_NEGATIVE_CASES, and five of them drive the validator, each expecting a specific error fragment. The component declares ten negative cases in EXPECTED_NEGATIVE_CASES. Five drive the validator and expect a specific error fragment:
\n
dangling_graph_edge expects $.graph_slice.edges[edge1].target is dangling.
silent_omission_count expects silent_omission_count must be 0.
\n
Freshness and related cases expect states
\n
Four freshness cases expect a blocked or stale state and the last expects an empty cohort, and a missing rejection is recorded as not observed. Four drive the freshness overlay and expect a state rather than a green pass: freshness_missing_readiness_artifact and freshness_success_shortfall and freshness_blocker_list each expect blocked_missing_artifact, and freshness_stale_not_fresh expects stale_green_feed with a non-zero staleness count. The last, related_no_overlap_different_type, expects an empty cohort. When the source does not produce the expected rejection, the case is recorded as not observed, which blocks that mechanism.
\n
Prior Art Grounding
\n
CQRS and dashboard-observability practice
\n
The shape comes from CQRS read models and dashboard observability, with an added overclaim gate and a freshness state that refuses green from stale evidence. The read-model shape comes from CQRS and dashboard-observability practice: derive presentation-ready projections from source data, make feed freshness visible, and keep the read surface separate from any mutation authority. Two public anchors are relevant. Microsoft's CQRS pattern describes read models optimized for queries rather than command handling. Grafana dashboards query and transform data sources into operational panels. This component borrows the shape and adds an explicit overclaim gate and a freshness state that refuses to report green from stale evidence.
\n
Validation Result record Path
\n
Reader-verifiable commands, run from the microcosm-substrate/ public root:
\n
What the two commands write and prove
\n
The fixture and bundle commands validate the copied source, digests, and freshness rows, and a pass means every check behaved with a metadata-only result record. The fixture command writes the read-model result and sign-off JSON. The bundle command validates the copied source, its manifest digests, the freshness rows, the related-situation cohorts, and the metadata-only posture. A pass means the three checks behaved as declared over the synthetic fixtures: the clean payload validated, each negative fragment or state appeared, and the result record stayed metadata-only. These are validation result records for the fixtures, nothing more.
\n
Scope boundary
\n
Scope limit
\n
The fixture-bound proof boundary
\n
Over synthetic fixtures the copied checks emit the expected errors, classify freshness honestly, and keep no-overlap cohorts empty, and that is the whole boundary. The strongest claim the evidence supports: over public synthetic fixtures, the copied read-model checks emit the expected structural and overclaim errors, classify feed freshness into honest fresh, stale, and blocked states, and build entity-overlap cohorts that stay empty when nothing overlaps. That is the proof boundary. It is fixture-bound mechanism evidence.
\n
What it refuses past that line
\n
Past the boundary it is not market-level conclusions, advice, or a external model access, and grants no launch, mutation, or whole-system authority. It refuses everything past that line. It is not live market-level conclusions, not investment-related actions, and not a external model access. It excludes launch or public sharing, does not claim whole-system equivalence, excludes source-file changes, and does not assert whole-system correctness. That is the scope limit.
\n",
"components": [
{
"anchor": "component-batch12_market_dashboard_read_model_capsule",
"component_id": "batch12_market_dashboard_read_model_capsule",
"public_label": "Market Dashboard Read-Model Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 2,
"principle": 4
},
"link_count": 13,
"links": [
{
"id": "batch12_market_dashboard_read_model_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Market Dashboard Read-Model Bundle",
"relation": "explains",
"url": "component-batch12-market-dashboard-read-model-capsule.html"
},
{
"id": "mechanism.batch12_market_dashboard_read_model_capsule.validates_public_market_dashboard_read_model_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public market dashboard read model bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch12-market-dashboard-read-model-capsule-validates-public-market-dashboard-read-model-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
},
{
"id": "paper_module.prediction_oracle_reconciliation",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Prediction Oracle Reconciliation",
"relation": "depends on",
"url": "paper-module-prediction-oracle-reconciliation.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch6_unsurfaced_primitives_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 6 Unsurfaced Primitives Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[75:paper_module.batch6_unsurfaced_primitives_capsule].depends_on",
"url": "paper-module-batch6-unsurfaced-primitives-capsule.html"
}
],
"paper_module_id": "paper_module.batch12_market_dashboard_read_model_capsule",
"reader_packet": {
"abstract": "Set 12 Market Dashboard Read-Model Bundle validates copied read-model helpers over public fixtures without market-level conclusions, external model access, or launch-scope decision.",
"authority_ceiling": "Fixture-bound market-dashboard read-model evidence and copied source refs only; no launch-scope decision, external model access, whole-system equivalence, market-level conclusions, investment-related actions, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch12_market_dashboard_read_model_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch12-market-dashboard-read-model-capsule-diagram-1",
"html_url": "docs/paper-module-batch12-market-dashboard-read-model-capsule.html#paper-module-batch12-market-dashboard-read-model-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Synthetic dashboard, freshness, related fixtures\"] --> B[\"_load_source_module (stubbed import of market_dashboard_read_model.py)\"] B --> C[\"validate_market_dashboard_read_model\"] C --> C1[\"Structure: schema, queue-to-detail, graph edges, drilldown route safety\"] C --> C2[\"Ceiling: no auto-apply, review-gated, no-advice, no trading language, zero silent omissions\"] B --> D[\"_runtime_feed_freshness_overlay\"] D --> D1[\"fresh_green_feed\"] D --> D2[\"stale_green_feed\"] D --> D3[\"blocked_missing_artifact\"] B --> E[\"_related_situations\"] E --> E1[\"Entity or type match; self-excluded, capped; no overlap means empty\"] C1 --> F[\"metadata-only result record (refs, digests, counts, verdicts)\"] C2 --> F D1 --> F D2 --> F D3 --> F E1 --> F",
"source_format": "mermaid",
"source_hash": "sha256:523d4c0f70c2bb83e74b2ddfef203b97ad1f7f05c8352f4080316a432e85223b",
"source_line_end": 79,
"source_line_start": 61
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
}
],
"relation": "abides by"
},
{
"count": 2,
"links": [
{
"id": "paper_module.prediction_oracle_reconciliation",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Prediction Oracle Reconciliation",
"relation": "depends on",
"url": "paper-module-prediction-oracle-reconciliation.html"
},
{
"id": "paper_module.world_model_projection_drift_control_room",
"kind": "paper_module",
"kind_label": "paper module",
"label": "World-Model Projection Drift Control Room",
"relation": "depends on",
"url": "paper-module-world-model-projection-drift-control-room.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch12_market_dashboard_read_model_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Market Dashboard Read-Model Bundle",
"relation": "explains",
"url": "component-batch12-market-dashboard-read-model-capsule.html"
},
{
"id": "mechanism.batch12_market_dashboard_read_model_capsule.validates_public_market_dashboard_read_model_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public market dashboard read model bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch12-market-dashboard-read-model-capsule-validates-public-market-dashboard-read-model-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch12_market_dashboard_read_model_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 135,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch12_market_dashboard_read_model_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "CQRS pattern",
"section_id": "cqrs-and-dashboard-observability-practice",
"url": "https://learn.microsoft.com/en-us/azure/architecture/patterns/cqrs"
},
{
"kind": "external_reference",
"label": "Grafana dashboards",
"section_id": "cqrs-and-dashboard-observability-practice",
"url": "https://grafana.com/docs/grafana/latest/visualizations/dashboards/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "A dashboard read model turns a market graph into screen shapes, and it is the easiest place for a stale feed or overclaim to slip in. A market dashboard is a presentation layer. It takes a generated graph of market situations and turns it into the shapes a screen consumes: a trust strip, a ranked queue, a detail index, a graph slice, facets, and drilldowns. That read model is also the easiest place for an overclaim to slip in. A stale feed can render green. An autoapplyallowed flag can stay true. A copied label can read \"strong buy\" without anyone deciding it should.",
"level": 4,
"order": 1,
"section_id": "the-dashboard-read-model-is-exercised-as-copied-code",
"source_line_end": 6,
"source_line_start": 3,
"title": "The dashboard read model is exercised as copied code"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It imports the copied read-model source and asks one thing: handed a planted overclaim, does the source flag it or stay quiet? It imports the copied, read-model source as runnable code and drives its checks over small synthetic fixtures. The question it answers is narrow: when the read model is handed a planted overclaim, does the source flag it, or does the surface stay quiet? A pass is fixture evidence that the copied checks behave. It is not market-level conclusions, investment-related actions, or a live feed.",
"level": 4,
"order": 2,
"section_id": "three-checks-bind-rows-filters-and-claim-limits",
"source_line_end": 10,
"source_line_start": 7,
"title": "Three checks bind rows, filters, and claim limits"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Three copied checks run, and every result is metadata-only: refs, digests, counts, and verdicts only, never the source text. The component runs three copied checks and records what each one did. Every result is metadata-only: refs, digests, counts, and verdicts only, never the copied source text or any model-output data.",
"level": 4,
"order": 3,
"section_id": "dashboard-evidence-remains-body-free-and-non-live",
"source_line_end": 14,
"source_line_start": 11,
"title": "Dashboard evidence remains metadata-only and non-live"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A loose harness that runs the copied source once on clean input proves nothing; the real risk is a read model silent on a planted overclaim. The copied source is system/lib/marketdashboardreadmodel.py. It carries three functions this component cares about: validatemarketdashboardreadmodel, runtimefeedfreshnessoverlay, and relatedsituations, plus the constant TRADINGCLAIM_PATTERN. A loose harness would import that source, run it once on a clean input, watch it pass, and call it validated. That proves nothing about the boundary. The failure mode worth catching is a read model that passes its own structure checks and then stays silent on a planted overclaim.",
"level": 4,
"order": 5,
"section_id": "the-failure-mode-worth-catching",
"source_line_end": 20,
"source_line_start": 17,
"title": "The failure mode worth catching"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Each check is fed one input per failure shape, and a missing rejection, not a passing run, is what gets recorded as blocked. So this component feeds the checks inputs designed to trip them, one input per failure shape, and treats a missing rejection as the finding. If the validator does not emit the expected error for a dangling edge or an auto-apply overclaim, that is what gets recorded as blocked. The interesting evidence is what the checks refuse.",
"level": 4,
"order": 6,
"section_id": "a-missing-rejection-is-the-finding",
"source_line_end": 24,
"source_line_start": 21,
"title": "A missing rejection is the finding"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component imports the copied source in a sandbox, runs all three checks over fixture cases, and collects their findings. The component imports the copied source in a sandbox, then drives the three checks over fixture cases and collects findings.",
"level": 4,
"order": 8,
"section_id": "one-sandboxed-import-feeds-three-checks",
"source_line_end": 31,
"source_line_start": 27,
"title": "One sandboxed import feeds three checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The copied file is executed under throwaway system stubs in sys.modules, so it imports without pulling in any private host dependency. loadsourcemodule resolves the copied file through the exported bundle manifest with sourcetarget and executes it under marketdashboardimportstubs. The stub context registers throwaway system, system.lib, and system.lib.marketsituation_graph modules in sys.modules so the copied source imports without pulling in any private dependency, then restores the previous module table on exit. Nothing from the host system is loaded; the graph module is a stand-in with fixed filename and schema constants.",
"level": 4,
"order": 9,
"section_id": "stubbed-sandbox-import-in-load-source-module",
"source_line_end": 35,
"source_line_start": 32,
"title": "Stubbed sandbox import in `_load_source_module`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatemarketdashboardreadmodel first checks the read model is well formed, then enforces the claim boundary, and must emit the matching error for five negatives. computevalidatorcases drives validatemarketdashboardreadmodel. That copied check is the structural and overclaim gate. It first checks the read model is well formed: the schema version matches, every queued situation resolves to a detail entry, every graph-slice edge points at a node that exists, and each drilldown route returns metadata only with no .. traversal. It then enforces the claim boundary: autoapplyallowed must be false, reviewgated must be true, noadvicemode must be on, the silent-omission count must be zero, and copi\u2026",
"level": 4,
"order": 10,
"section_id": "structure-and-overclaim-gate-over-five-negatives",
"source_line_end": 39,
"source_line_start": 36,
"title": "Structure and overclaim gate over five negatives"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The overlay classifies each synthetic run as fresh, stale, or blocked, and a stale or absent run never reports green. computefreshnesscases drives runtimefeedfreshnessoverlay. For each case it writes a synthetic readiness file with writereadiness, then asks the overlay to classify the run into one of three states. freshgreenfeed needs the run ready, all targets met, no blockers, and same-day generation. stalegreenfeed is artifact-backed but no longer same-day; the fixture ages one case 14 days. blockedmissing_artifact covers a run with no readiness file, a success shortfall, or a blocker present. A stale or absent run never reports green. Historical proof does not stand in for a live feed.",
"level": 4,
"order": 11,
"section_id": "three-freshness-states-no-green-from-stale",
"source_line_end": 43,
"source_line_start": 40,
"title": "Three freshness states, no green from stale"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "relatedsituations builds a capped see-also cohort from shared entities or type, and a situation with no overlap returns empty, not an invented link. computerelatedcases drives relatedsituations, the \"see also\" cohort builder. It collects situations that share an entity or match the situation type, excludes the focus situation, and caps the list. The component checks one boundary in particular: a situation with no entity overlap and a different type returns an empty cohort, not an invented link.",
"level": 4,
"order": 12,
"section_id": "no-overlap-cohort-stays-empty",
"source_line_end": 47,
"source_line_start": 44,
"title": "No-overlap cohort stays empty"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "evaluate runs a clean payload and the three case sets, emits one row per check, and resultcard projects the counts and floors into a card. evaluate runs a clean payload first (a clean fixture that produces validator errors is itself a finding), then runs the three case sets, folds their findings together, and emits one mechanism row per check with a pass or blocked status and its negative-case rows. evaluatenegativecase is the per-case entry the runner uses for the ten declared negative cases; it re-derives the source result and matches it against the expected fragment or state. run and runbatch12marketdashboardreadmodelbundle wrap evaluate through runcrownjewelorgan for fixture and exporte\u2026",
"level": 4,
"order": 13,
"section_id": "clean-payload-case-sets-and-the-result-card",
"source_line_end": 80,
"source_line_start": 48,
"title": "Clean payload, case sets, and the result card"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 82,
"source_line_start": 81,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "Ten cases live in EXPECTEDNEGATIVECASES, and five of them drive the validator, each expecting a specific error fragment. The component declares ten negative cases in EXPECTEDNEGATIVECASES. Five drive the validator and expect a specific error fragment: danglinggraphedge expects $.graph_slice.edges[edge1].target is dangling. traversalrouteref expects route_ref contains traversal. oracleautoapplyoverclaim expects autoapply_allowed must be false. stricttradingclaim_language expects contains trading/action claim language. silentomissioncount expects silentomissioncount must be 0.",
"level": 4,
"order": 15,
"section_id": "ten-declared-negative-cases",
"source_line_end": 92,
"source_line_start": 83,
"title": "Ten declared negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Four freshness cases expect a blocked or stale state and the last expects an empty cohort, and a missing rejection is recorded as not observed. Four drive the freshness overlay and expect a state rather than a green pass: freshnessmissingreadinessartifact and freshnesssuccessshortfall and freshnessblockerlist each expect blockedmissingartifact, and freshnessstalenotfresh expects stalegreenfeed with a non-zero staleness count. The last, relatednooverlapdifferenttype, expects an empty cohort. When the source does not produce the expected rejection, the case is recorded as not observed, which blocks that mechanism.",
"level": 4,
"order": 16,
"section_id": "freshness-and-related-cases-expect-states",
"source_line_end": 96,
"source_line_start": 93,
"title": "Freshness and related cases expect states"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 98,
"source_line_start": 97,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The shape comes from CQRS read models and dashboard observability, with an added overclaim gate and a freshness state that refuses green from stale evidence. The read-model shape comes from CQRS and dashboard-observability practice: derive presentation-ready projections from source data, make feed freshness visible, and keep the read surface separate from any mutation authority. Two public anchors are relevant. Microsoft's CQRS pattern describes read models optimized for queries rather than command handling. Grafana dashboards query and transform data sources into operational panels. This component borrows the shape and adds an explicit overclaim gate and a freshness state that refuses to r\u2026",
"level": 4,
"order": 18,
"section_id": "cqrs-and-dashboard-observability-practice",
"source_line_end": 102,
"source_line_start": 99,
"title": "CQRS and dashboard-observability practice"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Reader-verifiable commands, run from the microcosm-substrate/ public root:",
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 120,
"source_line_start": 103,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The fixture and bundle commands validate the copied source, digests, and freshness rows, and a pass means every check behaved with a metadata-only result record. The fixture command writes the read-model result and sign-off JSON. The bundle command validates the copied source, its manifest digests, the freshness rows, the related-situation cohorts, and the metadata-only posture. A pass means the three checks behaved as declared over the synthetic fixtures: the clean payload validated, each negative fragment or state appeared, and the result record stayed metadata-only. These are validation result records for the fixtures, nothing more.",
"level": 4,
"order": 20,
"section_id": "what-the-two-commands-write-and-prove",
"source_line_end": 124,
"source_line_start": 121,
"title": "What the two commands write and prove"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 126,
"source_line_start": 125,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Over synthetic fixtures the copied checks emit the expected errors, classify freshness honestly, and keep no-overlap cohorts empty, and that is the whole boundary. The strongest claim the evidence supports: over public synthetic fixtures, the copied read-model checks emit the expected structural and overclaim errors, classify feed freshness into honest fresh, stale, and blocked states, and build entity-overlap cohorts that stay empty when nothing overlaps. That is the proof boundary. It is fixture-bound mechanism evidence.",
"level": 4,
"order": 22,
"section_id": "the-fixture-bound-proof-boundary",
"source_line_end": 130,
"source_line_start": 127,
"title": "The fixture-bound proof boundary"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Past the boundary it is not market-level conclusions, advice, or a external model access, and grants no launch, mutation, or whole-system authority. It refuses everything past that line. It is not live market-level conclusions, not investment-related actions, and not a external model access. It excludes launch or public sharing, does not claim whole-system equivalence, excludes source-file changes, and does not assert whole-system correctness. That is the scope limit.",
"level": 4,
"order": 23,
"section_id": "what-it-refuses-past-that-line",
"source_line_end": 134,
"source_line_start": 131,
"title": "What it refuses past that line"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch12_market_dashboard_read_model_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 12 Market Dashboard Read-Model Bundle validates copied read-model helpers over public fixtures without market-level conclusions, external model access, or launch-scope decision."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch12_market_dashboard_read_model_capsule.py",
"path": "src/microcosm_core/organs/batch12_market_dashboard_read_model_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for market-dashboard read-model fixtures, source digest checks, freshness and related-situation cases, negative cases, and scope limits.",
"symbols": [
"_market_dashboard_import_stubs",
"_compute_validator_cases",
"_compute_freshness_cases",
"_compute_related_cases",
"_evaluate",
"run",
"run_batch12_market_dashboard_read_model_bundle",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch12_market_dashboard_read_model_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch12_market_dashboard_read_model_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch12_market_dashboard_read_model_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 12 Market Dashboard Read-Model Bundle runs market-dashboard read-model over public synthetic fixtures. It validates market dashboard import stubs, validator-case derivation, runtime feed freshness overlays, related-situation rows, source anchors, digest checks, negative cases, and scope limits while excluding launch-scope decision, external model access, whole-system equivalence, live market-level conclusions, investment-related actions, public sharing, and whole-system correctness.",
"summary_line": "Set 12 Market Dashboard Read-Model Bundle validates copied read-model helpers over public fixtures without market-level conclusions, external model access, or launch-scope decision.",
"title": "Set 12 Market Dashboard Read-Model Bundle"
},
{
"anchor": "paper-module-batch12-prediction-market-board-capsule",
"authority_ceiling": "Fixture-bound prediction-board and quant-mart diagnostic evidence plus copied source refs only; no market-level conclusions, provider truth, investment-related actions, external model access, whole-system equivalence, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_html": "
The board bundle runs one copied market projection
\n
A copied presentation-mart source runs over synthetic prediction-market and source-feed fixtures to check the board keeps guesses labelled as guesses. This component takes a copied presentation-mart source file, runs it over synthetic prediction-market and source-feed fixtures, and checks that the board keeps guesses labelled as guesses. It does not read live markets and it does not say what any price means.
\n
Fixture rows exercise ordering and visibility rules
\n
The bundle stubs the private imports, loads quant_presentation_mart.py from disk, and calls seven helpers, each against a fixture built to expose one way a board misleads. The copied source is system/lib/quant_presentation_mart.py. That file has private imports the public clone cannot satisfy, so the bundle stubs those imports, loads the module from disk, and calls seven of its helper functions directly. Each call is checked against a fixture built to expose one specific way a board can mislead.
\n
A passing board is not a live market
\n
The output carries per-mechanism status, computed negative cases, and an scope limit as diagnostic rows, never source bodies or model-output data. The output is a result record with per-mechanism status, the named negative cases it computed, and an scope limit. It carries diagnostic rows, not source bodies and not model-output data.
\n
Purpose
\n
How loose aggregation fakes a fact
\n
The shared failure mode is a row that looks like a fact but is an aggregation artifact: a doubled volume, a fabricated id, a false-healthy feed. Market and source dashboards share a failure mode: a row looks like a fact when it is really an artifact of loose aggregation. A duplicate listing doubles a volume figure. An unmatched market slug is handed a fabricated event identity. A feed returns zero rows but shows as healthy. A "change since last time" number appears with no prior run to compare against. A vintage date is displayed when the lifecycle data never carried one.
\n
Asserting the exact conservative behaviour
\n
The bundle runs the real helper bodies over synthetic inputs and asserts the exact conservative behaviour rather than trusting a silent default. The single question here is whether the copied presentation logic keeps those distinctions honest when run over public synthetic inputs. Loose aggregation fails silently: it defaults a missing value to a confident zero, or gives an orphan slug a synthetic id. This bundle runs the real helper bodies and asserts the exact conservative behaviour instead.
\n
How it works
\n
The pipeline and its two entry points
\n
The pipeline is a fixture load, stubbed import, seven helper calls, and eight invariant checks, and both entry points forward to run_crown_jewel_organ. The pipeline is a fixture load, a stubbed source import, seven helper calls, and eight invariant checks. run and run_batch12_prediction_market_board_bundle both forward to run_crown_jewel_organ with _evaluate as the evaluator and evaluate_negative_case as the negative-case evaluator. The bundle entrypoint differs only by input mode.
\n
Resolving and importing the copied source
\n
_source_manifest_for_input resolves the manifest and _load_source_module reads it to find and import quant_presentation_mart.py from its file location._source_manifest_for_input resolves the source manifest. It prefers a source_module_manifest.json sitting in the input directory, and otherwise points at the exported bundle manifest under the public root. _load_source_module reads that manifest through _source_target to find system/lib/quant_presentation_mart.py, then imports it from its file location.
\n
Scoped stand-in modules in _quant_mart_import_stubs
\n
Lightweight stand-in modules are injected into sys.modules so the copied source loads, then restored on exit so the stub stays scoped to the import._quant_mart_import_stubs is the context manager that makes the copied source loadable. The source imports system.lib.market_feed_run_evidence, system.lib.market_fusion_readiness, and system.lib.finance_numeric_assurance. The bundle injects lightweight stand-in modules for those names into sys.modules, including a FEED_NODE_IDS list of global_stock_feed, global_news_feed, and global_macro_feed. On exit it restores whatever modules were there before, so the stub is scoped to the import.
\n
Three prediction-market invariants in _evaluate
\n
_evaluate calls _prediction_market_board and checks three invariants: the duplicate fold, the orphan with no fabricated identity, and the deduped aggregate._evaluate runs the prediction-market half. It loads prediction_market_rows.json, polymarket_identity_artifact.json, and quant_mart_helper_cases.json, then calls the source function _prediction_market_board with the Polymarket identity artifact. The board rows are keyed by event_identity_status. The available event and the missing_from_feed_artifact orphan drive three checks: the surviving market has volume 900000.0 with a market count of 1 (the duplicate fold), the orphan has a null event_id and an aggregate max_liquidity of 0.0 (no fabricated identity), and the deduped aggregate reports market_count1 and max_volume900000.0.
\n
Five source-feed invariants across four helpers
\n
Four source helpers check drift flags kept separate from FRED flags, missingness states, an unavailable delta with no baseline, and vintage bound to the structured source record._evaluate_quant_mart_helpers runs the source-feed half against four more source functions. _provider_drift_monitor produces per-feed drift flags: the stock feed must carry provider_fallback_used, html_response_seen, and fetch_failures, the news feed must stay empty, and the source feed must carry the FRED-specific fred_invalid_series and fred_network_warning. Keeping the generic transport flags separate from the FRED flags means a source data-source fault is not laundered into a generic warning. _missingness_board lists only feeds that are not both populated and ok: the healthy feed is absent, the empty feed reads zero_rows, the degraded feed reads quality_degraded. _delta_since_previous_green returns status: unavailable with an empty row_deltas_by_lane when no prior green run exists. _macro_regime_board, fed by _macro_lifecycle_by_slug, binds each bucket's vintage_status to whether the lifecycle structured source record carries it: the inflation bucket reads available with observation date 2026-05-01, and the growth bucket, whose lifecycle row is absent, reads missing_from_feed_artifact.
\n
Findings, negative-case scoring, and the card
\n
A failed check becomes a finding that flips the status to blocked, evaluate_negative_case scores each case, and result_card projects a metadata-only card. Any check that does not hold becomes a structured finding through finding, and the result status flips from pass to blocked. evaluate_negative_case re-runs _evaluate, pulls the matching case row with _computed_case_row, and returns a blocked result with the expected stable error codes when the invariant held, a pass when it did not, or an error record if the run raised. result_card projects the result into a public card carrying mechanism_count, computed_negative_case_count, and an authority floor read from AUTHORITY_CEILING, with body_in_receipt false throughout.
\n
Function
Role
run, run_batch12_prediction_market_board_bundle
Entry points; forward to the crown-jewel runner with the evaluators
_source_manifest_for_input, _load_source_module
Resolve and import the copied presentation-mart source
_quant_mart_import_stubs
Inject and restore stand-in modules so the source loads standalone
_evaluate
Call _prediction_market_board; check the three prediction-market invariants
_evaluate_quant_mart_helpers
Call the four source-feed helpers; check the five source invariants
evaluate_negative_case
Score one named case against its expected error codes
result_card
Project the result into a metadata-only public card
\nDiagram of the mechanism (11 steps).Diagram source & refs
flowchart TD Input["Fixture inputs rows, identity, helper cases"] --> Manifest["_source_manifest_for_input"] Manifest --> Load["_load_source_module via _quant_mart_import_stubs"] Load --> Eval["_evaluate"] Eval --> Board["_prediction_market_board dedup + orphan + aggregate"] Eval --> Helpers["_evaluate_quant_mart_helpers"] Helpers --> Drift["_provider_drift_monitor generic vs FRED flags"] Helpers --> Miss["_missingness_board zero_rows vs quality_degraded"] Helpers --> Delta["_delta_since_previous_green unavailable with no baseline"] Helpers --> Source["_macro_regime_board vintage bound to structured source record"] Board --> Card["result_card status, negative cases, ceiling"] Drift --> Card Miss --> Card Delta --> Card Source --> Card
\n
Negative cases
\n
Eight named cases as required rejections
\n
EXPECTED_NEGATIVE_CASES binds three prediction-market and five source-feed cases to stable error codes, each a rejection the board must make or fail loudly.EXPECTED_NEGATIVE_CASES declares eight named cases, each bound to a stable error code. The three prediction-market cases are duplicate_lower_volume_retained_higher, orphan_slug_no_identity_fabrication, and aggregate_count_and_max_volume_deduped. The five source-feed cases are provider_drift_multisignal_flags, provider_drift_fred_diagnostics_flags, missingness_zero_row_lane_flagged, delta_no_previous_green_unavailable, and macro_lifecycle_vintage_status_bound. A case is a rejection the board must make: keep the higher-volume duplicate, refuse to invent an event id, flag the empty feed, withhold the delta with no baseline, and show a vintage date only when it is real. When an invariant does not hold, the run emits BATCH12_PREDICTION_CASE_NOT_OBSERVED or BATCH12_QUANT_MART_HELPER_CASE_NOT_OBSERVED rather than reporting a silent pass.
\n
Prior Art Grounding
\n
Information-aggregation and provider-join shape
\n
The board borrows the information-aggregation and provider-join shape from prediction-market practice but stays explicitly diagnostic, treating no price as truth. The component borrows the information-aggregation and provider-join shape from prediction-market and public market-data practice, then keeps the board explicitly diagnostic. Relevant anchors include Robin Hanson's information markets framing, where markets aggregate dispersed information about uncertain events; the CFTC's prediction markets explainer, which frames event contracts, market prices, and consumer cautions; and the Polymarket API documentation as a concrete public API family for market, event, and series data. This bundle uses the join and identity-matching shape and nothing else: no price is treated as truth.
\n
Validation Result record Path
\n
Reader-verifiable commands, run from the microcosm-substrate/ public root:
\n
What the commands and test verify
\n
The fixture and bundle commands and the focused test validate the manifest, seven helper calls, and eight cases, and a pass means conservative diagnostics held. The fixture command writes the board result and sign-off record. The bundle command validates the copied source manifest, the seven helper calls, and metadata-only posture. The focused test checks fixture validation, bundle validation, and the eight named cases. A pass means the copied logic produced the expected conservative diagnostics over the synthetic fixtures. Validation result records for this module are the records those commands write.
\n
Scope boundary
\n
Scope limit
\n
The narrow boundary and its false flags
\n
The copied logic produces the expected diagnostics over synthetic fixtures, and AUTHORITY_CEILING sets every market-truth, forecast, advice, and launch flag to false. The strongest claim this supports is narrow: the copied quant_presentation_mart presentation logic runs standalone over public synthetic fixtures and produces the expected identity, drift, missingness, delta, and source-lifecycle diagnostics, with source bodies and model-output data kept out of the result record. That is the proof boundary. The scope limit stops there. AUTHORITY_CEILING sets live_prediction_market_truth, provider_truth, forecast_correctness, calibration_claim, investment_advice, provider_dispatch, release_authorized, publication_authorized, private_root_equivalence_claim, and whole_system_correctness_claim all to false. The board is a diagnostic surface over fixtures. It is not market-level conclusions, not provider truth, not investment-related actions, and not launch-scope decision.
\n",
"components": [
{
"anchor": "component-batch12_prediction_market_board_capsule",
"component_id": "batch12_prediction_market_board_capsule",
"public_label": "Prediction Market Board Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 14,
"links": [
{
"id": "batch12_prediction_market_board_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Prediction Market Board Bundle",
"relation": "explains",
"url": "component-batch12-prediction-market-board-capsule.html"
},
{
"id": "mechanism.batch12_prediction_market_board_capsule.validates_public_prediction_market_board_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public prediction market board bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch12-prediction-market-board-capsule-validates-public-prediction-market-board-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
},
{
"id": "paper_module.prediction_oracle_reconciliation",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Prediction Oracle Reconciliation",
"relation": "depends on",
"url": "paper-module-prediction-oracle-reconciliation.html"
},
{
"id": "paper_module.batch8_structural_theses_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Structural Theses Bundle",
"relation": "depends on",
"url": "paper-module-batch8-structural-theses-capsule.html"
},
{
"id": "paper_module.finance_forecast_evaluation_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Finance Forecast Evaluation Spine",
"relation": "depends on",
"url": "paper-module-finance-forecast-evaluation-spine.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch8_structural_theses_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Structural Theses Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[61:paper_module.batch8_structural_theses_capsule].depends_on",
"url": "paper-module-batch8-structural-theses-capsule.html"
}
],
"paper_module_id": "paper_module.batch12_prediction_market_board_capsule",
"reader_packet": {
"abstract": "Set 12 Prediction Market Board Bundle validates copied prediction-board and quant-mart diagnostics over public fixtures without market-level conclusions or provider authority.",
"authority_ceiling": "Fixture-bound prediction-board and quant-mart diagnostic evidence plus copied source refs only; no market-level conclusions, provider truth, investment-related actions, external model access, whole-system equivalence, launch-scope decision, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch12_prediction_market_board_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch12-prediction-market-board-capsule-diagram-1",
"html_url": "docs/paper-module-batch12-prediction-market-board-capsule.html#paper-module-batch12-prediction-market-board-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Input[\"Fixture inputs rows, identity, helper cases\"] --> Manifest[\"_source_manifest_for_input\"] Manifest --> Load[\"_load_source_module via _quant_mart_import_stubs\"] Load --> Eval[\"_evaluate\"] Eval --> Board[\"_prediction_market_board dedup + orphan + aggregate\"] Eval --> Helpers[\"_evaluate_quant_mart_helpers\"] Helpers --> Drift[\"_provider_drift_monitor generic vs FRED flags\"] Helpers --> Miss[\"_missingness_board zero_rows vs quality_degraded\"] Helpers --> Delta[\"_delta_since_previous_green unavailable with no baseline\"] Helpers --> Source[\"_macro_regime_board vintage bound to structured source record\"] Board --> Card[\"result_card status, negative cases, ceiling\"] Drift --> Card Miss --> Card Delta --> Card Source --> Card",
"source_format": "mermaid",
"source_hash": "sha256:70342b3ee610258e4575a59388a2e6f7df08b6452bea4f3f98f9c8163198e782",
"source_line_end": 77,
"source_line_start": 61
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.prediction_oracle_reconciliation",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Prediction Oracle Reconciliation",
"relation": "depends on",
"url": "paper-module-prediction-oracle-reconciliation.html"
},
{
"id": "paper_module.batch8_structural_theses_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Structural Theses Bundle",
"relation": "depends on",
"url": "paper-module-batch8-structural-theses-capsule.html"
},
{
"id": "paper_module.finance_forecast_evaluation_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Finance Forecast Evaluation Spine",
"relation": "depends on",
"url": "paper-module-finance-forecast-evaluation-spine.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch12_prediction_market_board_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Prediction Market Board Bundle",
"relation": "explains",
"url": "component-batch12-prediction-market-board-capsule.html"
},
{
"id": "mechanism.batch12_prediction_market_board_capsule.validates_public_prediction_market_board_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public prediction market board bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch12-prediction-market-board-capsule-validates-public-prediction-market-board-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "concept.research_and_science_replay_evidence_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "research and science replay validators as reproducibility scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-research-and-science-replay-evidence-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch12_prediction_market_board_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 119,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch12_prediction_market_board_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "information markets",
"section_id": "information-aggregation-and-provider-join-shape",
"url": "https://mason.gmu.edu/~rhanson/infomkts.html"
},
{
"kind": "external_reference",
"label": "prediction markets explainer",
"section_id": "information-aggregation-and-provider-join-shape",
"url": "https://www.cftc.gov/LearnandProtect/PredictionMarkets"
},
{
"kind": "external_reference",
"label": "Polymarket API documentation",
"section_id": "information-aggregation-and-provider-join-shape",
"url": "https://docs.polymarket.com/api-reference"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "A copied presentation-mart source runs over synthetic prediction-market and source-feed fixtures to check the board keeps guesses labelled as guesses. This component takes a copied presentation-mart source file, runs it over synthetic prediction-market and source-feed fixtures, and checks that the board keeps guesses labelled as guesses. It does not read live markets and it does not say what any price means.",
"level": 4,
"order": 1,
"section_id": "the-board-capsule-runs-one-copied-market-projection",
"source_line_end": 6,
"source_line_start": 3,
"title": "The board bundle runs one copied market projection"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The bundle stubs the private imports, loads quantpresentationmart.py from disk, and calls seven helpers, each against a fixture built to expose one way a board misleads. The copied source is system/lib/quantpresentationmart.py. That file has private imports the public clone cannot satisfy, so the bundle stubs those imports, loads the module from disk, and calls seven of its helper functions directly. Each call is checked against a fixture built to expose one specific way a board can mislead.",
"level": 4,
"order": 2,
"section_id": "fixture-rows-exercise-ordering-and-visibility-rules",
"source_line_end": 10,
"source_line_start": 7,
"title": "Fixture rows exercise ordering and visibility rules"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The output carries per-mechanism status, computed negative cases, and an scope limit as diagnostic rows, never source bodies or model-output data. The output is a result record with per-mechanism status, the named negative cases it computed, and an scope limit. It carries diagnostic rows, not source bodies and not model-output data.",
"level": 4,
"order": 3,
"section_id": "a-passing-board-is-not-a-live-market",
"source_line_end": 14,
"source_line_start": 11,
"title": "A passing board is not a live market"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The shared failure mode is a row that looks like a fact but is an aggregation artifact: a doubled volume, a fabricated id, a false-healthy feed. Market and source dashboards share a failure mode: a row looks like a fact when it is really an artifact of loose aggregation. A duplicate listing doubles a volume figure. An unmatched market slug is handed a fabricated event identity. A feed returns zero rows but shows as healthy. A \"change since last time\" number appears with no prior run to compare against. A vintage date is displayed when the lifecycle data never carried one.",
"level": 4,
"order": 5,
"section_id": "how-loose-aggregation-fakes-a-fact",
"source_line_end": 20,
"source_line_start": 17,
"title": "How loose aggregation fakes a fact"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The bundle runs the real helper bodies over synthetic inputs and asserts the exact conservative behaviour rather than trusting a silent default. The single question here is whether the copied presentation logic keeps those distinctions honest when run over public synthetic inputs. Loose aggregation fails silently: it defaults a missing value to a confident zero, or gives an orphan slug a synthetic id. This bundle runs the real helper bodies and asserts the exact conservative behaviour instead.",
"level": 4,
"order": 6,
"section_id": "asserting-the-exact-conservative-behaviour",
"source_line_end": 24,
"source_line_start": 21,
"title": "Asserting the exact conservative behaviour"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The pipeline is a fixture load, stubbed import, seven helper calls, and eight invariant checks, and both entry points forward to runcrownjewelorgan. The pipeline is a fixture load, a stubbed source import, seven helper calls, and eight invariant checks. run and runbatch12predictionmarketboardbundle both forward to runcrownjewelorgan with evaluate as the evaluator and evaluatenegativecase as the negative-case evaluator. The bundle entrypoint differs only by input mode.",
"level": 4,
"order": 8,
"section_id": "the-pipeline-and-its-two-entry-points",
"source_line_end": 30,
"source_line_start": 27,
"title": "The pipeline and its two entry points"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourcemanifestforinput resolves the manifest and loadsourcemodule reads it to find and import quantpresentationmart.py from its file location. sourcemanifestforinput resolves the source manifest. It prefers a sourcemodulemanifest.json sitting in the input directory, and otherwise points at the exported bundle manifest under the public root. loadsourcemodule reads that manifest through sourcetarget to find system/lib/quantpresentation_mart.py, then imports it from its file location.",
"level": 4,
"order": 9,
"section_id": "resolving-and-importing-the-copied-source",
"source_line_end": 34,
"source_line_start": 31,
"title": "Resolving and importing the copied source"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Lightweight stand-in modules are injected into sys.modules so the copied source loads, then restored on exit so the stub stays scoped to the import. quantmartimportstubs is the context manager that makes the copied source loadable. The source imports system.lib.marketfeedrunevidence, system.lib.marketfusionreadiness, and system.lib.financenumericassurance. The bundle injects lightweight stand-in modules for those names into sys.modules, including a FEEDNODEIDS list of globalstockfeed, globalnewsfeed, and globalmacro_feed. On exit it restores whatever modules were there before, so the stub is scoped to the import.",
"level": 4,
"order": 10,
"section_id": "scoped-stand-in-modules-in-quant-mart-import-stubs",
"source_line_end": 38,
"source_line_start": 35,
"title": "Scoped stand-in modules in `_quant_mart_import_stubs`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluate calls predictionmarketboard and checks three invariants: the duplicate fold, the orphan with no fabricated identity, and the deduped aggregate. evaluate runs the prediction-market half. It loads predictionmarketrows.json, polymarketidentityartifact.json, and quantmarthelpercases.json, then calls the source function predictionmarketboard with the Polymarket identity artifact. The board rows are keyed by eventidentitystatus. The available event and the missingfromfeedartifact orphan drive three checks: the surviving market has volume 900000.0 with a market count of 1 (the duplicate fold), the orphan has a null eventid and an aggregate maxliquidity of 0.0 (no fabricated identity), and\u2026",
"level": 4,
"order": 11,
"section_id": "three-prediction-market-invariants-in-evaluate",
"source_line_end": 42,
"source_line_start": 39,
"title": "Three prediction-market invariants in `_evaluate`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Four source helpers check drift flags kept separate from FRED flags, missingness states, an unavailable delta with no baseline, and vintage bound to the structured source record. evaluatequantmarthelpers runs the source-feed half against four more source functions. providerdriftmonitor produces per-feed drift flags: the stock feed must carry providerfallbackused, htmlresponseseen, and fetchfailures, the news feed must stay empty, and the source feed must carry the FRED-specific fredinvalidseries and frednetworkwarning. Keeping the generic transport flags separate from the FRED flags means a source data-source fault is not laundered into a generic warning. missingnessboard lists only feeds t\u2026",
"level": 4,
"order": 12,
"section_id": "five-macro-feed-invariants-across-four-helpers",
"source_line_end": 46,
"source_line_start": 43,
"title": "Five source-feed invariants across four helpers"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "A failed check becomes a finding that flips the status to blocked, evaluatenegativecase scores each case, and resultcard projects a metadata-only card. Any check that does not hold becomes a structured finding through finding, and the result status flips from pass to blocked. evaluatenegativecase re-runs evaluate, pulls the matching case row with computedcaserow, and returns a blocked result with the expected stable error codes when the invariant held, a pass when it did not, or an error record if the run raised. resultcard projects the result into a public card carrying mechanismcount, computednegativecasecount, and an authority floor read from AUTHORITYCEILING, with bodyin_receipt false t\u2026",
"level": 4,
"order": 13,
"section_id": "findings-negative-case-scoring-and-the-card",
"source_line_end": 78,
"source_line_start": 47,
"title": "Findings, negative-case scoring, and the card"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 80,
"source_line_start": 79,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVECASES binds three prediction-market and five source-feed cases to stable error codes, each a rejection the board must make or fail loudly. EXPECTEDNEGATIVECASES declares eight named cases, each bound to a stable error code. The three prediction-market cases are duplicatelowervolumeretainedhigher, orphanslugnoidentityfabrication, and aggregatecountandmaxvolumededuped. The five source-feed cases are providerdriftmultisignalflags, providerdriftfreddiagnosticsflags, missingnesszerorowlaneflagged, deltanopreviousgreenunavailable, and macrolifecyclevintagestatusbound. A case is a rejection the board must make: keep the higher-volume duplicate, refuse to invent an event id, flag th\u2026",
"level": 4,
"order": 15,
"section_id": "eight-named-cases-as-required-rejections",
"source_line_end": 84,
"source_line_start": 81,
"title": "Eight named cases as required rejections"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 86,
"source_line_start": 85,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The board borrows the information-aggregation and provider-join shape from prediction-market practice but stays explicitly diagnostic, treating no price as truth. The component borrows the information-aggregation and provider-join shape from prediction-market and public market-data practice, then keeps the board explicitly diagnostic. Relevant anchors include Robin Hanson's information markets framing, where markets aggregate dispersed information about uncertain events; the CFTC's prediction markets explainer, which frames event contracts, market prices, and consumer cautions; and the Polymarket API documentation as a concrete public API family for market, event, and series data. This bund\u2026",
"level": 4,
"order": 17,
"section_id": "information-aggregation-and-provider-join-shape",
"source_line_end": 90,
"source_line_start": 87,
"title": "Information-aggregation and provider-join shape"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Reader-verifiable commands, run from the microcosm-substrate/ public root:",
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 108,
"source_line_start": 91,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The fixture and bundle commands and the focused test validate the manifest, seven helper calls, and eight cases, and a pass means conservative diagnostics held. The fixture command writes the board result and sign-off record. The bundle command validates the copied source manifest, the seven helper calls, and metadata-only posture. The focused test checks fixture validation, bundle validation, and the eight named cases. A pass means the copied logic produced the expected conservative diagnostics over the synthetic fixtures. Validation result records for this module are the records those commands write.",
"level": 4,
"order": 19,
"section_id": "what-the-commands-and-test-verify",
"source_line_end": 112,
"source_line_start": 109,
"title": "What the commands and test verify"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 114,
"source_line_start": 113,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The copied logic produces the expected diagnostics over synthetic fixtures, and AUTHORITYCEILING sets every market-truth, forecast, advice, and launch flag to false. The strongest claim this supports is narrow: the copied quantpresentationmart presentation logic runs standalone over public synthetic fixtures and produces the expected identity, drift, missingness, delta, and source-lifecycle diagnostics, with source bodies and model-output data kept out of the result record. That is the proof boundary. The scope limit stops there. AUTHORITYCEILING sets livepredictionmarkettruth, providertruth, forecastcorrectness, calibrationclaim, investmentadvice, providerdispatch, releaseauthorized, publi\u2026",
"level": 4,
"order": 21,
"section_id": "the-narrow-boundary-and-its-false-flags",
"source_line_end": 118,
"source_line_start": 115,
"title": "The narrow boundary and its false flags"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch12_prediction_market_board_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 12 Prediction Market Board Bundle validates copied prediction-board and quant-mart diagnostics over public fixtures without market-level conclusions or provider authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch12_prediction_market_board_capsule.py",
"path": "src/microcosm_core/organs/batch12_prediction_market_board_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for prediction-board fixtures, quant-mart helper diagnostics, source digest checks, negative cases, and scope limits.",
"symbols": [
"_quant_mart_import_stubs",
"_evaluate_quant_mart_helpers",
"_evaluate",
"run",
"run_batch12_prediction_market_board_bundle",
"result_card",
"EXPECTED_NEGATIVE_CASES",
"AUTHORITY_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch12_prediction_market_board_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch12_prediction_market_board_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch12_prediction_market_board_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 12 Prediction Market Board Bundle runs prediction-market board and quant-mart diagnostics over public synthetic fixtures. It validates prediction-market joins, Polymarket identity by slug, provider drift monitors, missingness boards, prior green deltas, source lifecycle vintage enrichment, source-module digests, negative cases, and scope limits while excluding launch-scope decision, external model access, whole-system equivalence, market-level conclusions, provider truth, investment-related actions, public sharing, and whole-system correctness.",
"summary_line": "Set 12 Prediction Market Board Bundle validates copied prediction-board and quant-mart diagnostics over public fixtures without market-level conclusions or provider authority.",
"title": "Set 12 Prediction Market Board Bundle"
},
{
"anchor": "paper-module-batch12-release-claim-language-gate",
"authority_ceiling": "Fixture-bound launch-claim language gate evidence and copied source refs only; no launch-scope decision, publishing-scope decision, external model access, whole-system equivalence, market-level conclusions, investment-related actions, or whole-system correctness.",
"body_html": "
launch language is checked against evidence classes
\n
Public copy is scanned line by line for launch, licensing, maturity, and launch-authority wording it cannot support, failing non-zero under --assert-clear while any unsupported claim survives. This component checks public copy for language that claims more than the result records behind it can support. It reads a small public sharing manifest, opens each file the manifest lists, and matches every line against fixed families of launch, licensing, maturity, and launch-authority wording. It writes a metadata-only result record and, when run with --assert-clear, exits non-zero while any unsupported claim remains. It does not approve launch and it does not publish anything.
\n
The gate separates supported wording from overclaim
\n
The private gate module is absent on the public clone, so an in-file _PublicFallbackReleaseClaimGate stands in, making this a check on a checker over shipped fixtures. The gate logic it exercises is a separate module, release_claim_language_gate.py. On the public clone that module is not present, so the component replaces it with an in-file stand-in, _PublicFallbackReleaseClaimGate. The page you can run here is that fallback gate driven over shipped fixtures. This module is therefore a check on a checker: it proves the gate blocks bare launch wording and clears explicit boundary wording over a fixed set of inputs.
\n
Passing prose excludes launch
\n
The scan is a deterministic regular-expression pass that fails closed: any affirmative claim it cannot downgrade becomes an active blocker and forces exit 2. The scan is a deterministic regular-expression pass over lines, not a language model. It fails closed. An affirmative claim it cannot downgrade becomes an active blocker, and main --assert-clear returns exit 2 whenever an active blocker survives.
\n
Purpose
\n
How copy drifts into over-claiming
\n
Public copy drifts toward over-claiming as edits accumulate, and human review misses it because each phrase reads reasonably alone, so the gate asks whether a line's launch status outran the evidence. Public copy drifts toward over-claiming. A page that began as "fixture-proven, not yet published" gets edited over months until someone writes launch, licensing, or maturity wording without noticing that nothing underneath it changed. A loose human review misses this because each phrase reads reasonably in isolation. The gate answers one narrow question: does a line assert launch status that the evidence does not hold, and would an automated check catch it before the copy ships.
\n
A lexical check honest about its reach
\n
Being lexical, the check flags risky phrasing and its excusing context but never judges whether the underlying claim is true, only whether the wording outran the records. Because the check is lexical, it is honest about its own reach. It finds risky phrasing and the context that would excuse it. It does not decide whether the underlying claim is true, only whether the wording has outrun what the result records can back.
\n
How it works
\n
The stages run in this order inside _evaluate.
\n
_load_fixture reads or fails closed
\n
_load_fixture reads release_gate_fixture.json, turning a missing file, bad JSON, or non-object payload into a structured finding and the fail-closed envelope rather than a quiet pass. First, _load_fixture reads release_gate_fixture.json from the selected input directory. A missing file, invalid JSON, or a non-object payload becomes a structured finding, and the run returns the fail-closed envelope from _blocked_exercise rather than a quiet pass.
\n
_load_source_module returns the stub
\n
_load_source_module returns _PublicFallbackReleaseClaimGate when the manifest marks the public stub, still validating the copied module's digest and required anchors before any exercise. Second, _load_source_module resolves the gate. _source_module_public_stubbed inspects the copied-module manifest, and when it sees the public stub substitution it returns _PublicFallbackReleaseClaimGate instead of importing a private body. The copied module is still validated for digest equality and required anchors (RISKY_PHRASES, NEGATIVE_CONTEXT_MARKERS, META_FORBIDDEN_CONTEXT_MARKERS, private_control_plane_leak, def _classify_hit, def build_gate) before any exercise.
\n
_write_gate_fixture materialises each scenario
\n
_write_gate_fixture writes each scenario's docs/ text and a publication_manifest.yaml, while _fixture_doc_name rejects control characters, path separators, and duplicate document names. Third, _write_gate_fixture materialises each scenario on disk: it writes the fixture text into docs/, then writes a publication_manifest.yaml that lists those files. _fixture_doc_name rejects any file name with control characters (BATCH12_RELEASE_FIXTURE_PATH_CONTROL_CHAR) or path separators (BATCH12_RELEASE_FIXTURE_PATH_UNSAFE), and rejects an active document that would overwrite the safe one (BATCH12_RELEASE_FIXTURE_DUPLICATE_DOC_NAME).
\n
build_gate scans and classifies lines
\n
build_gate opens each manifest file and scans every line, recording each match as a metadata-only hit and counting active blockers, boundary hits, and claim surfaces into a status. Fourth, build_gate reads the manifest, opens each listed file, and scans every line against the phrase patterns. Each match is classified by _classify and recorded as a hit carrying its path, line number, phrase id, classification, and a body_in_receipt: false flag. The summary counts active blockers, boundary hits, and claim surfaces, and sets a status.
\n
_run_main_assert_clear captures the exit code
\n
_run_main_assert_clear runs the gate on the active scenario and returns its exit code, after which _evaluate computes the cases and hands off to the metadata-only writers. Fifth, _run_main_assert_clear runs main --assert-clear on the active scenario, captures the printed status, and returns the exit code. _evaluate then computes the negative cases and the positive boundary-clear check and hands the result to run or run_batch12_release_claim_language_gate_bundle, which write the metadata-only records.
\n
Function
Role
_load_fixture
Reads release_gate_fixture.json; converts missing or malformed input into a fail-closed finding.
_write_gate_fixture
Writes the docs/ files and publication_manifest.yaml for one scenario; rejects unsafe or duplicate names.
build_gate
Scans manifest files line by line for risky phrases and emits the classified hit summary.
_classify
Labels each hit as an active blocker or allowed boundary context.
main --assert-clear
Runs the gate and returns exit 2 when any active blocker remains.
_evaluate
Builds the four scenarios, computes the negative cases, and assembles the result record.
run_batch12_release_claim_language_gate_bundle
Wraps _evaluate and writes the metadata-only result, board, and validation records.
\n
_classify reads each hit in context
\n
_classify blocks any internal control-leak phrase outright, allows other hits only when a negation marker is present, and holds internal control vocabulary to a stricter rule that even a negation cannot clear. The interesting part is _classify. It reads a matched line two ways. If the phrase belongs to the private_control_plane_leak family, the line is an active blocker regardless of surrounding words. Otherwise, if the line contains a negative-context marker such as not, forbidden, blocked, or omitted, the hit is boundary-or-negative context and is allowed; with no such marker it is an active blocker. So "not a hosted product" clears and "deployment-posture" blocks. The internal control family is stricter on purpose: even "excludes launch" is blocked, because public reader copy should leave launch-authority vocabulary out entirely rather than negate it.
\n
Status and the ok flag from counts
\n
Status is set from the counts: any active blocker yields active_claim_blocked, and ok is true only at zero blockers, which is when --assert-clear avoids exit 2.build_gate then sets the status from the counts. Any active blocker gives active_claim_blocked; boundary hits with no active blocker give clear_boundary_only; nothing risky gives clear. The ok flag is true only at zero active blockers, and main --assert-clear returns 2 when ok is false.
\n
Four scenarios, one expected to clear
\n
_evaluate runs four scenarios from one fixture, expecting the safe boundary document to clear and the three over-claim documents to block._evaluate exercises four scenarios from one fixture: a safe boundary document, an affirmative over-claim document, a public sharing over-claim document, and a private internal control leak document. The safe scenario is expected to clear as boundary-only. The other three are expected to block.
\nDiagram of the mechanism (8 steps).Diagram source & refs
Source refs
safe and active copy text
release_gate_fixture.json
negation marker present => allowed
boundary_or_negative_context
flowchart TD Fixture["release_gate_fixture.json safe and active copy text"] Write["_write_gate_fixture docs files + publication_manifest.yaml"] Gate["build_gate scan manifest lines for risky phrases"] Classify{"_classify read each hit in context"} Boundary["boundary_or_negative_context negation marker present => allowed"] Active["active_claim_blocker bare launch or internal control wording => blocked"] Assert["main --assert-clear exit 2 when any active blocker remains"] Result["metadata-only result record result, board, validation"] Fixture --> Write Write --> Gate Gate --> Classify Classify -->|allowed| Boundary Classify -->|blocked| Active Active --> Assert Boundary --> Result Assert --> Result
\n
Negative cases
\n
Three fixed cases, each recomputed
\n
EXPECTED_NEGATIVE_CASES fixes three error-coded cases covering the affirmative blocker, the exit-2 assertion, and the internal control leak, blocking the run if any fails to compute.EXPECTED_NEGATIVE_CASES fixes three cases, each with a stable error code. affirmative_open_source_production_ready_blocks (BATCH12_RELEASE_CLAIM_ACTIVE_BLOCKER) requires the active scenario to reach active_claim_blocked with both the open_source and production_ready phrase ids present. assert_clear_returns_exit_2 (BATCH12_RELEASE_CLAIM_ASSERT_CLEAR_EXIT_2) requires main --assert-clear to return exit code 2 with ok false. private_control_plane_public_reader_blocks (BATCH12_RELEASE_PRIVATE_CONTROL_PLANE_LEAK) requires the internal control scenario to block on a private_control_plane_leak family hit. If any case does not compute, _evaluate records BATCH12_RELEASE_CLAIM_CASE_NOT_OBSERVED and the run is blocked.
\n
Prior Art Grounding
\n
Advertising-substantiation and communication rules
\n
The proof-consumer form comes from advertising-substantiation and regulated-communication practice, where FTC, FINRA, and SEC rules require a reasonable basis before a claim is disseminated. The pattern comes from advertising-substantiation and regulated-communication practice: public claims should stay within the evidence actually held, and stronger wording needs stronger support. The FTC advertising-substantiation policy (https://www.ftc.gov/legal-library/browse/ftc-policy-statement-regarding-advertising-substantiation) states that objective claims need a reasonable basis before dissemination. FINRA Rule 2210 and the SEC investment-adviser marketing rule carry the same shape for financial communications. This module borrows only the proof-consumer form of that idea. It does not implement legal compliance and it does not decide whether copy is fit to publish.
\n
Validation Result record Path
\n
Reader-verifiable commands, run from the microcosm-substrate/ public root:
\n
What a pass certifies
\n
A pass certifies the fallback gate cleared the boundary scenario, blocked the three over-claim scenarios, exited 2 under --assert-clear, and wrote metadata-only records. A pass means the fallback gate cleared the boundary scenario, blocked the three over-claim scenarios, returned exit 2 under --assert-clear, and wrote the result, board, and validation records with no copied body text inside them.
\n
Scope boundary
\n
Scope limit
\n
The lexical, fixture-bound proof boundary
\n
Over the shipped fixtures the fallback gate allows boundary-only wording, blocks bare launch, licensing, maturity, and launch-authority wording, and exits 2 while a blocker remains. The strongest claim the evidence supports is narrow: over the shipped fixtures, the fallback gate classifies boundary-only wording as allowed and classifies bare launch, licensing, maturity, and launch-authority wording as active blockers, and main --assert-clear exits 2 while any active blocker remains. That is the proof boundary. The evidence is lexical and fixture-bound.
\n
A pass is evidence, not permission to ship
\n
A green result grants no launch, public sharing, or hosted-product status, proves no semantic understanding or legal sufficiency, and stays evidence about the checker rather than permission to ship. The scope limit stops well short of that. A green result does not grant launch permission, public sharing, or hosted-product status. It does not establish semantic understanding of the copy, complete secret detection, whole-system equivalence, market-level conclusions, legal sufficiency, or whole-system correctness. It excludes editing the copied gate module or changing the fixtures, negative cases, or ceiling; those belong in the source and standard lanes first. A pass is evidence about the checker, not permission to ship.
\n",
"components": [
{
"anchor": "component-batch12_release_claim_language_gate",
"component_id": "batch12_release_claim_language_gate",
"public_label": "Release Public Wording Gate"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 14,
"links": [
{
"id": "batch12_release_claim_language_gate",
"kind": "organ",
"kind_label": "component",
"label": "Release Public Wording Gate",
"relation": "explains",
"url": "component-batch12-release-claim-language-gate.html"
},
{
"id": "mechanism.batch12_release_claim_language_gate.validates_public_release_claim_language_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public launch claim language gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch12-release-claim-language-gate-validates-public-release-claim-language-gate"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.public_reveal_walkthrough",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Public Reveal Walkthrough",
"relation": "depends on",
"url": "paper-module-public-reveal-walkthrough.html"
},
{
"id": "paper_module.proof_derived_governed_mutation_authorization",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof-Derived Governed Mutation Authorization",
"relation": "depends on",
"url": "paper-module-proof-derived-governed-mutation-authorization.html"
},
{
"id": "paper_module.batch8_validator_checker_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Validator Checker Bundle",
"relation": "depends on",
"url": "paper-module-batch8-validator-checker-capsule.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.batch12_release_claim_language_gate",
"reader_packet": {
"abstract": "Set 12 launch claim-Language Gate checks public claim language against result record-backed scope limits without approving launch or public sharing.",
"authority_ceiling": "Fixture-bound launch-claim language gate evidence and copied source refs only; no launch-scope decision, publishing-scope decision, external model access, whole-system equivalence, market-level conclusions, investment-related actions, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch12_release_claim_language_gate.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch12-release-claim-language-gate-diagram-1",
"html_url": "docs/paper-module-batch12-release-claim-language-gate.html#paper-module-batch12-release-claim-language-gate-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Fixture[\"release_gate_fixture.json safe and active copy text\"] Write[\"_write_gate_fixture docs files + publication_manifest.yaml\"] Gate[\"build_gate scan manifest lines for risky phrases\"] Classify{\"_classify read each hit in context\"} Boundary[\"boundary_or_negative_context negation marker present => allowed\"] Active[\"active_claim_blocker bare launch or internal control wording => blocked\"] Assert[\"main --assert-clear exit 2 when any active blocker remains\"] Result[\"metadata-only result record result, board, validation\"] Fixture --> Write Write --> Gate Gate --> Classify Classify -->|allowed| Boundary Classify -->|blocked| Active Active --> Assert Boundary --> Result Assert --> Result",
"source_format": "mermaid",
"source_hash": "sha256:7cd0ba6f377c1ee7b52f50edb1c28dbb233bff7b4be9a046b7c729c504bfba5f",
"source_line_end": 89,
"source_line_start": 71
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.public_reveal_walkthrough",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Public Reveal Walkthrough",
"relation": "depends on",
"url": "paper-module-public-reveal-walkthrough.html"
},
{
"id": "paper_module.proof_derived_governed_mutation_authorization",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof-Derived Governed Mutation Authorization",
"relation": "depends on",
"url": "paper-module-proof-derived-governed-mutation-authorization.html"
},
{
"id": "paper_module.batch8_validator_checker_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Validator Checker Bundle",
"relation": "depends on",
"url": "paper-module-batch8-validator-checker-capsule.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch12_release_claim_language_gate",
"kind": "organ",
"kind_label": "component",
"label": "Release Public Wording Gate",
"relation": "explains",
"url": "component-batch12-release-claim-language-gate.html"
},
{
"id": "mechanism.batch12_release_claim_language_gate.validates_public_release_claim_language_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public launch claim language gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch12-release-claim-language-gate-validates-public-release-claim-language-gate"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch12_release_claim_language_gate",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 135,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch12_release_claim_language_gate.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "Public copy is scanned line by line for launch, licensing, maturity, and launch-authority wording it cannot support, failing non-zero under --assert-clear while any unsupported claim survives. This component checks public copy for language that claims more than the result records behind it can support. It reads a small public sharing manifest, opens each file the manifest lists, and matches every line against fixed families of launch, licensing, maturity, and launch-authority wording. It writes a metadata-only result record and, when run with --assert-clear, exits non-zero while any unsupported claim remains. It does not approve launch and it does not publish anything.",
"level": 4,
"order": 1,
"section_id": "release-language-is-checked-against-evidence-classes",
"source_line_end": 6,
"source_line_start": 3,
"title": "launch language is checked against evidence classes"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The private gate module is absent on the public clone, so an in-file PublicFallbackReleaseClaimGate stands in, making this a check on a checker over shipped fixtures. The gate logic it exercises is a separate module, releaseclaimlanguagegate.py. On the public clone that module is not present, so the component replaces it with an in-file stand-in, _PublicFallbackReleaseClaimGate. The page you can run here is that fallback gate driven over shipped fixtures. This module is therefore a check on a checker: it proves the gate blocks bare launch wording and clears explicit boundary wording over a fixed set of inputs.",
"level": 4,
"order": 2,
"section_id": "the-gate-separates-supported-wording-from-overclaim",
"source_line_end": 10,
"source_line_start": 7,
"title": "The gate separates supported wording from overclaim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The scan is a deterministic regular-expression pass that fails closed: any affirmative claim it cannot downgrade becomes an active blocker and forces exit 2. The scan is a deterministic regular-expression pass over lines, not a language model. It fails closed. An affirmative claim it cannot downgrade becomes an active blocker, and main --assert-clear returns exit 2 whenever an active blocker survives.",
"level": 4,
"order": 3,
"section_id": "passing-prose-does-not-authorize-release",
"source_line_end": 14,
"source_line_start": 11,
"title": "Passing prose excludes launch"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Public copy drifts toward over-claiming as edits accumulate, and human review misses it because each phrase reads reasonably alone, so the gate asks whether a line's launch status outran the evidence. Public copy drifts toward over-claiming. A page that began as \"fixture-proven, not yet published\" gets edited over months until someone writes launch, licensing, or maturity wording without noticing that nothing underneath it changed. A loose human review misses this because each phrase reads reasonably in isolation. The gate answers one narrow question: does a line assert launch status that the evidence does not hold, and would an automated check catch it before the copy ships.",
"level": 4,
"order": 5,
"section_id": "how-copy-drifts-into-over-claiming",
"source_line_end": 20,
"source_line_start": 17,
"title": "How copy drifts into over-claiming"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Being lexical, the check flags risky phrasing and its excusing context but never judges whether the underlying claim is true, only whether the wording outran the records. Because the check is lexical, it is honest about its own reach. It finds risky phrasing and the context that would excuse it. It does not decide whether the underlying claim is true, only whether the wording has outrun what the result records can back.",
"level": 4,
"order": 6,
"section_id": "a-lexical-check-honest-about-its-reach",
"source_line_end": 24,
"source_line_start": 21,
"title": "A lexical check honest about its reach"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The stages run in this order inside _evaluate.",
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 28,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "loadfixture reads releasegatefixture.json, turning a missing file, bad JSON, or non-object payload into a structured finding and the fail-closed envelope rather than a quiet pass. First, loadfixture reads releasegatefixture.json from the selected input directory. A missing file, invalid JSON, or a non-object payload becomes a structured finding, and the run returns the fail-closed envelope from blocked_exercise rather than a quiet pass.",
"level": 4,
"order": 8,
"section_id": "load-fixture-reads-or-fails-closed",
"source_line_end": 32,
"source_line_start": 29,
"title": "`_load_fixture` reads or fails closed"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "loadsourcemodule returns PublicFallbackReleaseClaimGate when the manifest marks the public stub, still validating the copied module's digest and required anchors before any exercise. Second, loadsourcemodule resolves the gate. sourcemodulepublicstubbed inspects the copied-module manifest, and when it sees the public stub substitution it returns PublicFallbackReleaseClaimGate instead of importing a private body. The copied module is still validated for digest equality and required anchors (RISKYPHRASES, NEGATIVECONTEXTMARKERS, METAFORBIDDENCONTEXTMARKERS, privatecontrolplaneleak, def classifyhit, def build_gate) before any exercise.",
"level": 4,
"order": 9,
"section_id": "load-source-module-returns-the-stub",
"source_line_end": 36,
"source_line_start": 33,
"title": "`_load_source_module` returns the stub"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "writegatefixture writes each scenario's docs/ text and a publicationmanifest.yaml, while fixturedocname rejects control characters, path separators, and duplicate document names. Third, writegatefixture materialises each scenario on disk: it writes the fixture text into docs/, then writes a publicationmanifest.yaml that lists those files. fixturedocname rejects any file name with control characters (BATCH12RELEASEFIXTUREPATHCONTROLCHAR) or path separators (BATCH12RELEASEFIXTUREPATHUNSAFE), and rejects an active document that would overwrite the safe one (BATCH12RELEASEFIXTUREDUPLICATEDOC_NAME).",
"level": 4,
"order": 10,
"section_id": "write-gate-fixture-materialises-each-scenario",
"source_line_end": 40,
"source_line_start": 37,
"title": "`_write_gate_fixture` materialises each scenario"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildgate opens each manifest file and scans every line, recording each match as a metadata-only hit and counting active blockers, boundary hits, and claim surfaces into a status. Fourth, buildgate reads the manifest, opens each listed file, and scans every line against the phrase patterns. Each match is classified by classify and recorded as a hit carrying its path, line number, phrase id, classification, and a bodyin_receipt: false flag. The summary counts active blockers, boundary hits, and claim surfaces, and sets a status.",
"level": 4,
"order": 11,
"section_id": "build-gate-scans-and-classifies-lines",
"source_line_end": 44,
"source_line_start": 41,
"title": "`build_gate` scans and classifies lines"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "runmainassertclear runs the gate on the active scenario and returns its exit code, after which evaluate computes the cases and hands off to the metadata-only writers. Fifth, runmainassertclear runs main --assert-clear on the active scenario, captures the printed status, and returns the exit code. evaluate then computes the negative cases and the positive boundary-clear check and hands the result to run or runbatch12releaseclaimlanguagegate_bundle, which write the metadata-only records.",
"level": 4,
"order": 12,
"section_id": "run-main-assert-clear-captures-the-exit-code",
"source_line_end": 58,
"source_line_start": 45,
"title": "`_run_main_assert_clear` captures the exit code"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "classify blocks any internal control-leak phrase outright, allows other hits only when a negation marker is present, and holds internal control vocabulary to a stricter rule that even a negation cannot clear. The interesting part is classify. It reads a matched line two ways. If the phrase belongs to the privatecontrolplaneleak family, the line is an active blocker regardless of surrounding words. Otherwise, if the line contains a negative-context marker such as not , forbidden, blocked, or omitted, the hit is boundary-or-negative context and is allowed; with no such marker it is an active blocker. So \"not a hosted product\" clears and \"deployment-posture\" blocks. The internal control family\u2026",
"level": 4,
"order": 13,
"section_id": "classify-reads-each-hit-in-context",
"source_line_end": 62,
"source_line_start": 59,
"title": "`_classify` reads each hit in context"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Status is set from the counts: any active blocker yields activeclaimblocked, and ok is true only at zero blockers, which is when --assert-clear avoids exit 2. buildgate then sets the status from the counts. Any active blocker gives activeclaimblocked; boundary hits with no active blocker give clearboundary_only; nothing risky gives clear. The ok flag is true only at zero active blockers, and main --assert-clear returns 2 when ok is false.",
"level": 4,
"order": 14,
"section_id": "status-and-the-ok-flag-from-counts",
"source_line_end": 66,
"source_line_start": 63,
"title": "Status and the `ok` flag from counts"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "evaluate runs four scenarios from one fixture, expecting the safe boundary document to clear and the three over-claim documents to block. _evaluate exercises four scenarios from one fixture: a safe boundary document, an affirmative over-claim document, a public sharing over-claim document, and a private internal control leak document. The safe scenario is expected to clear as boundary-only. The other three are expected to block.",
"level": 4,
"order": 15,
"section_id": "four-scenarios-one-expected-to-clear",
"source_line_end": 90,
"source_line_start": 67,
"title": "Four scenarios, one expected to clear"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "negative-cases",
"source_line_end": 92,
"source_line_start": 91,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVECASES fixes three error-coded cases covering the affirmative blocker, the exit-2 assertion, and the internal control leak, blocking the run if any fails to compute. EXPECTEDNEGATIVECASES fixes three cases, each with a stable error code. affirmativeopensourceproductionreadyblocks (BATCH12RELEASECLAIMACTIVEBLOCKER) requires the active scenario to reach activeclaimblocked with both the opensource and productionready phrase ids present. assertclearreturnsexit2 (BATCH12RELEASECLAIMASSERTCLEAREXIT2) requires main --assert-clear to return exit code 2 with ok false. privatecontrolplanepublicreaderblocks (BATCH12RELEASEPRIVATECONTROLPLANELEAK) requires the internal control scenario t\u2026",
"level": 4,
"order": 17,
"section_id": "three-fixed-cases-each-recomputed",
"source_line_end": 96,
"source_line_start": 93,
"title": "Three fixed cases, each recomputed"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "prior-art-grounding",
"source_line_end": 98,
"source_line_start": 97,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The proof-consumer form comes from advertising-substantiation and regulated-communication practice, where FTC, FINRA, and SEC rules require a reasonable basis before a claim is disseminated. The pattern comes from advertising-substantiation and regulated-communication practice: public claims should stay within the evidence actually held, and stronger wording needs stronger support. The FTC advertising-substantiation policy (https://www.ftc.gov/legal-library/browse/ftc-policy-statement-regarding-advertising-substantiation) states that objective claims need a reasonable basis before dissemination. FINRA Rule 2210 and the SEC investment-adviser marketing rule carry the same shape for financial\u2026",
"level": 4,
"order": 19,
"section_id": "advertising-substantiation-and-communication-rules",
"source_line_end": 102,
"source_line_start": 99,
"title": "Advertising-substantiation and communication rules"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Reader-verifiable commands, run from the microcosm-substrate/ public root:",
"level": 2,
"order": 20,
"section_id": "validation-receipt-path",
"source_line_end": 120,
"source_line_start": 103,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass certifies the fallback gate cleared the boundary scenario, blocked the three over-claim scenarios, exited 2 under --assert-clear, and wrote metadata-only records. A pass means the fallback gate cleared the boundary scenario, blocked the three over-claim scenarios, returned exit 2 under --assert-clear, and wrote the result, board, and validation records with no copied body text inside them.",
"level": 4,
"order": 21,
"section_id": "what-a-pass-certifies",
"source_line_end": 124,
"source_line_start": 121,
"title": "What a pass certifies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 126,
"source_line_start": 125,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Over the shipped fixtures the fallback gate allows boundary-only wording, blocks bare launch, licensing, maturity, and launch-authority wording, and exits 2 while a blocker remains. The strongest claim the evidence supports is narrow: over the shipped fixtures, the fallback gate classifies boundary-only wording as allowed and classifies bare launch, licensing, maturity, and launch-authority wording as active blockers, and main --assert-clear exits 2 while any active blocker remains. That is the proof boundary. The evidence is lexical and fixture-bound.",
"level": 4,
"order": 23,
"section_id": "the-lexical-fixture-bound-proof-boundary",
"source_line_end": 130,
"source_line_start": 127,
"title": "The lexical, fixture-bound proof boundary"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A green result grants no launch, public sharing, or hosted-product status, proves no semantic understanding or legal sufficiency, and stays evidence about the checker rather than permission to ship. The scope limit stops well short of that. A green result does not grant launch permission, public sharing, or hosted-product status. It does not establish semantic understanding of the copy, complete secret detection, whole-system equivalence, market-level conclusions, legal sufficiency, or whole-system correctness. It excludes editing the copied gate module or changing the fixtures, negative cases, or ceiling; those belong in the source and standard lanes first. A pass is evidence about the che\u2026",
"level": 4,
"order": 24,
"section_id": "a-pass-is-evidence-not-permission-to-ship",
"source_line_end": 134,
"source_line_start": 131,
"title": "A pass is evidence, not permission to ship"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch12_release_claim_language_gate.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 12 launch claim-Language Gate checks public claim language against result record-backed scope limits without approving launch or public sharing."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch12_release_claim_language_gate.py",
"path": "src/microcosm_core/organs/batch12_release_claim_language_gate.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for launch-claim fixture generation, claim-language blocking, main assert-clear checks, negative cases, and scope limits.",
"symbols": [
"_blocked_exercise",
"_write_gate_fixture",
"_run_main_assert_clear",
"_evaluate",
"run",
"run_batch12_release_claim_language_gate_bundle",
"result_card",
"EXPECTED_NEGATIVE_CASES"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch12_release_claim_language_gate.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch12_release_claim_language_gate.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch12_release_claim_language_gate.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 12 launch claim-Language Gate runs launch claim-language over public fixtures. It classifies phrases against evidence class and scope limit, checks typed ordinal evidence ranks, real-system flags, fail-closed defaults, boundary-context negation, main --assert-clear behavior, source digests, negative cases, and scope limits while excluding launch-scope decision, external model access, whole-system equivalence, market-level conclusions, investment-related actions, public sharing, and whole-system correctness.",
"summary_line": "Set 12 launch claim-Language Gate checks public claim language against result record-backed scope limits without approving launch or public sharing.",
"title": "Set 12 launch claim-Language Gate"
},
{
"anchor": "paper-module-batch10-cold-eval-honesty-capsule",
"authority_ceiling": "Fixture-bound route-quality scorecard and copied source refs only; no live benchmark, navigation truth, hosted readiness, launch-scope decision, external model access, source-file changes, or whole-system correctness.",
"body_html": "
One copied simulator, one bounded scorecard
\n
It runs one real route-quality simulator, an exact copy of cold_eval.py, over a fixed public workspace and reports the scorecard without letting it become a claim about the live system.batch10_cold_eval_honesty_capsule takes one real route-quality simulator, runs it over a fixed public workspace, and reports the scorecard without letting that scorecard become a claim about the live system. The simulator is an exact copy of cold_eval.py, imported from the exported bundle rather than rewritten, so the numbers come from the same code the original produced.
\n
Four assertions without a wider win claim
\n
A green run asserts four things: the copied simulator ran, every task went to the idea-first arm, its extra route surface shows in the counts, and the forbidden claims stay forbidden. A green run says four things happened: the copied simulator executed, the recomputed scorecard has every task won by the idea-first arm, the extra route surface that arm reaches is visible in the counts, and the exercise manifest still forbids benchmark, navigation-truth, launch, and hosted-readiness claims. It does not say idea-first routing wins anywhere except this workspace.
\n
Counts and statuses without source bodies
\n
The result is a small metadata-only record carrying counts, engine statuses, and a scoring-policy name, but no source text or command output. The result is a small metadata-only record. It carries counts, engine statuses, and a scoring-policy name. It does not carry any source text or command output.
\n
Purpose
\n
A clean-sweep scorecard is easy to fake
\n
A scorecard where one arm wins every task is easy to fake, so this component runs the real simulator and checks none of those shortcuts happened before treating the win as accounting. A scorecard that shows one arm winning every task is easy to fake. You can hand the winning arm the answer, inject the expected references into its route, or quietly let a non-public paths leak into the fixture and call the shape "measured". This component exists to run the real simulator and then check that none of those shortcuts happened, so the win is an accounting fact about a declared fixture and nothing more.
\n
Expected references score coverage, never the route
\n
The honesty turns on one property: expected references only ever score coverage and are never added to either arm's route, so idea-first wins by genuinely declaring more relevant files. The honest part turns on one property of the copied scorer. Each task lists the route references an answer should reach, but those expected references are only ever used to score coverage. They are never added to either arm's route. The flat arm is scored on what a reader reaches from README.md, docs/quickstart.md, and pyproject.toml. The idea-first arm is scored on what its entry packets declare. The idea-first arm wins because it genuinely declares more of the relevant files, not because the scorer leaked the target into the route.
\n
How it works
\n
Both entrypoints share one evaluator
\n
Both entrypoints hand their fixture to run_crown_jewel_organ, where _evaluate turns each blocked engine into an explicit finding instead of a silent success.run and run_batch10_cold_eval_bundle both hand the fixture to run_crown_jewel_organ with _evaluate as the evaluator. _evaluate runs three engines in order and turns any non-pass into a structured finding, so a blocked engine surfaces as an error code rather than a silent success.
\n
Function
Engine id
What it does
_run_original_cold_eval
cold_eval_original_runner
Loads the copied cold_eval.py, copies the workspace to a temporary directory, calls its run_cold_eval at a fixed timestamp, then reads back the scorecard and result record.
_scorecard_shape_audit
cold_eval_scorecard_shape_audit
Checks the winner shape and route-count asymmetry, and scans the tasks file for private references.
_claim_ceiling_gate
cold_eval_claim_ceiling_gate
Checks that the exercise manifest forbids the named claims and keeps expected-reference injection off.
\n
The copied simulator runs at a fixed time
\n
_run_original_cold_eval requires the task file, runs the copied simulator at a fixed timestamp, and reads the scorecard and result record back from a temporary workspace._run_original_cold_eval first requires evals/cold_agent_ab/tasks.json in the workspace. Without it the engine returns blocked with BATCH10_COLD_EVAL_TASKS_REQUIRED and never runs the simulator. With it, the engine copies the workspace, calls run_cold_eval at 2026-05-31T00:00:00Z so the run is deterministic, and reads runs/cold_agent_ab/seed_scorecard.json plus receipts/cold_agent_ab_seed.json back from the temporary copy. It emits the task count, the idea-first win count, the flat win count, the per-task winners, and the scoring-policy name. In the shipped fixture the idea-first win count is 3.
\n
The scorecard audit rejects route and path leaks
\n
_scorecard_shape_audit accepts only an all-idea-first result with visible route-count asymmetry and no private reference in the task fixture._scorecard_shape_audit reads the same tasks file and walks its strings for any value containing a user-home prefix or the private source-repository root. It passes only when every winner is B.idea_first_packet, the idea-first route reference count is greater than the flat count, and no private reference was found. It records the two counts and the route_surface_asymmetry_visible flag so a reader can see the gap that produced the win.
\n
The ceiling gate blocks promoted claims
\n
_claim_ceiling_gate requires expected-reference injection to stay off and the exercise manifest to keep all four promoted claims explicitly forbidden._claim_ceiling_gate loads batch10_cold_eval_exercise_manifest.json and passes only when at least one row was checked, expected_ref_injection_allowed is false, and the manifest's forbidden-claim list contains live_agent_benchmark_win, navigation_truth, release_approval, and hosted_public_readiness. This is where the run refuses to promote the shape.
\n
The source manifest is checked before execution
\n
Before any engine runs, the shared runner verifies the copied simulator's digest and required anchors so a tampered source copy cannot produce a passing scorecard. The source copy is checked before any of this. The shared runner validates the bundle's source_module_manifest.json against the copied file: the digest must match and the required anchors (SCORING_POLICY =, def _score_task(, def run_cold_eval(, expected_ref_injection_used, idea_first_packet_wins_fixture) must be present. A tampered digest fails with CROWN_JEWEL_SOURCE_DIGEST_MISMATCH before the engines run.
\nDiagram of the mechanism (6 steps).Diagram source & refs
Source refs
run copied run_cold_eval, read back scorecard
cold_eval_original_runner
all-B winner + route asymmetry + no non-public refs
cold_eval_scorecard_shape_audit
injection off, forbidden claims named
cold_eval_claim_ceiling_gate
flowchart TD A["Fixture or exported bundle (workspace + exercise manifest)"] --> M["Source manifest check digest + required anchors"] M --> R["cold_eval_original_runner run copied run_cold_eval, read back scorecard"] R --> S["cold_eval_scorecard_shape_audit all-B winner + route asymmetry + no non-public refs"] S --> C["cold_eval_claim_ceiling_gate injection off, forbidden claims named"] C --> O["metadata-only result record and card"]
\n
The compact card preserves the same result
\n
result_card adds engine and copied-module counts to the full record, while bundle mode follows the same evaluator path with exported inputs.result_card projects a compact card from the full record and adds the engine count and the copied source-module count. run_batch10_cold_eval_bundle is the same path with the exported-bundle input mode.
\n
Negative cases
\n
Four perturbations must move the verdict
\n
Each negative fixture mutates one input, reruns the same evaluator path, and must turn the result blocked with its named error code. The fixture ships four perturbations. Each one copies the input, applies a named mutation, re-runs the engines, and confirms the run turns blocked with the right code.
\n
Missing tasks, route ties, injection, and non-public refs
\n
The four mutations remove tasks, erase route asymmetry, enable expected-ref injection, or add a non-public paths, and each has a distinct required rejection.missing_tasks removes the tasks file, so _run_original_cold_eval returns BATCH10_COLD_EVAL_TASKS_REQUIRED. flat_route_can_win rewrites the route counts to be equal, so the shape audit loses its asymmetry and reports BATCH10_COLD_EVAL_NOT_ALWAYS_B_WIN. expected_ref_injection sets expected_ref_injection_allowed to true in the manifest, and the scope limit gate rejects it with BATCH10_COLD_EVAL_EXPECTED_REF_INJECTION_FORBIDDEN. private_fixture_ref injects a /Users/ probe into the tasks file, and the shape audit rejects it with BATCH10_COLD_EVAL_PRIVATE_REF_REJECTED. Each mutation runs through the same manifest and evaluator path as the positive case, so a perturbation that fails to move the outcome is itself a failure.
\n
Prior Art Grounding
\n
Evaluation reports need scenario and limit context
\n
HELM and Model Cards motivate reporting what was measured, under which fixture assumptions, and which wider capability claims the score cannot support. This component follows evaluation-transparency and benchmark-hygiene practice: a scorecard should state what was measured, what fixture assumptions went in, and what the result can and cannot support. HELM frames evaluation as a scenario-bound surface rather than a single global capability claim. Model Cards for Model Reporting established pairing performance numbers with intended use and limitations. This component borrows that scorecard-plus-limits shape and narrows it to a deterministic route-quality fixture.
\n
Validation Result record Path
\n
Reader-verifiable commands, run from the microcosm-substrate/ public root:
\n
A pass covers three engines and four rejections
\n
The commands prove the copied simulator and all three engines ran under the declared fixture while every negative case blocked with its expected code. The fixture command writes the scorecard record and sign-off JSON. The bundle command validates the copied source, its manifest, the metadata-only cards, the injection policy, and the private-reference negative case. A pass means the copied simulator ran, the three engines passed, and the four negative cases each turned blocked with the expected code.
\n
Scope boundary
\n
Scope limit
\n
The fixture supports one exact accounting claim
\n
The strongest claim is that this copied simulator produced the declared all-idea-first scorecard without expected-ref injection or private-path leakage, while the manifest kept promoted claims forbidden. The strongest claim here is small and exact. Over this fixture, the copied cold_eval.py ran, its scorecard was recomputed with every task won by the idea-first arm, expected-reference injection stayed off, no private reference leaked in, and the manifest still forbade the promoted claims. That is the whole proof boundary.
\n
The workspace result grants no live authority
\n
The scope limit excludes live benchmarks, navigation truth, hosted readiness, route superiority, external model access, source-file changes, launch, and whole-system correctness. The scope limit stops there. This component does not claim live benchmark results, navigation truth, hosted readiness, route-quality superiority, external model access, source-file changes, launch-scope decision, or whole-system correctness. The fixture win is accounting for one declared workspace, not evidence that idea-first routing wins in the live system. Validation result records back the claim above and nothing beyond it.
\n",
"components": [
{
"anchor": "component-batch10_cold_eval_honesty_capsule",
"component_id": "batch10_cold_eval_honesty_capsule",
"public_label": "Cold Evaluation Honesty Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 14,
"links": [
{
"id": "batch10_cold_eval_honesty_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Cold Evaluation Honesty Bundle",
"relation": "explains",
"url": "component-batch10-cold-eval-honesty-capsule.html"
},
{
"id": "mechanism.batch10_cold_eval_honesty_capsule.validates_public_cold_eval_honesty_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public cold eval honesty bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch10-cold-eval-honesty-capsule-validates-public-cold-eval-honesty-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
},
{
"id": "paper_module.agent_benchmark_integrity_anti_gaming_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Benchmark Integrity Anti-Gaming Replay",
"relation": "depends on",
"url": "paper-module-agent-benchmark-integrity-anti-gaming-replay.html"
},
{
"id": "paper_module.research_replication_rubric_artifact_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Research Replication Rubric Artifact Replay",
"relation": "depends on",
"url": "paper-module-research-replication-rubric-artifact-replay.html"
},
{
"id": "paper_module.batch8_validator_checker_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Validator Checker Bundle",
"relation": "depends on",
"url": "paper-module-batch8-validator-checker-capsule.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.batch10_cold_eval_honesty_capsule",
"reader_packet": {
"abstract": "Set 10 Cold Eval Honesty Bundle runs cold-eval over public fixtures without benchmark, navigation-truth, or launch-scope decision.",
"authority_ceiling": "Fixture-bound route-quality scorecard and copied source refs only; no live benchmark, navigation truth, hosted readiness, launch-scope decision, external model access, source-file changes, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch10_cold_eval_honesty_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch10-cold-eval-honesty-capsule-diagram-1",
"html_url": "docs/paper-module-batch10-cold-eval-honesty-capsule.html#paper-module-batch10-cold-eval-honesty-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Fixture or exported bundle (workspace + exercise manifest)\"] --> M[\"Source manifest check digest + required anchors\"] M --> R[\"cold_eval_original_runner run copied run_cold_eval, read back scorecard\"] R --> S[\"cold_eval_scorecard_shape_audit all-B winner + route asymmetry + no non-public refs\"] S --> C[\"cold_eval_claim_ceiling_gate injection off, forbidden claims named\"] C --> O[\"metadata-only result record and card\"]",
"source_format": "mermaid",
"source_hash": "sha256:32af93306fedb4244c67ff7f648853a1b0344fd5062235aff7836668e6ae2d8c",
"source_line_end": 133,
"source_line_start": 126
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.agent_benchmark_integrity_anti_gaming_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Benchmark Integrity Anti-Gaming Replay",
"relation": "depends on",
"url": "paper-module-agent-benchmark-integrity-anti-gaming-replay.html"
},
{
"id": "paper_module.research_replication_rubric_artifact_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Research Replication Rubric Artifact Replay",
"relation": "depends on",
"url": "paper-module-research-replication-rubric-artifact-replay.html"
},
{
"id": "paper_module.batch8_validator_checker_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Validator Checker Bundle",
"relation": "depends on",
"url": "paper-module-batch8-validator-checker-capsule.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch10_cold_eval_honesty_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Cold Evaluation Honesty Bundle",
"relation": "explains",
"url": "component-batch10-cold-eval-honesty-capsule.html"
},
{
"id": "mechanism.batch10_cold_eval_honesty_capsule.validates_public_cold_eval_honesty_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public cold eval honesty bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch10-cold-eval-honesty-capsule-validates-public-cold-eval-honesty-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "concept.agent_reliability_and_safety_validator_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "agent reliability and safety validators as bounded public scope limits",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-agent-reliability-and-safety-validator-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch10_cold_eval_honesty_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 238,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch10_cold_eval_honesty_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "HELM",
"section_id": "evaluation-reports-need-scenario-and-limit-context",
"url": "https://crfm.stanford.edu/helm/index.html"
},
{
"kind": "external_reference",
"label": "Model Cards for Model Reporting",
"section_id": "evaluation-reports-need-scenario-and-limit-context",
"url": "https://arxiv.org/abs/1810.03993"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "It runs one real route-quality simulator, an exact copy of coldeval.py, over a fixed public workspace and reports the scorecard without letting it become a claim about the live system. batch10coldevalhonesty_capsule takes one real route-quality simulator, runs it over a fixed public workspace, and reports the scorecard without letting that scorecard become a claim about the live system. The simulator is an exact copy of cold_eval.py, imported from the exported bundle rather than rewritten, so the numbers come from the same code the original produced.",
"level": 4,
"order": 1,
"section_id": "one-copied-simulator-one-bounded-scorecard",
"source_line_end": 13,
"source_line_start": 3,
"title": "One copied simulator, one bounded scorecard"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A green run asserts four things: the copied simulator ran, every task went to the idea-first arm, its extra route surface shows in the counts, and the forbidden claims stay forbidden.** A green run says four things happened: the copied simulator executed, the recomputed scorecard has every task won by the idea-first arm, the extra route surface that arm reaches is visible in the counts, and the exercise manifest still forbids benchmark, navigation-truth, launch, and hosted-readiness claims. It does not say idea-first routing wins anywhere except this workspace.",
"level": 4,
"order": 2,
"section_id": "four-assertions-without-a-wider-win-claim",
"source_line_end": 24,
"source_line_start": 14,
"title": "Four assertions without a wider win claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The result is a small metadata-only record carrying counts, engine statuses, and a scoring-policy name, but no source text or command output.** The result is a small metadata-only record. It carries counts, engine statuses, and a scoring-policy name. It does not carry any source text or command output.",
"level": 4,
"order": 3,
"section_id": "counts-and-statuses-without-source-bodies",
"source_line_end": 31,
"source_line_start": 25,
"title": "Counts and statuses without source bodies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 33,
"source_line_start": 32,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A scorecard where one arm wins every task is easy to fake, so this component runs the real simulator and checks none of those shortcuts happened before treating the win as accounting.** A scorecard that shows one arm winning every task is easy to fake. You can hand the winning arm the answer, inject the expected references into its route, or quietly let a non-public paths leak into the fixture and call the shape \"measured\". This component exists to run the real simulator and then check that none of those shortcuts happened, so the win is an accounting fact about a declared fixture and nothing more.",
"level": 4,
"order": 5,
"section_id": "a-clean-sweep-scorecard-is-easy-to-fake",
"source_line_end": 45,
"source_line_start": 34,
"title": "A clean-sweep scorecard is easy to fake"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The honesty turns on one property: expected references only ever score coverage and are never added to either arm's route, so idea-first wins by genuinely declaring more relevant files.** The honest part turns on one property of the copied scorer. Each task lists the route references an answer should reach, but those expected references are only ever used to score coverage. They are never added to either arm's route. The flat arm is scored on what a reader reaches from README.md, docs/quickstart.md, and pyproject.toml. The idea-first arm is scored on what its entry packets declare. The idea-first arm wins because it genuinely declares more of the relevant files, not because the scorer lea\u2026",
"level": 4,
"order": 6,
"section_id": "expected-references-score-coverage-never-the-route",
"source_line_end": 59,
"source_line_start": 46,
"title": "Expected references score coverage, never the route"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 61,
"source_line_start": 60,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "Both entrypoints hand their fixture to runcrownjewelorgan, where _evaluate turns each blocked engine into an explicit finding instead of a silent success.** run and runbatch10coldevalbundle both hand the fixture to runcrownjewelorgan with evaluate as the evaluator. _evaluate runs three engines in order and turns any non-pass into a structured finding, so a blocked engine surfaces as an error code rather than a silent success.",
"level": 4,
"order": 8,
"section_id": "both-entrypoints-share-one-evaluator",
"source_line_end": 77,
"source_line_start": 62,
"title": "Both entrypoints share one evaluator"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "runoriginalcoldeval requires the task file, runs the copied simulator at a fixed timestamp, and reads the scorecard and result record back from a temporary workspace.** runoriginalcoldeval first requires evals/coldagentab/tasks.json in the workspace. Without it the engine returns blocked with BATCH10COLDEVALTASKSREQUIRED and never runs the simulator. With it, the engine copies the workspace, calls runcoldeval at 2026-05-31T00:00:00Z so the run is deterministic, and reads runs/coldagentab/seed_scorecard.json plus result records/coldagentab_seed.json back from the temporary copy. It emits the task count, the idea-first win count, the flat win count, the per-task winners, and the scoring-polic\u2026",
"level": 4,
"order": 9,
"section_id": "the-copied-simulator-runs-at-a-fixed-time",
"source_line_end": 92,
"source_line_start": 78,
"title": "The copied simulator runs at a fixed time"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "scorecardshape_audit accepts only an all-idea-first result with visible route-count asymmetry and no private reference in the task fixture.** scorecardshape_audit reads the same tasks file and walks its strings for any value containing a user-home prefix or the private source-repository root. It passes only when every winner is B.ideafirstpacket, the idea-first route reference count is greater than the flat count, and no private reference was found. It records the two counts and the routesurfaceasymmetry_visible flag so a reader can see the gap that produced the win.",
"level": 4,
"order": 10,
"section_id": "the-scorecard-audit-rejects-route-and-path-leaks",
"source_line_end": 103,
"source_line_start": 93,
"title": "The scorecard audit rejects route and path leaks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "claimceiling_gate requires expected-reference injection to stay off and the exercise manifest to keep all four promoted claims explicitly forbidden.** claimceilinggate loads batch10coldevalexercise_manifest.json and passes only when at least one row was checked, expectedrefinjection_allowed is false, and the manifest's forbidden-claim list contains liveagentbenchmarkwin, navigationtruth, release_approval, and hostedpublicreadiness. This is where the run refuses to promote the shape.",
"level": 4,
"order": 11,
"section_id": "the-ceiling-gate-blocks-promoted-claims",
"source_line_end": 113,
"source_line_start": 104,
"title": "The ceiling gate blocks promoted claims"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "**Before any engine runs, the shared runner verifies the copied simulator's digest and required anchors so a tampered source copy cannot produce a passing scorecard.** The source copy is checked before any of this. The shared runner validates the bundle's sourcemodulemanifest.json against the copied file: the digest must match and the required anchors (SCORINGPOLICY =, def score_task(, def runcoldeval(, expectedrefinjection_used, ideafirstpacketwinsfixture) must be present. A tampered digest fails with CROWNJEWELSOURCEDIGESTMISMATCH before the engines run.",
"level": 4,
"order": 12,
"section_id": "the-source-manifest-is-checked-before-execution",
"source_line_end": 134,
"source_line_start": 114,
"title": "The source manifest is checked before execution"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "resultcard adds engine and copied-module counts to the full record, while bundle mode follows the same evaluator path with exported inputs.** result_card projects a compact card from the full record and adds the engine count and the copied source-module count. runbatch10coldevalbundle is the same path with the exported-bundle input mode.",
"level": 4,
"order": 13,
"section_id": "the-compact-card-preserves-the-same-result",
"source_line_end": 142,
"source_line_start": 135,
"title": "The compact card preserves the same result"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 144,
"source_line_start": 143,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Each negative fixture mutates one input, reruns the same evaluator path, and must turn the result blocked with its named error code.** The fixture ships four perturbations. Each one copies the input, applies a named mutation, re-runs the engines, and confirms the run turns blocked with the right code.",
"level": 4,
"order": 15,
"section_id": "four-perturbations-must-move-the-verdict",
"source_line_end": 152,
"source_line_start": 145,
"title": "Four perturbations must move the verdict"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The four mutations remove tasks, erase route asymmetry, enable expected-ref injection, or add a non-public paths, and each has a distinct required rejection.** missingtasks removes the tasks file, so runoriginalcold_eval returns BATCH10COLDEVALTASKSREQUIRED. flatroutecan_win rewrites the route counts to be equal, so the shape audit loses its asymmetry and reports BATCH10COLDEVALNOTALWAYSBWIN. expectedrefinjection sets expectedrefinjection_allowed to true in the manifest, and the scope limit gate rejects it with BATCH10COLDEVALEXPECTEDREFINJECTIONFORBIDDEN. shape audit rejects it with BATCH10COLDEVALPRIVATEREF_REJECTED. Each mutation runs through the same manifest and evaluator path as the\u2026",
"level": 4,
"order": 16,
"section_id": "missing-tasks-route-ties-injection-and-private-refs",
"source_line_end": 167,
"source_line_start": 153,
"title": "Missing tasks, route ties, injection, and non-public refs"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 169,
"source_line_start": 168,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**HELM and Model Cards motivate reporting what was measured, under which fixture assumptions, and which wider capability claims the score cannot support.** This component follows evaluation-transparency and benchmark-hygiene practice: a scorecard should state what was measured, what fixture assumptions went in, and what the result can and cannot support. HELM frames evaluation as a scenario-bound surface rather than a single global capability claim. Model Cards for Model Reporting established pairing performance numbers with intended use and limitations. This component borrows that scorecard-plus-limits shape and narrows it to a deterministic route-quality fixture.",
"level": 4,
"order": 18,
"section_id": "evaluation-reports-need-scenario-and-limit-context",
"source_line_end": 184,
"source_line_start": 170,
"title": "Evaluation reports need scenario and limit context"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Reader-verifiable commands, run from the microcosm-substrate/ public root:",
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 202,
"source_line_start": 185,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The commands prove the copied simulator and all three engines ran under the declared fixture while every negative case blocked with its expected code.** The fixture command writes the scorecard record and sign-off JSON. The bundle command validates the copied source, its manifest, the metadata-only cards, the injection policy, and the private-reference negative case. A pass means the copied simulator ran, the three engines passed, and the four negative cases each turned blocked with the expected code.",
"level": 4,
"order": 20,
"section_id": "a-pass-covers-three-engines-and-four-rejections",
"source_line_end": 212,
"source_line_start": 203,
"title": "A pass covers three engines and four rejections"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 214,
"source_line_start": 213,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is that this copied simulator produced the declared all-idea-first scorecard without expected-ref injection or private-path leakage, while the manifest kept promoted claims forbidden.** The strongest claim here is small and exact. Over this fixture, the copied cold_eval.py ran, its scorecard was recomputed with every task won by the idea-first arm, expected-reference injection stayed off, no private reference leaked in, and the manifest still forbade the promoted claims. That is the whole proof boundary.",
"level": 4,
"order": 22,
"section_id": "the-fixture-supports-one-exact-accounting-claim",
"source_line_end": 225,
"source_line_start": 215,
"title": "The fixture supports one exact accounting claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The scope limit excludes live benchmarks, navigation truth, hosted readiness, route superiority, external model access, source-file changes, launch, and whole-system correctness.** The scope limit stops there. This component does not claim live benchmark results, navigation truth, hosted readiness, route-quality superiority, external model access, source-file changes, launch-scope decision, or whole-system correctness. The fixture win is accounting for one declared workspace, not evidence that idea-first routing wins in the live system. Validation result records back the claim above and nothing beyond it.",
"level": 4,
"order": 23,
"section_id": "the-workspace-result-grants-no-live-authority",
"source_line_end": 237,
"source_line_start": 226,
"title": "The workspace result grants no live authority"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch10_cold_eval_honesty_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 10 Cold Eval Honesty Bundle runs cold-eval over public fixtures without benchmark, navigation-truth, or launch-scope decision."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch10_cold_eval_honesty_capsule.py",
"path": "src/microcosm_core/organs/batch10_cold_eval_honesty_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for copied cold-eval execution, scorecard shape auditing, expected-ref policy checks, negative cases, and scope limits.",
"symbols": [
"_run_original_cold_eval",
"_scorecard_shape_audit",
"_claim_ceiling_gate",
"_evaluate",
"run",
"run_batch10_cold_eval_bundle",
"result_card",
"EXPECTED_NEGATIVE_CASES"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch10_cold_eval_honesty_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch10_cold_eval_honesty_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch10_cold_eval_honesty_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 10 Cold Eval Honesty Bundle imports the real cold_eval.py source body and runs its route-quality simulator over a synthetic public workspace. It audits the all-B idea-first scorecard shape, expected-ref injection policy, private fixture refs, missing-task refusals, negative cases, and scope limits while excluding live benchmark results, navigation truth, hosted readiness, launch-scope decision, external model access, source-file changes, and whole-system correctness.",
"summary_line": "Set 10 Cold Eval Honesty Bundle runs cold-eval over public fixtures without benchmark, navigation-truth, or launch-scope decision.",
"title": "Set 10 Cold Eval Honesty Bundle"
},
{
"anchor": "paper-module-batch10-live-source-drift-capsule",
"authority_ceiling": "Fixture-bound source-digest, anchor, compile, and scope limit evidence only; no route authority, Work or work log mutation, mission execution, git approval, source-file changes, launch, public sharing, external model access, or non-public runtime export.",
"body_html": "
Live-source drift is recomputed across copied inputs
\n
copies of four internal control source files are checked against the source they were copied from. This component checks whether a small set of copies of internal control source files still match the current source they were copied from. The copies are exact Python bodies for four files: the option-surface router, the mission-transaction landing preflight, the work landing controller, and the work log controller.
\n
Three engines separate source, projection, and ceiling
\n
Three questions at once: do the bytes still match the current digest, do the copies still compile, and do they still carry their named surfaces. It answers three questions about those copies at once. Do their bytes still match the current source digest? Do they still compile as Python without being imported or run? Do they still contain the named command and contract surfaces they are supposed to carry? It emits a result record with the digest rows, compile status, anchor checks, and the scope limit, and no copied source body.
\n
A green result remains fixture-bound
\n
It is built to go red: when the source moves ahead of the copies, the failed digest check is the signal to refresh them. The component is built to go red. When the current source moves ahead of the copies, the digest check fails on purpose, and that red result is the signal to refresh the copies through the exact-copy source lane. It is a drift sentinel, not a one-off snapshot.
\n
Purpose
\n
Why a stale copy looks fine
\n
A copy can compile and match one recorded digest while being weeks behind the file it mirrors, so weak checks pass it. Copied source drifts silently. A public copy of a control file can look fine, compile fine, and still be weeks behind the file it claims to mirror. A loose check that only asks "is this valid Python?" passes a stale copy. A check that only compares one recorded digest can pass a copy whose recorded digest was itself never refreshed.
\n
Current and stale digests in one row
\n
Each row holds the current and the earlier stale digest side by side, so proof of freshness and evidence of past drift cannot be erased. This component makes drift visible instead of hiding it. Each digest row keeps two fingerprints side by side: the current source digest, and the stale digest that was recorded earlier. The row passes only when the copied body matches the current digest, and it separately asserts that the stale digest differs from the current one. The proof of freshness and the evidence of the earlier drift sit in the same row, so a refresh cannot quietly erase the fact that the copy was once behind.
\n
How it works
\n
Two entry points into _evaluate
\n
Both entry points forward to run_crown_jewel_organ, and _evaluate loads the manifest and copied texts, then runs four engines in order.run and run_batch10_live_source_drift_bundle are the two entry points. Both forward to run_crown_jewel_organ with the shared SPEC, an _evaluate function, and evaluate_negative_case. _evaluate loads the probe manifest and the copied source texts, then runs four engines in order and collects their findings.
\n
Per-row digest agreement in _digest_refresh_matrix
\n
A row passes only when copied, target, and current digests all agree and the stale digest differs from the current one._digest_refresh_matrix builds one row per copied file. For each row it reads the current source digest and the stale recorded digest from the probe manifest, and the copied and target digests from the source manifest. A row's current_matches_copy is true only when the current digest is non-empty and the copied digest, the target digest, and the current digest all agree and the manifest marks the target as matching. Its stale_differs_from_current is true when a stale digest exists and differs from the current one. The engine rejects a copy whose body no longer matches the current digest, a row whose stale digest was allowed to equal the current one, duplicate rows, missing rows, unexpected rows, and a row count that does not match the four expected files.
\n
Compile without import in _compile_gate
\n
Each copied file is compiled with py_compile into a throwaway directory but never imported, so malformed bodies are caught without running any code._compile_gate compiles each copied file with py_compile.compile into a throwaway temporary directory with doraise=True. It never imports the module, so a malformed body is caught without executing any copied code. import_executed stays false in the result. It rejects a missing target, a non-.py target, and any body that fails to compile.
\n
Named surfaces checked by substring
\n
Substring checks confirm each copy still carries its named command and contract surfaces, so a compiling body that quietly dropped one still fails._control_surface_anchor_matrix runs substring checks over each copied body. It confirms that each file still exposes the command and contract surfaces it is supposed to carry, for example def build_option_surface and option_surface_lens_packet_v0 in the option-surface router, or def cmd_session_finalize and read_receipt_id in the work log controller. A copy that compiles but has quietly lost a command or contract surface still fails here.
\n
Authority flags held false by _claim_ceiling_gate
\n
Eleven authority flags must be false and body_in_receipt false, keeping every copied body a validator input and nothing more._claim_ceiling_gate reads the scope limit policy from the probe manifest and checks that eleven authority flags are all false, that body_in_receipt is false, and that body_import_authority is copied_source_body_only. This is the gate that keeps a copied body a validator input and nothing more.
\n
Aggregation and the compact result card
\n
_evaluate blocks the whole result if any engine or structural check fails, and result_card projects the status plus four counts into a card._evaluate then confirms all four engine ids are present and that the source manifest declares exactly four modules. If any engine is blocked or any structural check fails, the whole result is blocked. result_card projects the pass or blocked status plus engine_count, copied_macro_source_module_count, stale_digest_count, and compiled_module_count into a compact card.
\nDiagram of the mechanism (8 steps).Diagram source & refs
flowchart TD A["Probe manifest current + stale digests, authority policy"] --> C B["Copied bodies and source manifest"] --> C C["_digest_refresh_matrix copied = target = current, stale differs from current"] --> F B --> D["_compile_gate py_compile, no import"] B --> E["_control_surface_anchor_matrix named surfaces present"] A --> G["_claim_ceiling_gate authority flags false"] D --> F E --> F G --> F F["_evaluate all engines pass, four modules present"] --> H["metadata-only result and card"] C -. mismatch .-> I["Blocked: refresh copies via exact-copy source lane"]
\n
Negative cases
\n
Perturbing a copy, never the fixtures
\n
The bundle is copied into a temporary workspace and perturbed there, so the on-disk fixtures stay untouched while four cases must be rejected.evaluate_negative_case copies the bundle into a temporary workspace and perturbs it, so the fixtures on disk are never mutated. Four cases must all be rejected.
\n
stale_digest_replay collapses the two digests
\n
Setting a row's stale digest equal to its current one must make all_stale_digests_differ report false.stale_digest_replay sets a row's stale digest equal to its current digest. The digest engine must then report all_stale_digests_differ as false. Expected code BATCH10_LIVE_SOURCE_DRIFT_STALE_DIGEST_REPLAY_REFUSED.
A body with the right anchors but invalid Python, digest refreshed, must be caught only by the compile engine counting fewer modules.compile_bypass overwrites one copied body with text that carries the right anchor comments but is invalid Python, and refreshes its digest so only the compile check can catch it. The compile engine must then compile fewer modules than it counted. Expected code BATCH10_LIVE_SOURCE_DRIFT_COMPILE_BYPASS_REFUSED.
\n
private_runtime_state_export flips an authority flag
\n
Turning on the runtime-state-export authorization flag in the policy must be rejected by the scope limit gate.private_runtime_state_export sets private_runtime_state_export_authorized to true in the policy. The scope limit gate must reject it. Expected code BATCH10_LIVE_SOURCE_DRIFT_PRIVATE_RUNTIME_EXPORT_REFUSED.
\n
live_mutation_authority_claim raises three mutation flags
\n
Setting the work log mutation, source-file changes, and mission execution flags true must all be rejected by the scope limit gate.live_mutation_authority_claim sets the work log mutation, source-file changes, and mission execution flags to true. The scope limit gate must reject all three. Expected code BATCH10_LIVE_SOURCE_DRIFT_MUTATION_AUTHORITY_REFUSED.
\n
Prior Art Grounding
\n
Reproducible-build and supply-chain provenance practice
\n
The discipline comes from reproducible builds and supply-chain provenance: declare inputs, fingerprint them, check artifacts against the fingerprints, and keep the stale digest visible as a fixture. The pattern comes from reproducible-build and supply-chain provenance practice: declare the source inputs, fingerprint them, check copied or generated artifacts against those fingerprints, and keep private state out of the result record. Two useful anchors are Bazel hermeticity, for its emphasis on declared inputs and repeatable outputs, and SLSA provenance, for recording how an artifact relates to its build inputs. This component applies that discipline to live source-copy drift, and keeps the stale digest visible as a regression fixture rather than overwriting it.
\n
Validation Result record Path
\n
Run from the microcosm-substrate/ public root.
\n
What a pass and a blocked result mean
\n
A pass writes four record types and rejects the negatives, while a blocked result is the valid drift signal to refresh the copies and rerun. A pass writes digest-refresh, compile-gate, anchor-matrix, and scope limit records and confirms the four negative cases are rejected. A blocked result is a valid outcome when the copies have fallen behind the current source: it is the signal to refresh the copied bodies and manifest digests through the exact-copy source lane and rerun, not a defect in this page. When the sentinel is blocked, the focused test is expected to fail on pass-status or exact-body equality and should be reported with the same refresh residual.
\n
Scope boundary
\n
Scope limit
\n
The narrow proof boundary
\n
The strongest claim is narrow: four copies can be compared, compiled, surface-checked, and policy-guarded, with four negatives rejected when the copy lane is current. The strongest claim this evidence supports is narrow. Selected copies of four internal control source files can be compared against the current source digest, compiled without import execution, checked for their required command and contract surfaces, and guarded by a policy whose authority flags are all false, with four named negative cases rejected when the exact-copy lane is current. That is the proof boundary.
\n
What a green result does not grant
\n
A green result grants no route, mutation, execution, commit, launch, or export authority, and the copied body is never an authority to run anything. The scope limit is lower than the claim. A green result does not grant route authority, work log or work log mutation authority, mission execution, git staging or commit approval, source-file changes, launch, public sharing, external model access, or private runtime state export. It does not establish whole-system correctness. The copied body is a validator input, never an authority to run anything.
\n",
"components": [
{
"anchor": "component-batch10_live_source_drift_capsule",
"component_id": "batch10_live_source_drift_capsule",
"public_label": "Live Source Drift Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 14,
"links": [
{
"id": "batch10_live_source_drift_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Live Source Drift Bundle",
"relation": "explains",
"url": "component-batch10-live-source-drift-capsule.html"
},
{
"id": "mechanism.batch10_live_source_drift_capsule.validates_public_live_source_drift_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public live source drift bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch10-live-source-drift-capsule-validates-public-live-source-drift-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.concurrency_mission_control",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Concurrency Mission Control",
"relation": "depends on",
"url": "paper-module-concurrency-mission-control.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch10_governance_compilers_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 10 Governance And Compilers Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[72:paper_module.batch10_governance_compilers_capsule].depends_on",
"url": "paper-module-batch10-governance-compilers-capsule.html"
}
],
"paper_module_id": "paper_module.batch10_live_source_drift_capsule",
"reader_packet": {
"abstract": "Set 10 Live Source Drift Bundle validates copied current internal control source bodies without route, ledger-mutation, or launch-scope decision.",
"authority_ceiling": "Fixture-bound source-digest, anchor, compile, and scope limit evidence only; no route authority, Work or work log mutation, mission execution, git approval, source-file changes, launch, public sharing, external model access, or non-public runtime export.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch10_live_source_drift_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch10-live-source-drift-capsule-diagram-1",
"html_url": "docs/paper-module-batch10-live-source-drift-capsule.html#paper-module-batch10-live-source-drift-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Probe manifest current + stale digests, authority policy\"] --> C B[\"Copied bodies and source manifest\"] --> C C[\"_digest_refresh_matrix copied = target = current, stale differs from current\"] --> F B --> D[\"_compile_gate py_compile, no import\"] B --> E[\"_control_surface_anchor_matrix named surfaces present\"] A --> G[\"_claim_ceiling_gate authority flags false\"] D --> F E --> F G --> F F[\"_evaluate all engines pass, four modules present\"] --> H[\"metadata-only result and card\"] C -. mismatch .-> I[\"Blocked: refresh copies via exact-copy source lane\"]",
"source_format": "mermaid",
"source_hash": "sha256:36857a9081bdbf752ce09ad0c2dc44c526748c2cb1de09749a08000be08f5a5a",
"source_line_end": 132,
"source_line_start": 119
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.concurrency_mission_control",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Concurrency Mission Control",
"relation": "depends on",
"url": "paper-module-concurrency-mission-control.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch10_live_source_drift_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Live Source Drift Bundle",
"relation": "explains",
"url": "component-batch10-live-source-drift-capsule.html"
},
{
"id": "mechanism.batch10_live_source_drift_capsule.validates_public_live_source_drift_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public live source drift bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch10-live-source-drift-capsule-validates-public-live-source-drift-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch10_live_source_drift_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 248,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch10_live_source_drift_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Bazel hermeticity",
"section_id": "reproducible-build-and-supply-chain-provenance-practice",
"url": "https://bazel.build/concepts/hermeticity"
},
{
"kind": "external_reference",
"label": "SLSA provenance",
"section_id": "reproducible-build-and-supply-chain-provenance-practice",
"url": "https://slsa.dev/spec/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**copies of four internal control source files are checked against the source they were copied from.** This component checks whether a small set of copies of internal control source files still match the current source they were copied from. The copies are exact Python bodies for four files: the option-surface router, the mission-transaction landing preflight, the work landing controller, and the work log controller.",
"level": 4,
"order": 1,
"section_id": "live-source-drift-is-recomputed-across-copied-inputs",
"source_line_end": 11,
"source_line_start": 3,
"title": "Live-source drift is recomputed across copied inputs"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Three questions at once: do the bytes still match the current digest, do the copies still compile, and do they still carry their named surfaces.** It answers three questions about those copies at once. Do their bytes still match the current source digest? Do they still compile as Python without being imported or run? Do they still contain the named command and contract surfaces they are supposed to carry? It emits a result record with the digest rows, compile status, anchor checks, and the scope limit, and no copied source body.",
"level": 4,
"order": 2,
"section_id": "three-engines-separate-source-projection-and-ceiling",
"source_line_end": 21,
"source_line_start": 12,
"title": "Three engines separate source, projection, and ceiling"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is built to go red: when the source moves ahead of the copies, the failed digest check is the signal to refresh them.** The component is built to go red. When the current source moves ahead of the copies, the digest check fails on purpose, and that red result is the signal to refresh the copies through the exact-copy source lane. It is a drift sentinel, not a one-off snapshot.",
"level": 4,
"order": 3,
"section_id": "a-green-result-remains-fixture-bound",
"source_line_end": 29,
"source_line_start": 22,
"title": "A green result remains fixture-bound"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 31,
"source_line_start": 30,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A copy can compile and match one recorded digest while being weeks behind the file it mirrors, so weak checks pass it.** Copied source drifts silently. A public copy of a control file can look fine, compile fine, and still be weeks behind the file it claims to mirror. A loose check that only asks \"is this valid Python?\" passes a stale copy. A check that only compares one recorded digest can pass a copy whose recorded digest was itself never refreshed.",
"level": 4,
"order": 5,
"section_id": "why-a-stale-copy-looks-fine",
"source_line_end": 40,
"source_line_start": 32,
"title": "Why a stale copy looks fine"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Each row holds the current and the earlier stale digest side by side, so proof of freshness and evidence of past drift cannot be erased.** This component makes drift visible instead of hiding it. Each digest row keeps two fingerprints side by side: the current source digest, and the stale digest that was recorded earlier. The row passes only when the copied body matches the current digest, and it separately asserts that the stale digest differs from the current one. The proof of freshness and the evidence of the earlier drift sit in the same row, so a refresh cannot quietly erase the fact that the copy was once behind.",
"level": 4,
"order": 6,
"section_id": "current-and-stale-digests-in-one-row",
"source_line_end": 51,
"source_line_start": 41,
"title": "Current and stale digests in one row"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 53,
"source_line_start": 52,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Both entry points forward to runcrownjewelorgan, and _evaluate loads the manifest and copied texts, then runs four engines in order.** run and runbatch10livesourcedrift_bundle are the two entry points. Both forward to runcrownjewelorgan with the shared SPEC, an evaluate function, and evaluatenegativecase. _evaluate loads the probe manifest and the copied source texts, then runs four engines in order and collects their findings.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-into-evaluate",
"source_line_end": 62,
"source_line_start": 54,
"title": "Two entry points into `_evaluate`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A row passes only when copied, target, and current digests all agree and the stale digest differs from the current one. digestrefreshmatrix builds one row per copied file. For each row it reads the current source digest and the stale recorded digest from the probe manifest, and the copied and target digests from the source manifest. A row's currentmatchescopy is true only when the current digest is non-empty and the copied digest, the target digest, and the current digest all agree and the manifest marks the target as matching. Its stalediffersfrom_current is true when a stale digest exists and differs from the current one. The engine rejects a copy whose body no longer matches the curren\u2026",
"level": 4,
"order": 9,
"section_id": "per-row-digest-agreement-in-digest-refresh-matrix",
"source_line_end": 77,
"source_line_start": 63,
"title": "Per-row digest agreement in `_digest_refresh_matrix`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Each copied file is compiled with pycompile into a throwaway directory but never imported, so malformed bodies are caught without running any code.** compilegate compiles each copied file with py_compile.compile into a throwaway temporary directory with doraise=True. It never imports the module, so a malformed body is caught without executing any copied code. import_executed stays false in the result. It rejects a missing target, a non-.py target, and any body that fails to compile.",
"level": 4,
"order": 10,
"section_id": "compile-without-import-in-compile-gate",
"source_line_end": 87,
"source_line_start": 78,
"title": "Compile without import in `_compile_gate`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Substring checks confirm each copy still carries its named command and contract surfaces, so a compiling body that quietly dropped one still fails.** controlsurfaceanchormatrix runs substring checks over each copied body. It confirms that each file still exposes the command and contract surfaces it is supposed to carry, for example def buildoptionsurface and optionsurfacelenspacketv0 in the option-surface router, or def cmdsessionfinalize and readreceiptid in the work log controller. A copy that compiles but has quietly lost a command or contract surface still fails here.",
"level": 4,
"order": 11,
"section_id": "named-surfaces-checked-by-substring",
"source_line_end": 99,
"source_line_start": 88,
"title": "Named surfaces checked by substring"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Eleven authority flags must be false and bodyin_receipt false, keeping every copied body a validator input and nothing more. claimceilinggate reads the scope limit policy from the probe manifest and checks that eleven authority flags are all false, that bodyinreceipt is false, and that bodyimportauthority is copiedsourcebody_only. This is the gate that keeps a copied body a validator input and nothing more.",
"level": 4,
"order": 12,
"section_id": "authority-flags-held-false-by-claim-ceiling-gate",
"source_line_end": 108,
"source_line_start": 100,
"title": "Authority flags held false by `_claim_ceiling_gate`"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "**_evaluate blocks the whole result if any engine or structural check fails, and resultcard projects the status plus four counts into a card. _evaluate then confirms all four engine ids are present and that the source manifest declares exactly four modules. If any engine is blocked or any structural check fails, the whole result is blocked. result_card projects the pass or blocked status plus enginecount, copiedmacrosourcemodule_count, staledigestcount, and compiledmodulecount into a compact card.",
"level": 4,
"order": 13,
"section_id": "aggregation-and-the-compact-result-card",
"source_line_end": 133,
"source_line_start": 109,
"title": "Aggregation and the compact result card"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 135,
"source_line_start": 134,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The bundle is copied into a temporary workspace and perturbed there, so the on-disk fixtures stay untouched while four cases must be rejected.** evaluatenegativecase copies the bundle into a temporary workspace and perturbs it, so the fixtures on disk are never mutated. Four cases must all be rejected.",
"level": 4,
"order": 15,
"section_id": "perturbing-a-copy-never-the-fixtures",
"source_line_end": 143,
"source_line_start": 136,
"title": "Perturbing a copy, never the fixtures"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Setting a row's stale digest equal to its current one must make allstaledigestsdiffer report false. staledigestreplay sets a row's stale digest equal to its current digest. The digest engine must then report allstaledigests_differ as false. Expected code BATCH10LIVESOURCEDRIFTSTALEDIGESTREPLAY_REFUSED.",
"level": 4,
"order": 16,
"section_id": "stale-digest-replay-collapses-the-two-digests",
"source_line_end": 151,
"source_line_start": 144,
"title": "`stale_digest_replay` collapses the two digests"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A body with the right anchors but invalid Python, digest refreshed, must be caught only by the compile engine counting fewer modules. compilebypass overwrites one copied body with text that carries the right anchor comments but is invalid Python, and refreshes its digest so only the compile check can catch it. The compile engine must then compile fewer modules than it counted. Expected code BATCH10LIVESOURCEDRIFTCOMPILEBYPASSREFUSED.",
"level": 4,
"order": 17,
"section_id": "compile-bypass-hides-invalid-python-behind-anchors",
"source_line_end": 160,
"source_line_start": 152,
"title": "`compile_bypass` hides invalid Python behind anchors"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Turning on the runtime-state-export authorization flag in the policy must be rejected by the scope limit gate. privateruntimestateexport sets privateruntimestateexportauthorized to true in the policy. The scope limit gate must reject it. Expected code BATCH10LIVESOURCEDRIFTPRIVATERUNTIMEEXPORT_REFUSED.",
"level": 4,
"order": 18,
"section_id": "private-runtime-state-export-flips-an-authority-flag",
"source_line_end": 168,
"source_line_start": 161,
"title": "`private_runtime_state_export` flips an authority flag"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Setting the work log mutation, source-file changes, and mission execution flags true must all be rejected by the scope limit gate.** livemutationauthority_claim sets the work log mutation, source-file changes, and mission execution flags to true. The scope limit gate must reject all three. Expected code BATCH10LIVESOURCEDRIFTMUTATIONAUTHORITYREFUSED.",
"level": 4,
"order": 19,
"section_id": "live-mutation-authority-claim-raises-three-mutation-flags",
"source_line_end": 176,
"source_line_start": 169,
"title": "`live_mutation_authority_claim` raises three mutation flags"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "prior-art-grounding",
"source_line_end": 178,
"source_line_start": 177,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The discipline comes from reproducible builds and supply-chain provenance: declare inputs, fingerprint them, check artifacts against the fingerprints, and keep the stale digest visible as a fixture.** The pattern comes from reproducible-build and supply-chain provenance practice: declare the source inputs, fingerprint them, check copied or generated artifacts against those fingerprints, and keep private state out of the result record. Two useful anchors are Bazel hermeticity, for its emphasis on declared inputs and repeatable outputs, and SLSA provenance, for recording how an artifact relates to its build inputs. This component applies that discipline to live source-copy drift, and keeps\u2026",
"level": 4,
"order": 21,
"section_id": "reproducible-build-and-supply-chain-provenance-practice",
"source_line_end": 193,
"source_line_start": 179,
"title": "Reproducible-build and supply-chain provenance practice"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from the microcosm-substrate/ public root.",
"level": 2,
"order": 22,
"section_id": "validation-receipt-path",
"source_line_end": 211,
"source_line_start": 194,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass writes four record types and rejects the negatives, while a blocked result is the valid drift signal to refresh the copies and rerun.** A pass writes digest-refresh, compile-gate, anchor-matrix, and scope limit records and confirms the four negative cases are rejected. A blocked result is a valid outcome when the copies have fallen behind the current source: it is the signal to refresh the copied bodies and manifest digests through the exact-copy source lane and rerun, not a defect in this page. When the sentinel is blocked, the focused test is expected to fail on pass-status or exact-body equality and should be reported with the same refresh residual.",
"level": 4,
"order": 23,
"section_id": "what-a-pass-and-a-blocked-result-mean",
"source_line_end": 223,
"source_line_start": 212,
"title": "What a pass and a blocked result mean"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 24,
"section_id": "claim-ceiling",
"source_line_end": 225,
"source_line_start": 224,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is narrow: four copies can be compared, compiled, surface-checked, and policy-guarded, with four negatives rejected when the copy lane is current.** The strongest claim this evidence supports is narrow. Selected copies of four internal control source files can be compared against the current source digest, compiled without import execution, checked for their required command and contract surfaces, and guarded by a policy whose authority flags are all false, with four named negative cases rejected when the exact-copy lane is current. That is the proof boundary.",
"level": 4,
"order": 25,
"section_id": "the-narrow-proof-boundary",
"source_line_end": 236,
"source_line_start": 226,
"title": "The narrow proof boundary"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A green result grants no route, mutation, execution, commit, launch, or export authority, and the copied body is never an authority to run anything.** The scope limit is lower than the claim. A green result does not grant route authority, work log or work log mutation authority, mission execution, git staging or commit approval, source-file changes, launch, public sharing, external model access, or private runtime state export. It does not establish whole-system correctness. The copied body is a validator input, never an authority to run anything.",
"level": 4,
"order": 26,
"section_id": "what-a-green-result-does-not-grant",
"source_line_end": 247,
"source_line_start": 237,
"title": "What a green result does not grant"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch10_live_source_drift_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 10 Live Source Drift Bundle validates copied current internal control source bodies without route, ledger-mutation, or launch-scope decision."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch10_live_source_drift_capsule.py",
"path": "src/microcosm_core/organs/batch10_live_source_drift_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for live source-drift digest checks, copied-source compile gates, control-surface anchor checks, negative cases, and scope limits.",
"symbols": [
"_digest_refresh_matrix",
"_compile_gate",
"_control_surface_anchor_matrix",
"_claim_ceiling_gate",
"_evaluate",
"run",
"run_batch10_live_source_drift_bundle",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch10_live_source_drift_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch10_live_source_drift_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch10_live_source_drift_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 10 Live Source Drift Bundle imports exact current internal control Python bodies after source drift. It validates stale-versus-current digest rows, copied-source compileability without import execution, required behavioral anchors, source-manifest boundaries, negative cases, and scope limits while excluding route authority, work log or work log mutation authority, mission execution, git approval, source-file changes, launch, public sharing, external model access, and non-public runtime export.",
"summary_line": "Set 10 Live Source Drift Bundle validates copied current internal control source bodies without route, ledger-mutation, or launch-scope decision.",
"title": "Set 10 Live Source Drift Bundle"
},
{
"anchor": "paper-module-batch7-macro-engines-capsule",
"authority_ceiling": "Fixture-bound public source-body import and deterministic exercise evidence only; no launch-scope decision, whole-system equivalence, semantic truth, investment-related actions, complete sandbox proof, selected-test sufficiency proof, external model access, or source-file changes.",
"body_html": "
Eight copied engines under one public fixture
\n
batch7_macro_engines_capsule takes eight unrelated engines that were copied out of the wider system as source and checks that each one still behaves the way its own tests and invariants say it should.batch7_macro_engines_capsule takes eight unrelated engines that were copied out of the wider system as source and checks that each one still behaves the way its own tests and invariants say it should. The engines have nothing in common at runtime: a trace parser written in JavaScript, a dependency-graph scheduler, an AST source index, a patch applier, a network-isolation guard, a robust numeric scorer, a personalized PageRank ranker, and a regression-test selector.
\n
The component runs each engine through one small deterministic exercise with a known answer, then records a metadata-only result row for it. The component runs each engine through one small deterministic exercise with a known answer, then records a metadata-only result row for it. A pass means every engine produced its expected output and every paired negative case kept failing.
\n
It writes batch7_macro_engines_capsule_result.json, a board, a validation result record, and a bundle result under the output directory the caller names. It writes batch7_macro_engines_capsule_result.json, a board, a validation result record, and a bundle result under the output directory the caller names. Those records carry status, counts, digests, and engine ids. They never carry the copied source or captured command output: every row sets body_in_receipt to false.
\n
Purpose
\n
Copied code is easy to trust and easy to be wrong about. A source body can be copied faithfully, pass a digest check, and still be broken or subtly different from the engine it was copied from. Copied code is easy to trust and easy to be wrong about. A source body can be copied faithfully, pass a digest check, and still be broken or subtly different from the engine it was copied from. A digest match proves the bytes arrived. It says nothing about whether the code runs.
\n
This component refuses to treat a digest match as behavior. Each engine has to execute and return the right value before its row is allowed to pass, and each engine's failure mode has to stay a failure. This component refuses to treat a digest match as behavior. Each engine has to execute and return the right value before its row is allowed to pass, and each engine's failure mode has to stay a failure. The point is not that any single engine is interesting on its own. It is that eight engines with quite different runtimes, an external Node process, in-process Python calls, and static AST reads, can be exercised together under one fixture and one scope limit.
\n
How it works
\n
_evaluate selects one of two paths from the input directory name. When the directory is exported_batch7_macro_engines_capsule_bundle, it calls _source_open_bundle_exercises, which gates all eight engine rows on two manifest fields, all_expected_digests_matched and all_required_anchors_present._evaluate selects one of two paths from the input directory name. When the directory is exported_batch7_macro_engines_capsule_bundle, it calls _source_open_bundle_exercises, which gates all eight engine rows on two manifest fields, all_expected_digests_matched and all_required_anchors_present. If either is false, every row is blocked. Otherwise _evaluate runs the eight live exercises in order. It records a finding for any row that did not pass and for any engine in EXPECTED_ENGINES that is missing from the result.
\n
Each live exercise is a single function. Each returns a status plus the concrete facts it checked. Each live exercise is a single function. Each returns a status plus the concrete facts it checked.
\n
Function
Engine
What it checks
_agent_trace_exercise
agent_trace_ir_compiler
Runs node --test parser.test.mjs against the copied parser.mjs. The suite includes a commit claim with no diff evidence, which the parser rejects.
_dag_exercise
constitutional_dag_kernel
Calls compute_waves on a six-node graph and checks the schedule is [["a","f"], ["b","c"], ["d"], ["e"]]. A two-node cycle must raise ValueError; audit_config_purity must flag an impure config.
_release_root_exercise
release_root_compiler
Parses the copied release_root_compiler.py with ast.parse and checks build_std_python_report and build_release_root_compiler are defined and the source reports missing_ref_count.
_source_surgeon_exercise
source_surgeon_patch
Uses SourceSurgeon._apply_unified_diff_hunks. A one-line diff must yield a = 'B'; a mismatched-context diff must raise ApplyError; malformed Python must raise SyntaxError.
_clean_clone_exercise
hermetic_clean_clone
Swaps socket.create_connection and socket.socket for a raiser, confirms an outbound connection raises NetworkDisabled, then restores both.
_calculator_exercise
calculator_standard_actor
Feeds [1, 2, 3, 4, 5, 100] to StandardActor._compute_center_scale. The robust center stays at 3.5 while the naive mean rises above 19.
_pagerank_exercise
personalized_pagerank_ranker
Calls personalized_pagerank on a four-node graph. Scores must sum to 1.0; an unknown source node must return {}.
_regression_selection_exercise
regression_test_selection
Runs select_impacted_tests.py --json, or reads the copied selector statically in the exported-bundle path, and checks the fallback test set is never empty.
\n
Three helpers hold the orchestration together. _safe_engine wraps each runner and converts any exception into a blocked row rather than crashing the run. Three helpers hold the orchestration together. _safe_engine wraps each runner and converts any exception into a blocked row rather than crashing the run. _semantic_runtime_exercises assembles the eight rows keyed by engine id and caches the result for a given input. evaluate_negative_case and _observed_negative_case confirm a declared negative case was actually seen before it counts, so a passing engine plus a witnessed rejection is what marks a case as covered.
\n
Shared crown-jewel entrypoint wiring
\n
run and run_batch7_bundle forward SPEC to run_crown_jewel_organ with _evaluate and evaluate_negative_case. result_card projects the result into a compact card that carries engine_count and copied_macro_source_module_count and nothing else from the run.run and run_batch7_bundle forward SPEC to run_crown_jewel_organ with _evaluate and evaluate_negative_case. result_card projects the result into a compact card that carries engine_count and copied_macro_source_module_count and nothing else from the run.
\nDiagram of the mechanism (5 steps).Diagram source & refs
flowchart TD input["Input directory"] bundle["Exported bundle: gate 8 rows on manifest digests and anchors"] live["Live fixture: run 8 engine exercises"] neg["Negative cases: must keep failing"] result["metadata-only result: status, counts, digests"] input --> bundle input --> live live --> neg bundle --> result neg --> result
\n
Negative cases
\n
Each engine has a paired case that must be rejected, listed in EXPECTED_NEGATIVE_CASES with its stable error code: Each engine has a paired case that must be rejected, listed in EXPECTED_NEGATIVE_CASES with its stable error code:
A case only counts once _observed_negative_case sees the matching engine pass and the specific rejection field set, so a case cannot be marked covered by an engine that did not run. A case only counts once _observed_negative_case sees the matching engine pass and the specific rejection field set, so a case cannot be marked covered by an engine that did not run.
\n
Prior Art Grounding
\n
The engines sit on standard practice in trace instrumentation and graph analysis: parse execution traces into structured units of work, preserve graph-rank invariants, schedule a dependency graph into waves, and choose focused tests without claiming they are sufficient. The engines sit on standard practice in trace instrumentation and graph analysis: parse execution traces into structured units of work, preserve graph-rank invariants, schedule a dependency graph into waves, and choose focused tests without claiming they are sufficient. Two anchors are directly relevant. OpenTelemetry documents traces and spans as a vendor-neutral model for units of work and their relationships, which is the shape the trace parser produces. NetworkX PageRank documents the PageRank family this component's ranker follows. The regression selector follows ordinary test-impact-analysis practice with a never-empty fallback.
\n
Validation Result record Path
\n
Run from the microcosm-substrate/ public root:
\n
The fixture command runs the eight live exercises and writes the result and sign-off JSON. The bundle command validates the copied source witnesses against the manifest without emitting any source body. The fixture command runs the eight live exercises and writes the result and sign-off JSON. The bundle command validates the copied source witnesses against the manifest without emitting any source body. The focused test covers the runtime path, the exported-bundle path, the negative cases, and the card's body omission. A pass across all of these means the fixture-bound run held; it does not extend past that.
\n
Scope boundary
\n
Scope limit
\n
The strongest claim this component supports is bounded and fixture-scoped: eight copied engines were exercised through public fixtures, each produced its expected output, and each paired negative case kept failing. The strongest claim this component supports is bounded and fixture-scoped: eight copied engines were exercised through public fixtures, each produced its expected output, and each paired negative case kept failing. That is validator evidence under the AUTHORITY_CEILING constant, which sets release_authorized, provider_dispatch, source_mutation_authorized, investment_advice, semantic_truth_authority, and test_completeness_proof all to false.
\n
The proof boundary stops at the copied source and the deterministic exercises. A pass is not launch-scope decision, hosted-public authority, whole-system equivalence, a complete sandbox, or proof that the selected tests are sufficient. The proof boundary stops at the copied source and the deterministic exercises. A pass is not launch-scope decision, hosted-public authority, whole-system equivalence, a complete sandbox, or proof that the selected tests are sufficient. The calculator is a numeric primitive, not market data or investment-related actions. The ranker checks a rank invariant, not meaning. The clean-clone check proves one hermetic baseline, not full isolation. The scope limit excludes raw operator transcripts, model-output data, account secrets, and live fetches, and none of those appear in any result record.
\n",
"components": [
{
"anchor": "component-batch7_macro_engines_capsule",
"component_id": "batch7_macro_engines_capsule",
"public_label": "Trace, Code-Map & Scheduling Engines Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 14,
"links": [
{
"id": "batch7_macro_engines_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Trace, Code-Map & Scheduling Engines Bundle",
"relation": "explains",
"url": "component-batch7-macro-engines-capsule.html"
},
{
"id": "mechanism.batch7_macro_engines_capsule.validates_public_macro_engines_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public source engines bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch7-macro-engines-capsule-validates-public-macro-engines-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.computer_use_action_trace_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Computer-Use Action Trace Replay",
"relation": "depends on",
"url": "paper-module-computer-use-action-trace-replay.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch8_policy_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Policy Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[60:paper_module.batch8_policy_engines_capsule].depends_on",
"url": "paper-module-batch8-policy-engines-capsule.html"
},
{
"id": "paper_module.batch8_tools_tail_primitives_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Tools-Tail Primitives Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[62:paper_module.batch8_tools_tail_primitives_capsule].depends_on",
"url": "paper-module-batch8-tools-tail-primitives-capsule.html"
},
{
"id": "paper_module.batch9_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 9 Source Engines Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[70:paper_module.batch9_macro_engines_capsule].depends_on",
"url": "paper-module-batch9-macro-engines-capsule.html"
}
],
"paper_module_id": "paper_module.batch7_macro_engines_capsule",
"reader_packet": {
"abstract": "Set 7 Source Engines Bundle imports source engine bodies and exercises trace, graph, scheduling, source-index, patch, numeric, rank, and regression-selection invariants.",
"authority_ceiling": "Fixture-bound public source-body import and deterministic exercise evidence only; no launch-scope decision, whole-system equivalence, semantic truth, investment-related actions, complete sandbox proof, selected-test sufficiency proof, external model access, or source-file changes.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch7_macro_engines_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch7-macro-engines-capsule-diagram-1",
"html_url": "docs/paper-module-batch7-macro-engines-capsule.html#paper-module-batch7-macro-engines-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD input[\"Input directory\"] bundle[\"Exported bundle: gate 8 rows on manifest digests and anchors\"] live[\"Live fixture: run 8 engine exercises\"] neg[\"Negative cases: must keep failing\"] result[\"metadata-only result: status, counts, digests\"] input --> bundle input --> live live --> neg bundle --> result neg --> result",
"source_format": "mermaid",
"source_hash": "sha256:656f1b5bc74d924b9d84aca0ccde80d7ebc6bbc1a4c4760bc57cb1364f8c72f9",
"source_line_end": 112,
"source_line_start": 100
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.computer_use_action_trace_replay",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Computer-Use Action Trace Replay",
"relation": "depends on",
"url": "paper-module-computer-use-action-trace-replay.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch7_macro_engines_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Trace, Code-Map & Scheduling Engines Bundle",
"relation": "explains",
"url": "component-batch7-macro-engines-capsule.html"
},
{
"id": "mechanism.batch7_macro_engines_capsule.validates_public_macro_engines_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public source engines bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch7-macro-engines-capsule-validates-public-macro-engines-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch7_macro_engines_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 209,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch7_macro_engines_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "OpenTelemetry",
"section_id": "the-engines-sit-on-standard-practice-in-trace-instrumentation-and-graph-analysis",
"url": "https://opentelemetry.io/docs/"
},
{
"kind": "external_reference",
"label": "NetworkX PageRank",
"section_id": "the-engines-sit-on-standard-practice-in-trace-instrumentation-and-graph-analysis",
"url": "https://networkx.org/documentation/stable/reference/algorithms/generated/networkx.algorithms.link_analysis.pagerank_alg.pagerank.html"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "batch7macroenginescapsule takes eight unrelated engines that were copied out of the wider system as source and checks that each one still behaves the way its own tests and invariants say it should.** batch7macroengines_capsule takes eight unrelated engines that were copied out of the wider system as source and checks that each one still behaves the way its own tests and invariants say it should. The engines have nothing in common at runtime: a trace parser written in JavaScript, a dependency-graph scheduler, an AST source index, a patch applier, a network-isolation guard, a robust numeric scorer, a personalized PageRank ranker, and a regression-test selector.",
"level": 4,
"order": 1,
"section_id": "eight-copied-engines-under-one-public-fixture",
"source_line_end": 13,
"source_line_start": 3,
"title": "Eight copied engines under one public fixture"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component runs each engine through one small deterministic exercise with a known answer, then records a metadata-only result row for it. The component runs each engine through one small deterministic exercise with a known answer, then records a metadata-only result row for it. A pass means every engine produced its expected output and every paired negative case kept failing.",
"level": 4,
"order": 2,
"section_id": "the-component-runs-each-engine-through-one-small-deterministic-exercise",
"source_line_end": 20,
"source_line_start": 14,
"title": "The component runs each engine through one small deterministic exercise"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It writes batch7macroenginescapsuleresult.json, a board, a validation result record, and a bundle result under the output directory the caller names. It writes batch7macroenginescapsuleresult.json, a board, a validation result record, and a bundle result under the output directory the caller names. Those records carry status, counts, digests, and engine ids. They never carry the copied source or captured command output: every row sets bodyinreceipt to false.",
"level": 4,
"order": 3,
"section_id": "it-writes-batch7-macro-engines-capsule-result-json",
"source_line_end": 29,
"source_line_start": 21,
"title": "It writes `batch7_macro_engines_capsule_result.json`"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 31,
"source_line_start": 30,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Copied code is easy to trust and easy to be wrong about. A source body can be copied faithfully, pass a digest check, and still be broken or subtly different from the engine it was copied from. Copied code is easy to trust and easy to be wrong about. A source body can be copied faithfully, pass a digest check, and still be broken or subtly different from the engine it was copied from. A digest match proves the bytes arrived. It says nothing about whether the code runs.",
"level": 4,
"order": 5,
"section_id": "copied-code-is-easy-to-trust-and-easy-to-be-wrong-about",
"source_line_end": 39,
"source_line_start": 32,
"title": "Copied code is easy to trust and easy to be wrong about"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "This component refuses to treat a digest match as behavior. Each engine has to execute and return the right value before its row is allowed to pass, and each engine's failure mode has to stay a failure. This component refuses to treat a digest match as behavior. Each engine has to execute and return the right value before its row is allowed to pass, and each engine's failure mode has to stay a failure. The point is not that any single engine is interesting on its own. It is that eight engines with quite different runtimes, an external Node process, in-process Python calls, and static AST reads, can be exercised together under one fixture and one scope limit.",
"level": 4,
"order": 6,
"section_id": "this-component-refuses-to-treat-a-digest-match-as-behavior",
"source_line_end": 49,
"source_line_start": 40,
"title": "This component refuses to treat a digest match as behavior"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 51,
"source_line_start": 50,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluate selects one of two paths from the input directory name. When the directory is exportedbatch7macroenginescapsulebundle, it calls sourceopenbundleexercises, which gates all eight engine rows on two manifest fields, allexpecteddigestsmatched and allrequiredanchorspresent.** _evaluate selects one of two paths from the input directory name. When the directory is exportedbatch7macroenginescapsule_bundle, it calls sourceopenbundleexercises, which gates all eight engine rows on two manifest fields, allexpecteddigests_matched and allrequiredanchors_present. If either is false, every row is blocked. Otherwise _evaluate runs the eight live exercises in order. It records a finding for any row\u2026",
"level": 4,
"order": 8,
"section_id": "evaluate-selects-one-of-two-paths-from-the-input-directory-name",
"source_line_end": 63,
"source_line_start": 52,
"title": "`_evaluate` selects one of two paths from the input directory name"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "Each live exercise is a single function. Each returns a status plus the concrete facts it checked. Each live exercise is a single function. Each returns a status plus the concrete facts it checked.",
"level": 4,
"order": 9,
"section_id": "each-live-exercise-is-a-single-function",
"source_line_end": 80,
"source_line_start": 64,
"title": "Each live exercise is a single function"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Three helpers hold the orchestration together. safeengine wraps each runner and converts any exception into a blocked row rather than crashing the run. Three helpers hold the orchestration together. safeengine wraps each runner and converts any exception into a blocked row rather than crashing the run. semanticruntime_exercises assembles the eight rows keyed by engine id and caches the result for a given input. evaluatenegativecase and observednegative_case confirm a declared negative case was actually seen before it counts, so a passing engine plus a witnessed rejection is what marks a case as covered.",
"level": 4,
"order": 10,
"section_id": "three-helpers-hold-the-orchestration-together",
"source_line_end": 91,
"source_line_start": 81,
"title": "Three helpers hold the orchestration together"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "run and runbatch7bundle forward SPEC to runcrownjewelorgan with evaluate and evaluatenegativecase. resultcard projects the result into a compact card that carries enginecount and copiedmacrosourcemodulecount and nothing else from the run. run and runbatch7bundle forward SPEC to runcrownjewel_organ with evaluate and evaluatenegativecase. resultcard projects the result into a compact card that carries enginecount and copiedmacrosourcemodule_count and nothing else from the run.",
"level": 4,
"order": 11,
"section_id": "shared-crown-jewel-entrypoint-wiring",
"source_line_end": 113,
"source_line_start": 92,
"title": "Shared crown-jewel entrypoint wiring"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 115,
"source_line_start": 114,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "Each engine has a paired case that must be rejected, listed in EXPECTEDNEGATIVECASES with its stable error code: Each engine has a paired case that must be rejected, listed in EXPECTEDNEGATIVECASES with its stable error code: tracecommitwithoutdiff gives BATCH7TRACECOMMITWITHOUTDIFFREJECTED. dagcycle gives BATCH7DAGCYCLEREJECTED. releaserootbadref gives BATCH7RELEASEROOTBADREFREPORTED. sourcesurgeoncontextmismatch gives BATCH7SOURCESURGEONCONTEXT_MISMATCH. cleanclonenetworkcall gives BATCH7CLEANCLONENETWORK_DISABLED. calculatoroutlier gives BATCH7CALCULATOROUTLIERRESISTANT. pagerankmissingsource gives BATCH7PAGERANKMISSINGSOURCEREFUSED. regressionselectionempty gives BATCH7RTSNEVEREMPTYFALL\u2026",
"level": 4,
"order": 13,
"section_id": "each-engine-has-a-paired-case-that-must-be-rejected",
"source_line_end": 130,
"source_line_start": 116,
"title": "Each engine has a paired case that must be rejected"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A case only counts once observednegativecase sees the matching engine pass and the specific rejection field set, so a case cannot be marked covered by an engine that did not run.** A case only counts once observednegative_case sees the matching engine pass and the specific rejection field set, so a case cannot be marked covered by an engine that did not run.",
"level": 4,
"order": 14,
"section_id": "a-case-only-counts-once-observed-negative-case-sees-the-matching-engine-pass",
"source_line_end": 137,
"source_line_start": 131,
"title": "A case only counts once `_observed_negative_case` sees the matching engine pass"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 139,
"source_line_start": 138,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The engines sit on standard practice in trace instrumentation and graph analysis: parse execution traces into structured units of work, preserve graph-rank invariants, schedule a dependency graph into waves, and choose focused tests without claiming they are sufficient. The engines sit on standard practice in trace instrumentation and graph analysis: parse execution traces into structured units of work, preserve graph-rank invariants, schedule a dependency graph into waves, and choose focused tests without claiming they are sufficient. Two anchors are directly relevant. OpenTelemetry documents traces and spans as a vendor-neutral model for units of work and their relationships, which is the\u2026",
"level": 4,
"order": 16,
"section_id": "the-engines-sit-on-standard-practice-in-trace-instrumentation-and-graph-analysis",
"source_line_end": 154,
"source_line_start": 140,
"title": "The engines sit on standard practice in trace instrumentation and graph analysis"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from the microcosm-substrate/ public root:",
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 174,
"source_line_start": 155,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The fixture command runs the eight live exercises and writes the result and sign-off JSON. The bundle command validates the copied source witnesses against the manifest without emitting any source body. The fixture command runs the eight live exercises and writes the result and sign-off JSON. The bundle command validates the copied source witnesses against the manifest without emitting any source body. The focused test covers the runtime path, the exported-bundle path, the negative cases, and the card's body omission. A pass across all of these means the fixture-bound run held; it does not extend past that.",
"level": 4,
"order": 18,
"section_id": "the-fixture-command-runs-the-eight-live-exercises-and-writes-the-result-and-acceptance-json",
"source_line_end": 184,
"source_line_start": 175,
"title": "The fixture command runs the eight live exercises and writes the result and sign-off JSON"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "claim-ceiling",
"source_line_end": 186,
"source_line_start": 185,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim this component supports is bounded and fixture-scoped: eight copied engines were exercised through public fixtures, each produced its expected output, and each paired negative case kept failing. The strongest claim this component supports is bounded and fixture-scoped: eight copied engines were exercised through public fixtures, each produced its expected output, and each paired negative case kept failing. That is validator evidence under the AUTHORITYCEILING constant, which sets releaseauthorized, providerdispatch, sourcemutationauthorized, investmentadvice, semantictruthauthority, and testcompletenessproof all to false.",
"level": 4,
"order": 20,
"section_id": "the-strongest-claim-this-component-supports-is-bounded-and-fixture-scoped",
"source_line_end": 196,
"source_line_start": 187,
"title": "The strongest claim this component supports is bounded and fixture-scoped"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The proof boundary stops at the copied source and the deterministic exercises. A pass is not launch-scope decision, hosted-public authority, whole-system equivalence, a complete sandbox, or proof that the selected tests are sufficient. The proof boundary stops at the copied source and the deterministic exercises. A pass is not launch-scope decision, hosted-public authority, whole-system equivalence, a complete sandbox, or proof that the selected tests are sufficient. The calculator is a numeric primitive, not market data or investment-related actions. The ranker checks a rank invariant, not meaning. The clean-clone check proves one hermetic baseline, not full isolation. The scope limit excl\u2026",
"level": 4,
"order": 21,
"section_id": "the-proof-boundary-stops-at-the-copied-source-and-the-deterministic-exercises",
"source_line_end": 208,
"source_line_start": 197,
"title": "The proof boundary stops at the copied source and the deterministic exercises"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch7_macro_engines_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 7 Source Engines Bundle imports source engine bodies and exercises trace, graph, scheduling, source-index, patch, numeric, rank, and regression-selection invariants."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch7_macro_engines_capsule.py",
"path": "src/microcosm_core/organs/batch7_macro_engines_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for Set 7 public source-engine source imports, deterministic exercises, negative cases, and scope limits.",
"symbols": [
"_agent_trace_exercise",
"_dag_exercise",
"_release_root_exercise",
"_source_surgeon_exercise",
"_clean_clone_exercise",
"_calculator_exercise",
"_pagerank_exercise",
"_regression_selection_exercise"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch7_macro_engines_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch7_macro_engines_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch7_macro_engines_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 7 Source Engines Bundle binds the accepted batch7_macro_engines_capsule component to its public source-open bundle. It checks copied trace parser, codemap layout, DAG scheduling, launch-root, source-surgeon, clean-clone, calculator, PageRank, and regression-selection witnesses, negative cases, digest boundaries, and scope limits while excluding launch-scope decision, whole-system equivalence, semantic truth, investment-related actions, sandbox completeness, and selected-test sufficiency proof.",
"summary_line": "Set 7 Source Engines Bundle imports source engine bodies and exercises trace, graph, scheduling, source-index, patch, numeric, rank, and regression-selection invariants.",
"title": "Set 7 Source Engines Bundle"
},
{
"anchor": "paper-module-batch9-macro-engines-capsule",
"authority_ceiling": "Fixture-bound public source-body import and deterministic exercise evidence only; no live lineage truth, human approval authority, real market/news truth, host-state truth, work log truth, external model access, source-file changes, public sharing, launch-scope decision, or whole-system equivalence.",
"body_html": "
Ten copied engines run against public fixtures
\n
This component copies ten backend and governance engine bodies into a public bundle and runs each against fixtures, emitting metadata-only records of what ran and whether the paired rejection fired. This component copies ten backend and governance engine bodies out of the larger system into a public bundle, then runs each one against small fixtures to confirm the copied code still computes what it did before. It emits metadata-only records: which engine ran, what it produced, and whether the paired rejection fired.
\n
Execution tests behavior beyond matching digests
\n
A digest proves the bytes match but not that the logic still behaves, so the bundle loads each copied module and calls its real functions rather than trusting baked-in output. Copying a file into a public bundle proves that the bytes match. It does not establish that the logic still behaves. This component closes that gap by loading each copied module and calling its real functions, rather than comparing against output files baked once and trusted forever.
\n
One bundle spans ten bounded engine domains
\n
The result is one reproducible evidence bundle spanning ten domains, from provenance lineage and approval adjudication to host-pressure admission and milestone-relative quality accounting. The result is a reproducible evidence bundle for provenance lineage, approval adjudication, Python symbol indexing, mission-graph wave ordering, dependency pin drift, config-authority audit, host-pressure admission, doctrine enrichment, worker pre-dispatch gating, and milestone-relative quality accounting.
\n
Purpose
\n
A digest check cannot prove behavior
\n
An exported copy can pass a digest check and still be dead, because a digest confirms identical bytes but says nothing about whether the function still runs. An exported copy can pass a digest check and still be dead. A digest says the bytes are identical to what was copied. It says nothing about whether the function still runs, or whether a later edit quietly changed a branch. A result file that is written once and checked forever has the same weakness: it stops tracking the code it claims to describe.
\n
Live execution guards the loose-copy failure
\n
The component ties proof to live execution, loading each engine from its copied file so the check moves when the body moves and a digest-faithful but behavior-lost bundle fails. This component ties the proof to live execution. Each of the ten engines is loaded from its copied file and driven with a fixture, so the check moves when the copied body moves. That is the point of the "loose copy" failure it guards against: a bundle that looks faithful because the digest matches, while the behaviour it was supposed to preserve is gone.
\n
How it works
\n
The pipeline runs in four stages, in this order.
\n
Stage one resolves the source manifest
\n
First _source_rows reads source_module_manifest.json, one row per copied engine carrying a module id, target path, expected digest, and required anchors, while helpers locate each copied file. First it resolves the source manifest. _source_rows reads source_module_manifest.json, one row per copied engine, each with a module id, a copied target path, an expected digest, and required anchor strings. _source_module_path and _source_module_text locate the copied file for a given module id.
\n
Stage two checks digests and anchors
\n
Second, a copied body failing its manifest digest raises CROWN_JEWEL_SOURCE_DIGEST_MISMATCH, and one missing a required anchor raises CROWN_JEWEL_SOURCE_ANCHOR_MISSING even after the digest is refreshed. Second it checks digests and anchors. A copied body whose bytes no longer match its manifest digest fails CROWN_JEWEL_SOURCE_DIGEST_MISMATCH. A body that has lost one of its required anchors fails CROWN_JEWEL_SOURCE_ANCHOR_MISSING, even when the digest was refreshed to match the edit.
\n
Stage three runs the ten exercises
\n
Third, _run_all_exercises drives all ten: nine import the copied module and call its functions, while the text-backed mission graph gates a local reimplementation on anchors in the copied text. Third it runs the exercises. _run_all_exercises drives all ten. Nine are path-backed: the copied module is imported with _load_python_source_module and its own functions are called against a fixture in a temporary directory. One, the mission graph, is text-backed: _source_backed_mission_graph_contract scans the copied text for branch anchors and gates a local reimplementation on whether they are present.
\n
Stage four evaluates and emits records
\n
Fourth, evaluate collects exercise statuses, counts mechanisms and modules, and returns a body_in_receipt-false record, with run and run_batch9_bundle wrapping the two paths and result_card projecting a compact card. Fourth it evaluates and emits. evaluate collects the exercise statuses, counts the declared mechanisms against EXPECTED_MECHANISMS, counts copied modules, and returns a record with body_in_receipt set to false. run wraps this for the fixture path and run_batch9_bundle for the exported bundle path. result_card projects a compact card carrying an authority floor and a body floor.
\n
Each engine is exercised through its real imported function:
\n
Mechanism id
Imported source
Function called
lineage_temporal_provenance_chain_resolver
system/server/lineage.py
build_temporal_lineage, _resolve_primary_lineage
approval_sign_off_claim_adjudicator
system/lib/approval_registry.py
list_approvals, decide_approval
python_ast_symbol_index_doc_tree
system/lib/python_documentation_tree.py
build_file_entry
mission_graph_topological_compiler
system/server/graph.py
text anchors gate compile_mission_graph
dependency_pin_drift_auditor
tools/dev/check_pin_drift.py
parse_requirements, evaluate
config_authority_drift_audit
system/lib/config_authority_registry.py
validate_config_authority_registry
host_pressure_admission_decision_gate
system/lib/admission_consumer.py
build_admission_consumer_decision
doctrine_file_enrichment_multihop_join
system/server/doctrine_enrichment.py
DoctrineEnrichmentService.get_file_doctrine
worker_job_budget_forbidden_surface_gate
system/lib/type_a_worker_harness.py
_enforce_budget, _contains_forbidden_surface
milestone_relative_promotion_quality_accounting
system/lib/population_lane_metrics.py
classify_blockers_and_next_action
\n
What each check confirms:
\n
lineage_temporal_provenance_chain_resolver: Parent-chain resolution; resolve_lineage sets self_loop_pruned when a node names itself as parent and cycle_detected when a run id repeats.
approval_sign_off_claim_adjudicator: A pre-acquired claim with a different nonce makes decide_approval return claim_conflict, recorded as preacquired_claim_refused.
python_ast_symbol_index_doc_tree: Symbols are read back from written-out Python; a file that will not parse sets syntax_error_gap.
mission_graph_topological_compiler: Group closure and upstream walk build waves by peeling indegree-zero nodes; a missing target blocks with missing_target_error.
dependency_pin_drift_auditor: Requirements are bucketed ok, drifted, missing, or unparseable.
config_authority_drift_audit: A registry that marks a generated projection mutable is rejected.
host_pressure_admission_decision_gate: Under the auto policy a summary-first recommendation blocks new work, setting auto_policy_blocked.
doctrine_file_enrichment_multihop_join: An indexed path returns a populated envelope; an unindexed path returns an empty one.
worker_job_budget_forbidden_surface_gate: A job over budget or naming a forbidden surface is blocked before dispatch.
milestone_relative_promotion_quality_accounting: Milestone-relative promotion metrics are computed; runs missing a committed-at timestamp are counted into a bucket.
\n
Local reference implementations define expected shape
\n
The component also carries local reference implementations that define the expected shape, but for nine engines the imported copy runs, with the mission graph the sole case where the local body runs. The component also carries local reference implementations, such as resolve_lineage, build_ast_doc_tree, and compile_mission_graph. These define the expected shape. For nine engines the imported copy is what actually runs; the mission graph is the one case where the local body runs, gated by anchors found in the copied text.
\nDiagram of the mechanism (6 steps).Diagram source & refs
Each engine ships one paired rejection, and evaluate_negative_case patches the positive fixture, reruns the exercises, and blocks with a stable error code only when the rejection actually fired. Each engine ships one paired rejection. evaluate_negative_case takes a case id, loads <case_id>.json, applies its fixture_patch onto the positive fixture with _deep_update, reruns the exercises, and calls _negative_case_observed to inspect one field. If the rejection fired the case returns status blocked with its stable error code; if it did not, the case fails. The ten cases and their codes:
Two tamper modes sit alongside these: an unrefreshed digest fails CROWN_JEWEL_SOURCE_DIGEST_MISMATCH, and a removed anchor still fails CROWN_JEWEL_SOURCE_ANCHOR_MISSING, keeping fixture-only replacements outside the proof shape. Two tamper modes sit alongside these. An edit to a copied body with no manifest digest refresh fails CROWN_JEWEL_SOURCE_DIGEST_MISMATCH. Removing a required anchor while refreshing the digest still fails CROWN_JEWEL_SOURCE_ANCHOR_MISSING. Together these keep a fixture-only or static-record replacement outside the accepted proof shape.
\n
Prior Art Grounding
\n
Golden-master testing that runs the copy
\n
The component follows the characterization or golden-master tradition, pinning behavior with deterministic fixtures, but its variation runs the copied code rather than comparing against a frozen output file. This component follows the characterization, or golden-master, testing tradition (Feathers, Working Effectively with Legacy Code), which pins existing behaviour with deterministic fixtures before trusting it. The variation here is that the pin runs the copied code rather than comparing against a frozen output file. The result is fixture-bound import evidence, not lineage truth, human approval authority, or market-level conclusions.
\n
Validation Result record Path
\n
Run these from the microcosm-substrate/ public root:
\n
What the four commands together prove
\n
The four commands prove the fixture and bundle paths, the focused test covers imports and rejections, and the corpus check confirms the projections stay consistent, nothing wider. The first command proves the public fixture path. The second proves the exported bundle path. The focused test covers exact-copy imports, source-sensitive behaviour shifts, digest-mismatch blocking, required-anchor removal, and the semantic rejections above. The corpus check confirms the generated instance, Mermaid, Atlas card, and this Markdown stay mutually consistent. A pass means the copied bodies and fixtures ran and matched here, nothing wider.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim is narrow: the ten copied bodies still behave like their imported source under a bounded positive case and one paired rejection, checked in the public clone. The strongest claim this component supports is narrow: the ten copied engine bodies and their public fixtures show that each engine still behaves like its imported source under a bounded positive case and one paired rejection, checked here in the public clone. Its proof boundary is the copied-body import, the digest and anchor checks, the ten exercises, the ten negative cases, and the metadata-only records they emit.
\n
What the scope limit excludes
\n
It refuses everything past that: no live lineage, approval, market, host-state, or work log truth, no dispatch, mutation, or launch, since generated views are reader surfaces, source-linked only. It refuses everything past that. It is not live lineage truth, human approval authority, real market or news truth, host-state truth, work log truth, external model access, source-file changes, public sharing, launch-scope decision, whole-system equivalence, or whole-system correctness. The scope limit excludes all of these. The generated Markdown, JSON, and site views are reader surfaces; source authority stays in the JSON contract, the source manifest, the tests, and the records.
\n",
"components": [
{
"anchor": "component-batch9_macro_engines_capsule",
"component_id": "batch9_macro_engines_capsule",
"public_label": "Backend & Governance Engines Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 14,
"links": [
{
"id": "batch9_macro_engines_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Backend & Governance Engines Bundle",
"relation": "explains",
"url": "component-batch9-macro-engines-capsule.html"
},
{
"id": "mechanism.batch9_macro_engines_capsule.validates_public_macro_engines_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public source engines bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch9-macro-engines-capsule-validates-public-macro-engines-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.batch7_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Source Engines Bundle",
"relation": "depends on",
"url": "paper-module-batch7-macro-engines-capsule.html"
},
{
"id": "paper_module.work_landing_control_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Work Landing Control Spine",
"relation": "depends on",
"url": "paper-module-work-landing-control-spine.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch10_governance_compilers_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 10 Governance And Compilers Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[72:paper_module.batch10_governance_compilers_capsule].depends_on",
"url": "paper-module-batch10-governance-compilers-capsule.html"
}
],
"paper_module_id": "paper_module.batch9_macro_engines_capsule",
"reader_packet": {
"abstract": "Set 9 Source Engines Bundle imports backend and governance source-engine bodies and exercises provenance, approval, AST, finance-news, mission graph, dependency, config, edge, WorkAtlas, host-pressure, doctrine-enrichment, worker-budget, and milestone-quality invariants.",
"authority_ceiling": "Fixture-bound public source-body import and deterministic exercise evidence only; no live lineage truth, human approval authority, real market/news truth, host-state truth, work log truth, external model access, source-file changes, public sharing, launch-scope decision, or whole-system equivalence.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch9_macro_engines_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch9-macro-engines-capsule-diagram-1",
"html_url": "docs/paper-module-batch9-macro-engines-capsule.html#paper-module-batch9-macro-engines-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD manifest[\"source_module_manifest.json 10 copied engine bodies\"] fixtures[\"fixture input positive fixture + 10 negative cases\"] digest[\"digest + anchor check bytes match, anchors present\"] exercises[\"_run_all_exercises load each copy, call its functions\"] neg[\"negative cases apply fixture_patch, rerun\"] emit[\"evaluate metadata-only record + card\"] manifest --> digest fixtures --> exercises digest --> exercises exercises --> neg exercises --> emit neg --> emit",
"source_format": "mermaid",
"source_hash": "sha256:f0b4a36abbb40712f9eef12aba737736a2e3a6cc2f707ac8bbc53982eebbb382",
"source_line_end": 166,
"source_line_start": 152
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.batch7_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Source Engines Bundle",
"relation": "depends on",
"url": "paper-module-batch7-macro-engines-capsule.html"
},
{
"id": "paper_module.work_landing_control_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Work Landing Control Spine",
"relation": "depends on",
"url": "paper-module-work-landing-control-spine.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch9_macro_engines_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Backend & Governance Engines Bundle",
"relation": "explains",
"url": "component-batch9-macro-engines-capsule.html"
},
{
"id": "mechanism.batch9_macro_engines_capsule.validates_public_macro_engines_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public source engines bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch9-macro-engines-capsule-validates-public-macro-engines-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch9_macro_engines_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 278,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch9_macro_engines_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component copies ten backend and governance engine bodies into a public bundle and runs each against fixtures, emitting metadata-only records of what ran and whether the paired rejection fired.** This component copies ten backend and governance engine bodies out of the larger system into a public bundle, then runs each one against small fixtures to confirm the copied code still computes what it did before. It emits metadata-only records: which engine ran, what it produced, and whether the paired rejection fired.",
"level": 4,
"order": 1,
"section_id": "ten-copied-engines-run-against-public-fixtures",
"source_line_end": 12,
"source_line_start": 3,
"title": "Ten copied engines run against public fixtures"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A digest proves the bytes match but not that the logic still behaves, so the bundle loads each copied module and calls its real functions rather than trusting baked-in output.** Copying a file into a public bundle proves that the bytes match. It does not establish that the logic still behaves. This component closes that gap by loading each copied module and calling its real functions, rather than comparing against output files baked once and trusted forever.",
"level": 4,
"order": 2,
"section_id": "execution-tests-behavior-beyond-matching-digests",
"source_line_end": 21,
"source_line_start": 13,
"title": "Execution tests behavior beyond matching digests"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The result is one reproducible evidence bundle spanning ten domains, from provenance lineage and approval adjudication to host-pressure admission and milestone-relative quality accounting.** The result is a reproducible evidence bundle for provenance lineage, approval adjudication, Python symbol indexing, mission-graph wave ordering, dependency pin drift, config-authority audit, host-pressure admission, doctrine enrichment, worker pre-dispatch gating, and milestone-relative quality accounting.",
"level": 4,
"order": 3,
"section_id": "one-bundle-spans-ten-bounded-engine-domains",
"source_line_end": 31,
"source_line_start": 22,
"title": "One bundle spans ten bounded engine domains"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 33,
"source_line_start": 32,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**An exported copy can pass a digest check and still be dead, because a digest confirms identical bytes but says nothing about whether the function still runs.** An exported copy can pass a digest check and still be dead. A digest says the bytes are identical to what was copied. It says nothing about whether the function still runs, or whether a later edit quietly changed a branch. A result file that is written once and checked forever has the same weakness: it stops tracking the code it claims to describe.",
"level": 4,
"order": 5,
"section_id": "a-digest-check-cannot-prove-behavior",
"source_line_end": 43,
"source_line_start": 34,
"title": "A digest check cannot prove behavior"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component ties proof to live execution, loading each engine from its copied file so the check moves when the body moves and a digest-faithful but behavior-lost bundle fails.** This component ties the proof to live execution. Each of the ten engines is loaded from its copied file and driven with a fixture, so the check moves when the copied body moves. That is the point of the \"loose copy\" failure it guards against: a bundle that looks faithful because the digest matches, while the behaviour it was supposed to preserve is gone.",
"level": 4,
"order": 6,
"section_id": "live-execution-guards-the-loose-copy-failure",
"source_line_end": 53,
"source_line_start": 44,
"title": "Live execution guards the loose-copy failure"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The pipeline runs in four stages, in this order.",
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 57,
"source_line_start": 54,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "First sourcerows reads sourcemodule_manifest.json, one row per copied engine carrying a module id, target path, expected digest, and required anchors, while helpers locate each copied file.** First it resolves the source manifest. sourcerows reads sourcemodulemanifest.json, one row per copied engine, each with a module id, a copied target path, an expected digest, and required anchor strings. sourcemodulepath and sourcemoduletext locate the copied file for a given module id.",
"level": 4,
"order": 8,
"section_id": "stage-one-resolves-the-source-manifest",
"source_line_end": 67,
"source_line_start": 58,
"title": "Stage one resolves the source manifest"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Second, a copied body failing its manifest digest raises CROWNJEWELSOURCEDIGESTMISMATCH, and one missing a required anchor raises CROWNJEWELSOURCEANCHORMISSING even after the digest is refreshed.** Second it checks digests and anchors. A copied body whose bytes no longer match its manifest digest fails CROWNJEWELSOURCEDIGESTMISMATCH. A body that has lost one of its required anchors fails CROWNJEWELSOURCEANCHORMISSING, even when the digest was refreshed to match the edit.",
"level": 4,
"order": 9,
"section_id": "stage-two-checks-digests-and-anchors",
"source_line_end": 77,
"source_line_start": 68,
"title": "Stage two checks digests and anchors"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Third, runallexercises drives all ten: nine import the copied module and call its functions, while the text-backed mission graph gates a local reimplementation on anchors in the copied text.** Third it runs the exercises. runall_exercises drives all ten. Nine are path-backed: the copied module is imported with loadpythonsourcemodule and its own functions are called against a fixture in a temporary directory. One, the mission graph, is text-backed: sourcebackedmissiongraph_contract scans the copied text for branch anchors and gates a local reimplementation on whether they are present.",
"level": 4,
"order": 10,
"section_id": "stage-three-runs-the-ten-exercises",
"source_line_end": 88,
"source_line_start": 78,
"title": "Stage three runs the ten exercises"
},
{
"block_kinds": [
"list",
"paragraph",
"table"
],
"excerpt": "**Fourth, evaluate collects exercise statuses, counts mechanisms and modules, and returns a bodyinreceipt-false record, with run and runbatch9bundle wrapping the two paths and resultcard projecting a compact card. Fourth it evaluates and emits. evaluate collects the exercise statuses, counts the declared mechanisms against EXPECTED_MECHANISMS, counts copied modules, and returns a record with bodyinreceipt set to false. run wraps this for the fixture path and runbatch9bundle for the exported bundle path. result_card projects a compact card carrying an authority floor and a body floor. Each engine is exercised through its real imported function: What each check confirms: lineagetemporalproven\u2026",
"level": 4,
"order": 11,
"section_id": "stage-four-evaluates-and-emits-records",
"source_line_end": 141,
"source_line_start": 89,
"title": "Stage four evaluates and emits records"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "**The component also carries local reference implementations that define the expected shape, but for nine engines the imported copy runs, with the mission graph the sole case where the local body runs.** The component also carries local reference implementations, such as resolvelineage, buildastdoctree, and compilemissiongraph. These define the expected shape. For nine engines the imported copy is what actually runs; the mission graph is the one case where the local body runs, gated by anchors found in the copied text.",
"level": 4,
"order": 12,
"section_id": "local-reference-implementations-define-expected-shape",
"source_line_end": 167,
"source_line_start": 142,
"title": "Local reference implementations define expected shape"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 169,
"source_line_start": 168,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "Each engine ships one paired rejection, and evaluatenegative_case patches the positive fixture, reruns the exercises, and blocks with a stable error code only when the rejection actually fired.** Each engine ships one paired rejection. evaluatenegativecase takes a case id, loads .json, applies its fixturepatch onto the positive fixture with deep_update, reruns the exercises, and calls negativecase_observed to inspect one field. If the rejection fired the case returns status blocked with its stable error code; if it did not, the case fails. The ten cases and their codes: lineageselflooppruned -> BATCH9LINEAGESELFLOOP_PRUNED approvalpreacquiredclaimrefused -> BATCH9APPROVALPREACQUIRE\u2026",
"level": 4,
"order": 14,
"section_id": "each-engine-ships-one-paired-rejection",
"source_line_end": 191,
"source_line_start": 170,
"title": "Each engine ships one paired rejection"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Two tamper modes sit alongside these: an unrefreshed digest fails CROWNJEWELSOURCEDIGESTMISMATCH, and a removed anchor still fails CROWNJEWELSOURCEANCHORMISSING, keeping fixture-only replacements outside the proof shape.** Two tamper modes sit alongside these. An edit to a copied body with no manifest digest refresh fails CROWNJEWELSOURCEDIGESTMISMATCH. Removing a required anchor while refreshing the digest still fails CROWNJEWELSOURCEANCHORMISSING. Together these keep a fixture-only or static-record replacement outside the accepted proof shape.",
"level": 4,
"order": 15,
"section_id": "two-tamper-modes-on-the-copied-bodies",
"source_line_end": 202,
"source_line_start": 192,
"title": "Two tamper modes on the copied bodies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 204,
"source_line_start": 203,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component follows the characterization or golden-master tradition, pinning behavior with deterministic fixtures, but its variation runs the copied code rather than comparing against a frozen output file.** This component follows the characterization, or golden-master, testing tradition ([Feathers, Working Effectively with Legacy Code](https://en.wikipedia.org/wiki/Characterization_test)), which pins existing behaviour with deterministic fixtures before trusting it. The variation here is that the pin runs the copied code rather than comparing against a frozen output file. The result is fixture-bound import evidence, not lineage truth, human approval authority, or market-level conclusio\u2026",
"level": 4,
"order": 17,
"section_id": "golden-master-testing-that-runs-the-copy",
"source_line_end": 217,
"source_line_start": 205,
"title": "Golden-master testing that runs the copy"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run these from the microcosm-substrate/ public root:",
"level": 2,
"order": 18,
"section_id": "validation-receipt-path",
"source_line_end": 240,
"source_line_start": 218,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The four commands prove the fixture and bundle paths, the focused test covers imports and rejections, and the corpus check confirms the projections stay consistent, nothing wider.** The first command proves the public fixture path. The second proves the exported bundle path. The focused test covers exact-copy imports, source-sensitive behaviour shifts, digest-mismatch blocking, required-anchor removal, and the semantic rejections above. The corpus check confirms the generated instance, Mermaid, Atlas card, and this Markdown stay mutually consistent. A pass means the copied bodies and fixtures ran and matched here, nothing wider.",
"level": 4,
"order": 19,
"section_id": "what-the-four-commands-together-prove",
"source_line_end": 252,
"source_line_start": 241,
"title": "What the four commands together prove"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "claim-ceiling",
"source_line_end": 254,
"source_line_start": 253,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest supported claim is narrow: the ten copied bodies still behave like their imported source under a bounded positive case and one paired rejection, checked in the public clone.** The strongest claim this component supports is narrow: the ten copied engine bodies and their public fixtures show that each engine still behaves like its imported source under a bounded positive case and one paired rejection, checked here in the public clone. Its proof boundary is the copied-body import, the digest and anchor checks, the ten exercises, the ten negative cases, and the metadata-only records they emit.",
"level": 4,
"order": 21,
"section_id": "the-strongest-supported-claim-is-narrow",
"source_line_end": 265,
"source_line_start": 255,
"title": "The strongest supported claim is narrow"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It refuses everything past that: no live lineage, approval, market, host-state, or work log truth, no dispatch, mutation, or launch, since generated views are reader surfaces, source-linked only.** It refuses everything past that. It is not live lineage truth, human approval authority, real market or news truth, host-state truth, work log truth, external model access, source-file changes, public sharing, launch-scope decision, whole-system equivalence, or whole-system correctness. The scope limit excludes all of these. The generated Markdown, JSON, and site views are reader surfaces; source authority stays in the JSON contract, the source manifest, the tests, and the records.",
"level": 4,
"order": 22,
"section_id": "what-the-authority-ceiling-excludes",
"source_line_end": 277,
"source_line_start": 266,
"title": "What the scope limit excludes"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch9_macro_engines_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 9 Source Engines Bundle imports backend and governance source-engine bodies and exercises provenance, approval, AST, finance-news, mission graph, dependency, config, edge, WorkAtlas, host-pressure, doctrine-enrichment, worker-budget, and milestone-quality invariants."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch9_macro_engines_capsule.py",
"path": "src/microcosm_core/organs/batch9_macro_engines_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for Set 9 public source-engine source imports, thirteen deterministic exercises, negative cases, exact-copy digest checks, and scope limits.",
"symbols": [
"resolve_lineage",
"adjudicate_approval",
"build_ast_doc_tree",
"compile_mission_graph",
"_run_all_exercises",
"run",
"run_batch9_bundle",
"result_card"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch9_macro_engines_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch9_macro_engines_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch9_macro_engines_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 9 Source Engines Bundle binds the accepted batch9_macro_engines_capsule component to its public source-open bundle. It checks thirteen copied source engines for provenance lineage, approval adjudication, Python AST indexing, finance headline clustering, mission graph compilation, dependency pin drift, config authority audit, heterogeneous edge extraction, WorkAtlas aggregation, host-pressure admission, doctrine enrichment, worker job budget gating, and milestone-relative quality accounting while excluding live lineage truth, approval authority, market/news truth, host-state truth, work log truth, external model access, source-file changes, public sharing, and launch-scope decision.",
"summary_line": "Set 9 Source Engines Bundle imports backend and governance source-engine bodies and exercises provenance, approval, AST, finance-news, mission graph, dependency, config, edge, WorkAtlas, host-pressure, doctrine-enrichment, worker-budget, and milestone-quality invariants.",
"title": "Set 9 Source Engines Bundle"
},
{
"anchor": "paper-module-pattern-assimilation",
"authority_ceiling": "Public fixture metadata, exported assimilation bundle metadata, copied body-import digest evidence, and metadata-only result records only; no live work log or work log mutation, source note ingestion, global doctrine changes, launch or publishing-scope decision, external model access, behavior-change proof, non-public data equivalence, or whole-system correctness.",
"body_html": "
Checkable pattern assimilation from landed evidence
\n
pattern_assimilation_step turns a pass's claim to have learned something from uncheckable commit prose into a typed record a validator can read.pattern_assimilation_step checks whether a development pass that claims to have learned something actually filed an inspectable record of it. When a component lands in the system, someone is supposed to close it out with a decision: either this run changed a named artifact, or it changed nothing and here is why. That decision is usually prose in a commit message. Prose cannot be checked. This component forces the decision into a typed record and validates the record.
\n
Over landed-component rows and their completions, it confirms each landing has one same-lane completion that names its surface, carries its no-op fields, and never promotes a local lesson into global doctrine. It runs over fixture data, not live state. Given a set of landed-component rows and the completion records that go with them, it confirms that each landed component has exactly one same-lane completion, that a claimed change names the surface it touched, that a "nothing to refine" record carries its required fields, and that no record promotes a local lesson into global doctrine. The output is a result record listing which checks passed and which named negative cases fired.
\n
sign-off.py is the entry point, with a fixture lane over the completion records directly and a bundle lane over an exported package plus copied source bodies. The entry point is src/microcosm_core/validators/sign-off.py. It has two lanes: a fixture lane that validates the completion records directly, and a bundle lane that validates an exported package of the same records plus copied source bodies.
\n
Purpose
\n
The unbacked learning claim
\n
The failure guarded against is a learning claim nobody can audit, where loose prose about improving or finding nothing drifts free of the artifacts. The failure this guards against is the unbacked learning claim. A pass says it "improved the fixture" or "found nothing to refine", and nobody can tell whether that is true. Loose prose is easy to assert and impossible to audit, so the learning story drifts away from the artifacts.
\n
Typed fields a validator can read
\n
The fix makes each completion a validator-readable record: a change names its owner surface and artifact, a no-op proves stewardship, next-best lane, and re-entry, and a missing or mismatched completion is rejected. The fix is to make each completion resolve to a record with fields a validator can read. A concrete change record must name an owner_surface and the artifact it touched. A nothing_to_refine record must prove three things: stewardship was checked, the next-best lane was considered, and a re-entry condition was written down. A landed component with no completion, or with a completion that points at a record that does not exist or does not match, is rejected.
\n
A local lesson cannot self-promote to doctrine
\n
A lesson may route to the owner surface of its artifact but never promote itself to global doctrine; a record setting claims_global_doctrine_authority is rejected outright. One constraint the component will not relax: a local lesson may route to the owner surface that owns the affected artifact, but it may not promote itself into global doctrine. A change record that sets claims_global_doctrine_authority is rejected outright. Learning has to land on a specific board with a named steward, not become a free-floating rule.
\n
How it works
\n
The fixture lane in validate_pattern_assimilation
\n
validate_pattern_assimilation loads four inputs, scans them for non-public-state leakage, runs the row checker, and passes only when every expected negative case fired, the scan is clean, and nothing blocked. The fixture lane runs in validate_pattern_assimilation. It loads four input files (_load_inputs), scans them for non-public-state leakage (_scan_fixture_inputs), runs the row checker (_validate_rows), then compares the negative cases that fired against EXPECTED_NEGATIVE_CASES. The run passes only when every expected negative case was observed, the non-public-state scan is clean, and no blocking finding was raised. Results are written by write_outputs.
\n
Per-row validation in _validate_rows
\n
_validate_rows pre-filters change and no-op records down to the valid ones, then walks each landed row to raise a missing, unresolved, or duplicate finding._validate_rows is the mechanism. It first pre-filters the completion records into valid ones. A change record is valid when it has a result record id and component id, its refinement_result is fixture_manifest_refined or validator_contract_refined, it names an owner_surface, and it does not claim global doctrine authority. A no-op record is valid when its refinement_result is nothing_to_refine and it carries stewardship_checked, next_best_lane_checked, and a reentry_condition. It then walks each landed row: a row with no completion result raises MISSING_PATTERN_ASSIMILATION_CLOSEOUT; a row whose declared completion does not resolve to a matching valid record raises PATTERN_ASSIMILATION_CLOSEOUT_RECEIPT_REF_UNRESOLVED as a blocking finding. Duplicate result record ids across both record sets raise DUPLICATE_REFINEMENT_RECEIPT_ID so one lesson cannot be counted twice.
\n
Resolving the landing result record reference
\n
_validate_public_safe_receipt_ref confirms each landing's referenced record sits inside a public root, parses as .json or .jsonl, passes the scan, and hashes, blocking any absent or unsafe citation. Each landed row also names a landing result record, and _validate_public_safe_receipt_ref resolves that reference. It confirms the ref points inside a public result record root, is a .json or .jsonl file that parses, passes the non-public-state scan, and hashes it. A ref that escapes the public roots, is missing, has the wrong suffix, or trips the scan becomes a blocking finding, so a completion cannot cite a record that is absent or unsafe.
\n
The bundle lane in run_assimilation_bundle
\n
run_assimilation_bundle runs per-part checkers over an eight-part export, then verifies copied source bodies against their recorded sha256, passing only when the scan is clean and every checker is empty. The bundle lane runs in run_assimilation_bundle. It loads an eight-part export (_load_assimilation_bundle) and runs a chain of per-part checkers: validate_exported_organ_landings, validate_exported_refinement_receipts, validate_exported_nothing_to_refine_receipts, validate_exported_closeout_receipt_links, validate_exported_stewardship_checks, validate_exported_reentry_conditions, validate_exported_next_best_lane_checks, and validate_exported_assimilation_policy. It then verifies copied source bodies through validate_source_module_manifest, which reads source_module_manifest.json, requires body_in_receipt to be false, and, for each copied module, checks its declared class is in PUBLIC_SAFE_SOURCE_BODY_CLASSES, the target file exists, and its recorded sha256 matches the file on disk. A mismatch raises ASSIMILATION_BUNDLE_SOURCE_MODULE_INVALID. The bundle passes only when the scan is clean, every checker is empty of findings, and each record set is non-empty; a stable fingerprint hash of the inputs is stamped into the result.
\n
Atomic result writes and CLI dispatch
\n
_write_jsonl_upsert atomically replaces a run's own row instead of duplicating it, and main dispatches the bundle and fixture lanes, each returning 0 on pass and 1 on block. Result records are written by _write_jsonl_upsert for the append-only run log: it drops any prior line with the same run_id, appends the new row, and swaps the file in atomically through a temp file and os.replace, so a re-run replaces its own row instead of duplicating it. The CLI dispatch is main: the validate-assimilation-bundle subcommand runs the bundle lane, and --only pattern_assimilation_step --input --out runs the fixture lane. Each returns 0 on pass, 1 on block. The stamped scope limit is the constant PATTERN_ASSIMILATION_AUTHORITY_CEILING.
\nDiagram of the mechanism (7 steps).Diagram source & refs
flowchart TD inputs["Four fixture files landings + change + no-op + missing-case"] scan["non-public-state scan _scan_fixture_inputs"] rows["Row checker _validate_rows"] filter["Pre-filter valid records change: owner surface, no doctrine upgrade no-op: stewardship, next-best lane, re-entry"] walk{"Per landed row: one completion, ref resolves to a matching valid record?"} compare["Compare fired cases vs EXPECTED_NEGATIVE_CASES"] out["Result records write_outputs, _write_jsonl_upsert"] inputs --> scan scan --> rows rows --> filter filter --> walk walk -->|resolved| compare walk -->|missing, dangling, duplicate, upgraded| compare compare --> out
\n
Negative cases
\n
A silenced check blocks the run
\n
Every required negative case must fire on the fixture, and any case that stops firing blocks the run because a check has gone silent. The fixture is built to make every required negative case fire. If any stops firing, the run blocks, because that means a check has gone silent.
\n
The reason code each attempt raises
\n
Each gaming shape maps to a named code: missing completion, missing owner, stewardship, or re-entry fields, duplicate result record, doctrine self-upgrade, a seed body in the fixture, and a bundle source mismatch.MISSING_PATTERN_ASSIMILATION_CLOSEOUT fires for a landed component with no change record and no typed no-op record. MISSING_REFINEMENT_OWNER_SURFACE, MISSING_STEWARDSHIP_CHECK, and MISSING_REENTRY_CONDITION fire for records that cannot route the lesson to an owner surface and a re-entry condition. DUPLICATE_REFINEMENT_RECEIPT_ID fires for a repeated result record id. LOCAL_LESSON_AUTHORITY_UPGRADE fires for a local lesson that claims global doctrine authority. RAW_SEED_BODY_IN_ASSIMILATION_FIXTURE fires when a seed-origin body appears in the public fixture, and the body is rejected and redacted. In the bundle lane, ASSIMILATION_BUNDLE_SOURCE_MODULE_INVALID fires on a copied-source digest or class mismatch.
\n
Prior Art Grounding
\n
Pattern languages and provenance tracing
\n
The design draws on software pattern-language practice for naming context-bound lessons and on provenance and tracing practice for citing owner surface and touched artifact, applied here to completion bookkeeping only. The idea of naming recurring engineering lessons and binding each to its context comes from software pattern-language practice; the Hillside patterns library is the direct lineage for treating patterns as shared vocabulary rather than one-off notes. The record shape borrows from provenance and tracing practice: W3C PROV informs the rule that each change record cite its owner surface and the artifact it touched, and OpenTelemetry traces are a fair analogue for linking work into an inspectable chain. This component applies those ideas to completion bookkeeping only.
\n
Validation Result record Path
\n
Running the assimilation fixture lane
\n
From microcosm-substrate, run the fixture lane and the bundle lane into an isolated /tmp output so scratch records do not overwrite tracked files: From microcosm-substrate, run the fixture lane and the bundle lane into an isolated /tmp output so scratch records do not overwrite tracked files:
A pass means the fixture completions had the required typed shape and every negative case fired; a block means a completion was missing, dangling, duplicated, over-claiming, or drifted from its digest. A pass means the supplied fixture completions have the required typed shape and every required negative case fired. A block means a completion is missing, dangling, duplicated, or over-claiming, or a copied source body drifted from its recorded digest.
\n
Scope boundary
\n
Scope limit
\n
What the evidence supports
\n
Over the fixtures and copied bodies, the proof is that completion learning has a typed, same-lane, owner-routed shape and that invalid completion claims are rejected before they can pass as doctrine. Over the supplied fixtures and copied body imports, this component shows that completion learning has a typed, same-lane, owner-routed shape and that invalid completion claims are rejected before they can pass as doctrine. It checks owner-surface change evidence, typed nothing_to_refine records, stewardship and re-entry fields, duplicate result record rejection, seed-body exclusion, public relative paths, and copied-source digest verification.
\n
What stays outside the boundary
\n
The boundary reads no private lessons, mutates no ledger, promotes no doctrine, and excludes launch; the ceiling is public metadata plus metadata-only result records, nothing more. That is the whole proof boundary. It does not read private lessons or seed bodies, mutate live ledger state, promote global doctrine, include launch operations or public sharing, use external model services, claim non-public data equivalence, prove a behavior change, or certify whole-system correctness. The scope limit is public fixture and bundle metadata plus metadata-only result records, nothing more. The generated diagram and the Atlas card are navigation surfaces; they do not upgrade a local lesson into global doctrine.
\n",
"components": [
{
"anchor": "component-pattern_assimilation_step",
"component_id": "pattern_assimilation_step",
"public_label": "Pattern Assimilation Step"
}
],
"doctrine_feed": {
"counts": {
"axiom": 8,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 11
},
"link_count": 25,
"links": [
{
"id": "pattern_assimilation_step",
"kind": "organ",
"kind_label": "component",
"label": "Pattern Assimilation Step",
"relation": "explains",
"url": "component-pattern-assimilation-step.html"
},
{
"id": "mechanism.pattern_assimilation_step.validates_public_pattern_assimilation_step",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public pattern assimilation step",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-pattern-assimilation-step-validates-public-pattern-assimilation-step"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-7",
"kind": "principle",
"kind_label": "principle",
"label": "Track known unknowns without claiming the unknown is mapped",
"relation": "governed by",
"url": "doctrine.html#dcard-p-7"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
},
{
"id": "paper_module.voice_to_doctrine_self_improvement_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Voice to Doctrine Self-Improvement Loop",
"relation": "depends on",
"url": "paper-module-voice-to-doctrine-self-improvement-loop.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.pattern_assimilation",
"reader_packet": {
"abstract": "Pattern Assimilation validates public completion-learning metadata, owner-routed refinement result records, typed nothing-to-refine decisions, and copied body-import manifests without promoting local lessons into global doctrine.",
"authority_ceiling": "Public fixture metadata, exported assimilation bundle metadata, copied body-import digest evidence, and metadata-only result records only; no live work log or work log mutation, source note ingestion, global doctrine changes, launch or publishing-scope decision, external model access, behavior-change proof, non-public data equivalence, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.pattern_assimilation.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-pattern-assimilation-diagram-1",
"html_url": "docs/paper-module-pattern-assimilation.html#paper-module-pattern-assimilation-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD inputs[\"Four fixture files landings + change + no-op + missing-case\"] scan[\"non-public-state scan _scan_fixture_inputs\"] rows[\"Row checker _validate_rows\"] filter[\"Pre-filter valid records change: owner surface, no doctrine upgrade no-op: stewardship, next-best lane, re-entry\"] walk{\"Per landed row: one completion, ref resolves to a matching valid record?\"} compare[\"Compare fired cases vs EXPECTED_NEGATIVE_CASES\"] out[\"Result records write_outputs, _write_jsonl_upsert\"] inputs --> scan scan --> rows rows --> filter filter --> walk walk -->|resolved| compare walk -->|missing, dangling, duplicate, upgraded| compare compare --> out",
"source_format": "mermaid",
"source_hash": "sha256:940c7444b32d9badecca633c8560eccdacef766cf46a93c87fb453f2fbda9c6e",
"source_line_end": 168,
"source_line_start": 151
}
],
"grouped_relation_feeds": [
{
"count": 8,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
},
{
"id": "paper_module.voice_to_doctrine_self_improvement_loop",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Voice to Doctrine Self-Improvement Loop",
"relation": "depends on",
"url": "paper-module-voice-to-doctrine-self-improvement-loop.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "pattern_assimilation_step",
"kind": "organ",
"kind_label": "component",
"label": "Pattern Assimilation Step",
"relation": "explains",
"url": "component-pattern-assimilation-step.html"
},
{
"id": "mechanism.pattern_assimilation_step.validates_public_pattern_assimilation_step",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public pattern assimilation step",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-pattern-assimilation-step-validates-public-pattern-assimilation-step"
}
],
"relation": "explains"
},
{
"count": 12,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-7",
"kind": "principle",
"kind_label": "principle",
"label": "Track known unknowns without claiming the unknown is mapped",
"relation": "governed by",
"url": "doctrine.html#dcard-p-7"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.pattern_assimilation",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 278,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/pattern_assimilation.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Hillside patterns library",
"section_id": "pattern-languages-and-provenance-tracing",
"url": "https://hillside.net/patterns/"
},
{
"kind": "external_reference",
"label": "W3C PROV",
"section_id": "pattern-languages-and-provenance-tracing",
"url": "https://www.w3.org/TR/prov-overview/"
},
{
"kind": "external_reference",
"label": "OpenTelemetry traces",
"section_id": "pattern-languages-and-provenance-tracing",
"url": "https://opentelemetry.io/docs/concepts/signals/traces/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "patternassimilation_step turns a pass's claim to have learned something from uncheckable commit prose into a typed record a validator can read.** patternassimilationstep checks whether a development pass that claims to have learned something actually filed an inspectable record of it. When a component lands in the system, someone is supposed to close it out with a decision: either this run changed a named artifact, or it changed nothing and here is why. That decision is usually prose in a commit message. Prose cannot be checked. This component forces the decision into a typed record and validates the record.",
"level": 4,
"order": 1,
"section_id": "checkable-pattern-assimilation-from-landed-evidence",
"source_line_end": 13,
"source_line_start": 3,
"title": "Checkable pattern assimilation from landed evidence"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Over landed-component rows and their completions, it confirms each landing has one same-lane completion that names its surface, carries its no-op fields, and never promotes a local lesson into global doctrine.** It runs over fixture data, not live state. Given a set of landed-component rows and the completion records that go with them, it confirms that each landed component has exactly one same-lane completion, that a claimed change names the surface it touched, that a \"nothing to refine\" record carries its required fields, and that no record promotes a local lesson into global doctrine. The output is a result record listing which checks passed and which named negative cases fired.",
"level": 4,
"order": 2,
"section_id": "over-landed-component-rows-and-their-closeouts",
"source_line_end": 26,
"source_line_start": 14,
"title": "Over landed-component rows and their completions"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**sign-off.py is the entry point, with a fixture lane over the completion records directly and a bundle lane over an exported package plus copied source bodies.** The entry point is src/microcosm_core/validators/sign-off.py. It has two lanes: a fixture lane that validates the completion records directly, and a bundle lane that validates an exported package of the same records plus copied source bodies.",
"level": 4,
"order": 3,
"section_id": "acceptance-py-is-the-entry-point",
"source_line_end": 36,
"source_line_start": 27,
"title": "`sign-off.py` is the entry point"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 38,
"source_line_start": 37,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The failure guarded against is a learning claim nobody can audit, where loose prose about improving or finding nothing drifts free of the artifacts.** The failure this guards against is the unbacked learning claim. A pass says it \"improved the fixture\" or \"found nothing to refine\", and nobody can tell whether that is true. Loose prose is easy to assert and impossible to audit, so the learning story drifts away from the artifacts.",
"level": 4,
"order": 5,
"section_id": "the-unbacked-learning-claim",
"source_line_end": 47,
"source_line_start": 39,
"title": "The unbacked learning claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The fix makes each completion a validator-readable record: a change names its owner surface and artifact, a no-op proves stewardship, next-best lane, and re-entry, and a missing or mismatched completion is rejected.** The fix is to make each completion resolve to a record with fields a validator can read. A concrete change record must name an owner_surface and the artifact it touched. A nothingtorefine record must prove three things: stewardship was checked, the next-best lane was considered, and a re-entry condition was written down. A landed component with no completion, or with a completion that points at a record that does not exist or does not match, is rejected.",
"level": 4,
"order": 6,
"section_id": "typed-fields-a-validator-can-read",
"source_line_end": 59,
"source_line_start": 48,
"title": "Typed fields a validator can read"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A lesson may route to the owner surface of its artifact but never promote itself to global doctrine; a record setting claimsglobaldoctrine_authority is rejected outright.** One constraint the component will not relax: a local lesson may route to the owner surface that owns the affected artifact, but it may not promote itself into global doctrine. A change record that sets claimsglobaldoctrine_authority is rejected outright. Learning has to land on a specific board with a named steward, not become a free-floating rule.",
"level": 4,
"order": 7,
"section_id": "a-local-lesson-cannot-self-promote-to-doctrine",
"source_line_end": 70,
"source_line_start": 60,
"title": "A local lesson cannot self-promote to doctrine"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 72,
"source_line_start": 71,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatepattern_assimilation loads four inputs, scans them for non-public-state leakage, runs the row checker, and passes only when every expected negative case fired, the scan is clean, and nothing blocked.** The fixture lane runs in validatepatternassimilation. It loads four input files (loadinputs), scans them for non-public-state leakage (scanfixtureinputs), runs the row checker (validate_rows), then compares the negative cases that fired against EXPECTEDNEGATIVECASES. The run passes only when every expected negative case was observed, the non-public-state scan is clean, and no blocking finding was raised. Results are written by write_outputs.",
"level": 4,
"order": 9,
"section_id": "the-fixture-lane-in-validate-pattern-assimilation",
"source_line_end": 84,
"source_line_start": 73,
"title": "The fixture lane in `validate_pattern_assimilation`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validaterows pre-filters change and no-op records down to the valid ones, then walks each landed row to raise a missing, unresolved, or duplicate finding.** validaterows is the mechanism. It first pre-filters the completion records into valid ones. A change record is valid when it has a result record id and component id, its refinementresult is fixturemanifest_refined or validatorcontractrefined, it names an owner_surface, and it does not claim global doctrine authority. A no-op record is valid when its refinement_result is nothingtorefine and it carries stewardship_checked, nextbestlanechecked, and a reentrycondition. It then walks each landed row: a row with no completion result raises MI\u2026",
"level": 4,
"order": 10,
"section_id": "per-row-validation-in-validate-rows",
"source_line_end": 102,
"source_line_start": 85,
"title": "Per-row validation in `_validate_rows`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatepublicsafereceipt_ref confirms each landing's referenced record sits inside a public root, parses as .json or .jsonl, passes the scan, and hashes, blocking any absent or unsafe citation.** Each landed row also names a landing result record, and validatepublicsafereceipt_ref resolves that reference. It confirms the ref points inside a public result record root, is a .json or .jsonl file that parses, passes the non-public-state scan, and hashes it. A ref that escapes the public roots, is missing, has the wrong suffix, or trips the scan becomes a blocking finding, so a completion cannot cite a record that is absent or unsafe.",
"level": 4,
"order": 11,
"section_id": "resolving-the-landing-receipt-reference",
"source_line_end": 114,
"source_line_start": 103,
"title": "Resolving the landing result record reference"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "runassimilation_bundle runs per-part checkers over an eight-part export, then verifies copied source bodies against their recorded sha256, passing only when the scan is clean and every checker is empty.** The bundle lane runs in runassimilationbundle. It loads an eight-part export (loadassimilation_bundle) and runs a chain of per-part checkers: validateexportedorganlandings, validateexportedrefinementreceipts, validateexportednothingtorefine_receipts, validateexportedcompletionreceiptlinks, validateexportedstewardshipchecks, validateexportedreentryconditions, validateexportednextbestlane_checks, and validateexportedassimilation_policy. It then verifies copied source bodies through validates\u2026",
"level": 4,
"order": 12,
"section_id": "the-bundle-lane-in-run-assimilation-bundle",
"source_line_end": 136,
"source_line_start": 115,
"title": "The bundle lane in `run_assimilation_bundle`"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "writejsonl_upsert atomically replaces a run's own row instead of duplicating it, and main dispatches the bundle and fixture lanes, each returning 0 on pass and 1 on block.** Result records are written by writejsonl_upsert for the append-only run log: it drops any prior line with the same run_id, appends the new row, and swaps the file in atomically through a temp file and os.replace, so a re-run replaces its own row instead of duplicating it. The CLI dispatch is main: the validate-assimilation-bundle subcommand runs the bundle lane, and --only patternassimilationstep --input --out runs the fixture lane. Each returns 0 on pass, 1 on block. The stamped scope limit is the constant PATTERNASSIM\u2026",
"level": 4,
"order": 13,
"section_id": "atomic-result-writes-and-cli-dispatch",
"source_line_end": 169,
"source_line_start": 137,
"title": "Atomic result writes and CLI dispatch"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 171,
"source_line_start": 170,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Every required negative case must fire on the fixture, and any case that stops firing blocks the run because a check has gone silent.** The fixture is built to make every required negative case fire. If any stops firing, the run blocks, because that means a check has gone silent.",
"level": 4,
"order": 15,
"section_id": "a-silenced-check-blocks-the-run",
"source_line_end": 178,
"source_line_start": 172,
"title": "A silenced check blocks the run"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Each gaming shape maps to a named code: missing completion, missing owner, stewardship, or re-entry fields, duplicate result record, doctrine self-upgrade, a seed body in the fixture, and a bundle source mismatch.** MISSINGPATTERNASSIMILATION_Completion fires for a landed component with no change record and no typed no-op record. MISSINGREFINEMENTOWNER_SURFACE, MISSINGSTEWARDSHIPCHECK, and MISSINGREENTRYCONDITION fire for records that cannot route the lesson to an owner surface and a re-entry condition. DUPLICATEREFINEMENTRECEIPT_ID fires for a repeated result record id. LOCALLESSONAUTHORITY_UPGRADE fires for a local lesson that claims global doctrine authority. RAWSEEDBODYINASSIMILATION_\u2026",
"level": 4,
"order": 16,
"section_id": "the-reason-code-each-attempt-raises",
"source_line_end": 195,
"source_line_start": 179,
"title": "The reason code each attempt raises"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 197,
"source_line_start": 196,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The design draws on software pattern-language practice for naming context-bound lessons and on provenance and tracing practice for citing owner surface and touched artifact, applied here to completion bookkeeping only.** The idea of naming recurring engineering lessons and binding each to its context comes from software pattern-language practice; the Hillside patterns library is the direct lineage for treating patterns as shared vocabulary rather than one-off notes. The record shape borrows from provenance and tracing practice: W3C PROV informs the rule that each change record cite its owner surface and the artifact it touched, and OpenTelemetry traces are a fair analogue for linking work\u2026",
"level": 4,
"order": 18,
"section_id": "pattern-languages-and-provenance-tracing",
"source_line_end": 213,
"source_line_start": 198,
"title": "Pattern languages and provenance tracing"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 215,
"source_line_start": 214,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "From microcosm-substrate, run the fixture lane and the bundle lane into an isolated /tmp output so scratch records do not overwrite tracked files: From microcosm-substrate, run the fixture lane and the bundle lane into an isolated /tmp output so scratch records do not overwrite tracked files: Focused test re-entry:",
"level": 4,
"order": 20,
"section_id": "running-the-assimilation-fixture-lane",
"source_line_end": 240,
"source_line_start": 216,
"title": "Running the assimilation fixture lane"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the fixture completions had the required typed shape and every negative case fired; a block means a completion was missing, dangling, duplicated, over-claiming, or drifted from its digest.** A pass means the supplied fixture completions have the required typed shape and every required negative case fired. A block means a completion is missing, dangling, duplicated, or over-claiming, or a copied source body drifted from its recorded digest.",
"level": 4,
"order": 21,
"section_id": "what-a-pass-and-a-block-mean",
"source_line_end": 250,
"source_line_start": 241,
"title": "What a pass and a block mean"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 252,
"source_line_start": 251,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Over the fixtures and copied bodies, the proof is that completion learning has a typed, same-lane, owner-routed shape and that invalid completion claims are rejected before they can pass as doctrine.** Over the supplied fixtures and copied body imports, this component shows that completion learning has a typed, same-lane, owner-routed shape and that invalid completion claims are rejected before they can pass as doctrine. It checks owner-surface change evidence, typed nothingtorefine records, stewardship and re-entry fields, duplicate result record rejection, seed-body exclusion, public relative paths, and copied-source digest verification.",
"level": 4,
"order": 23,
"section_id": "what-the-evidence-supports",
"source_line_end": 264,
"source_line_start": 253,
"title": "What the evidence supports"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The boundary reads no private lessons, mutates no ledger, promotes no doctrine, and excludes launch; the ceiling is public metadata plus metadata-only result records, nothing more.** That is the whole proof boundary. It does not read private lessons or seed bodies, mutate live ledger state, promote global doctrine, include launch operations or public sharing, use external model services, claim non-public data equivalence, prove a behavior change, or certify whole-system correctness. The scope limit is public fixture and bundle metadata plus metadata-only result records, nothing more. The generated diagram and the Atlas card are navigation surfaces; they do not upgrade a local lesson into\u2026",
"level": 4,
"order": 24,
"section_id": "what-stays-outside-the-boundary",
"source_line_end": 277,
"source_line_start": 265,
"title": "What stays outside the boundary"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/pattern_assimilation.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Pattern Assimilation validates public completion-learning metadata, owner-routed refinement result records, typed nothing-to-refine decisions, and copied body-import manifests without promoting local lessons into global doctrine."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "acceptance.py",
"path": "src/microcosm_core/validators/acceptance.py",
"resolution": "resolved",
"role": "Validator and result record-writing locus for pattern-assimilation fixtures, exported assimilation bundle validation, source-module manifest checks, negative-case enforcement, and scope limits.",
"symbols": [
"validate_pattern_assimilation",
"run_assimilation_bundle",
"validate_source_module_manifest",
"_write_jsonl_upsert",
"EXPECTED_NEGATIVE_CASES",
"PATTERN_ASSIMILATION_AUTHORITY_CEILING",
"main"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/validators/acceptance.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/pattern_assimilation.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/pattern_assimilation.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Pattern Assimilation binds the accepted pattern_assimilation_step component to the public sign-off validator, first-wave fixture, exported assimilation bundle, source-module manifest, standard row, and metadata-only result records. It checks same-lane completion decisions, owner-surface refinement evidence, stewardship and re-entry fields, duplicate result record rejection, local-lesson scope limits, source note exclusion, copied body imports, and public-relative result record paths while excluding live ledger mutation, source notes ingestion, model-output data, global doctrine changes, launch, public sharing, behavior-change proof, and non-public data equivalence.",
"summary_line": "Pattern Assimilation validates public completion-learning metadata, owner-routed refinement result records, typed nothing-to-refine decisions, and copied body-import manifests without promoting local lessons into global doctrine.",
"title": "Pattern Assimilation"
},
{
"anchor": "paper-module-batch10-governance-compilers-capsule",
"authority_ceiling": "Fixture-bound public source-body import, source-faithful public refactor evidence, deterministic exercise evidence, and metadata-only result records only; no live work log truth, live work log truth, source-file changes, publishing-scope decision, launch-scope decision, external model access, market advice, whole-system equivalence, neutral benchmark claim, or whole-system correctness.",
"body_html": "
Four governance compilers run from copied bodies
\n
Eleven governance and compiler mechanisms are copied into a public bundle and re-run against bundle-built inputs, with a result record for each that separates a computed failure from a fixture-read one. This component takes eleven governance and compiler mechanisms that run in the wider system, copies their source bodies into a public bundle, and re-runs a small source-faithful port of each one against inputs it constructs itself. It writes a result record for every mechanism: what source it came from, what the positive run produced, what each planted failure case produced, and whether the bundle computed that failure itself or only read it from a fixture.
\n
One runner exercises all compiler fixtures
\n
The problem is trust: a cold reader cannot confirm the system's own claims, so this re-runs the logic in the open and puts the computed value beside the expected one. The problem it solves is trust. When the system says a mutation gate blocks a write, or a finance horizon rejects a bad date, a cold reader has no way to confirm that from the outside. This component re-runs the logic in the open and records the computed value next to the expected one.
\n
Compiler evidence stops before live governance
\n
It exercises the batch10_governance_compilers_capsule component and grants no launch, public sharing, live-ledger, or source-file changes. It exercises batch10_governance_compilers_capsule in src/microcosm_core/organs/batch10_governance_compilers_capsule.py. It is not launch-scope decision, publishing-scope decision, live ledger truth, or source-file changes permission.
\n
Purpose
\n
Why a declared error code proves nothing
\n
A fixture that names its own expected error code proves nothing, because the declaration can be right while the ported logic is broken or absent. A fixture that declares "this case should return error code X" proves nothing on its own. The declaration can be right while the ported logic is broken, or absent. A loose bundle would let a case grade itself green by assertion.
\n
Labels are input, not verdict
\n
Each case's on-disk label and error code are treated as input, recomputed in code and compared, and a case with no computed evidence raises fixture_verdict_echo_risk and blocks the run. This component refuses that. Every negative case carries a label and a stable error code on disk, but the bundle treats those as input, not verdict. For each case it recomputes the outcome in code and compares the computed value to the expectation. The result record marks each case fixture_role: negative_case_label_not_verdict_authority and stamps the verdict computed_by_batch10_governance_compilers_capsule_integrity_matrix. A case with no computed evidence raises fixture_verdict_echo_risk and blocks the run.
\n
How it works
\n
Two entry points over one evaluator
\n
run and run_batch10_governance_compilers_bundle both feed _evaluate, which builds eleven mechanism matrices before _build_integrity_matrix writes one row each.run and run_batch10_governance_compilers_bundle are the two entry points; both hand _evaluate to a shared runner. _evaluate builds eleven mechanism matrices, then _build_integrity_matrix writes one row per mechanism.
\n
Each compiler has a controlled matrix
\n
Every governance mechanism uses its own matrix function to construct controlled input and assert the copied logic's behavior. Each mechanism has its own matrix function that constructs a controlled input and asserts the ported logic behaves.
\n
Mechanism
Matrix function
What it exercises
mutation_governance_intent_gate
_mutation_governance_matrix
classify_latest_user_intent and latest_intent_gate: a diagnostic message blocks file writes, an implement message allows a patch
observe_apply_plan_compiler
_observe_apply_compiler_matrix
_extract_operations_from_artifact: a well-formed artifact yields operations, a malformed one yields none
flagship_reviewer_persona_gauntlet_adjudicator
_reviewer_gauntlet_matrix
six reviewer personas over a public proof bundle via _persona_results_from_public_artifact and _status_from_requirements
release_public_toggle_blocker_closure_triage
_release_toggle_triage_matrix
_blocker_class: a no-go toggle with an open blocker is not operator-review-ready
publication_manifest_selector_contract_verifier
_publication_manifest_contract_matrix
_hard_exclude_selectors and _path_matches_selector: a non-public paths against a hard-exclude rule is caught
constitution_workspace_receipt_reuse_decider
_receipt_reuse_matrix
_fingerprint_rows_by_path and _changed_stable_inputs: a changed stable input invalidates result record reuse
finance_no_lookahead_temporal_contract
_finance_temporal_matrix
resolve_horizon: an unparseable horizon is rejected
session_dependency_wave_executor
_session_wave_matrix
_session_node_state: a child whose dependency failed is skipped
claim_conflict_wait_tax_detector
_claim_conflict_wait_tax_matrix
_path_overlaps: a claim inside an active scope is a conflict
role_aware_dag_block_propagation
_role_aware_dag_matrix
_resolve_blocked_groups and _resolve_group_state: a quality error on an upstream probe does not over-block
weighted_lane_width_apportionment_binding_repair
_lane_width_binding_matrix
solve_root_object_lane_widths, held as a deferred repair, not a fresh import
\n
How _source_evidence records each source
\n
_source_evidence reads the manifest and records each mechanism as a verbatim copy or a source-faithful refactor, blocking the under-bound lane-width repair rather than presenting it as an import. Source evidence is resolved per mechanism by _source_evidence. It reads the bundle's source_module_manifest.json, looks up each mechanism's named source reference, and records what it finds. A verbatim copy carries its digest status and body-copied flag. A body that cannot ship as-is is recorded as a source-faithful public refactor with the original source digest, not a copy. publication_manifest_selector_contract_verifier is a refactor because its source carried a private home-path example. weighted_lane_width_apportionment_binding_repair is classified under_bound_binding_repair, so its integrity row is forced to current_action: block rather than presented as a Set 10 import.
\n
Recomputing every negative case
\n
_compute_negative_case_probe recomputes each case from its probe_input, and a missing computation, or fewer than ten copied source modules, turns into a blocking finding. Negative cases run through _compute_negative_case_probe. For each case it reads probe_input from the fixture, computes the observed value, and compares it to the binding's expected value. A row counts as proven only when the computed value matches. _build_integrity_matrix sets fixture_verdict_echo_risk on any mechanism whose cases are missing or uncomputed, and _evaluate turns any such count into a blocking finding. The run also requires exactly ten copied source modules; a thinned bundle raises BATCH10_SOURCE_MODULE_COUNT_INVALID and blocks. result_card projects the pass into a compact card with the mechanism count and passed-mechanism count.
\nDiagram of the mechanism (6 steps).Diagram source & refs
Source refs
require 10 copied modules
Validate source_module_manifest.json
flowchart TD Input["Public input dir or exported bundle"] Manifest["Validate source_module_manifest.json require 10 copied modules"] Matrices["Build 11 mechanism matrices each on bundle-constructed input"] Integrity["_build_integrity_matrix one row per mechanism recompute each negative case"] Echo["fixture_verdict_echo_risk? any uncomputed negative case"] Result["Result record: source evidence, computed outputs, scope limit body_in_receipt: false"] Input --> Manifest --> Matrices --> Integrity --> Echo Echo -->|no risk| Result Echo -->|risk found| Result
\n
Negative cases
\n
The eleven planted cases, each recomputed rather than trusted:
\n
mutation_status_intent_blocks_writes: a diagnostic message on an evolution route sets prohibit_file_writes to true.
observe_apply_malformed_artifact_refused: an artifact with a non-list operations field yields no operations.
reviewer_missing_boundary_detected: a missing boundary doc drops the safety persona to warn.
release_toggle_no_go_fail_closed: a no-go toggle with one blocker is not operator-review-ready.
publication_hard_exclude_rejected: private/raw_seed.md against a private/ hard-exclude is rejected.
receipt_reuse_stable_input_changed: a changed stable digest counts one changed input.
finance_invalid_horizon_rejected: the horizon not-a-date is rejected.
session_failed_dependency_skips_child: node C is skipped when its dependency B failed.
claim_parent_child_overlap_reported: a claim path inside an active scope is one conflict.
dag_quality_error_not_overblocked: a quality error on an upstream probe is softened, not blocked.
lane_width_binding_deferred_to_batch9: the lane-width repair resolves to under_bound_repair_deferred_to_batch9_claim.
\n
Prior Art Grounding
\n
Policy-as-code and supply-chain assurance
\n
The design borrows the gate-and-compiler shape from Open Policy Agent, Kubernetes admission policies, and SLSA with OpenSSF Scorecard, staying fixture-bound evidence over copied or refactored source. The component follows policy-as-code, admission-control, and supply-chain assurance patterns: compile rules into deterministic checks, reject unsupported actions before they mutate state, and keep provenance for the decision. Honest anchors are Open Policy Agent, which separates policy decisions from enforcement and evaluates structured input against machine-readable rules; Kubernetes validating admission policies, which block, warn, or audit API requests before admission; and SLSA with OpenSSF Scorecard, which represent typed assurance levels, checks, and provenance for a software supply chain. This component borrows the gate-and-compiler shape for the eleven mechanisms above and stays fixture-bound evidence over copied or refactored source.
\n
Validation Result record Path
\n
Run from the microcosm-substrate/ public root:
\n
What the two commands write and prove
\n
The first command writes the result record, the second validates source and rows, and a pass means computed evidence and metadata-only records, not launch or whole-system correctness. The first command writes the integrity-matrix result record and sign-off JSON. The second validates the copied and refactored source, the source evidence, the positive and negative rows, and the scope limit fields. A pass means the mechanisms had computed source evidence, computed positive and negative rows, and metadata-only records; it does not mean launch or whole-system correctness.
\n
Scope boundary
\n
Scope limit
\n
The fixture-bound proof boundary
\n
Over the public fixtures the eleven mechanisms carry source evidence, computed rows, and metadata-only records, and every negative case is recomputed so no fixture passes itself by declaration. The strongest claim: over the public fixtures and the refreshed source-open bundle, the eleven governance and compiler mechanisms have copied or declared source evidence, computed positive and negative exercise rows, and result records that carry references, digests, and counts but no copied bodies. Every negative case is recomputed by the bundle, so a fixture cannot pass itself by declaring an error code.
\n
What stays out of reach
\n
Past that line it establishes no live ledger truth, mutation, public sharing, launch, provider, benchmark, or production authority, only fixture-bound public source-open evidence. The proof boundary stops there. This component does not establish live work log truth, live work log truth, source-file changes, publishing-scope decision, launch-scope decision, external model access, neutral benchmark evidence, whole-system equivalence, investment-related actions, or deployment posture. The scope limit is fixture-bound public source-open evidence only.
\n",
"components": [
{
"anchor": "component-batch10_governance_compilers_capsule",
"component_id": "batch10_governance_compilers_capsule",
"public_label": "Governance & Compiler Mechanisms Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 3,
"principle": 4
},
"link_count": 14,
"links": [
{
"id": "batch10_governance_compilers_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Governance & Compiler Mechanisms Bundle",
"relation": "explains",
"url": "component-batch10-governance-compilers-capsule.html"
},
{
"id": "mechanism.batch10_governance_compilers_capsule.validates_public_governance_compilers_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public governance compilers bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch10-governance-compilers-capsule-validates-public-governance-compilers-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.batch10_live_source_drift_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 10 Live Source Drift Bundle",
"relation": "depends on",
"url": "paper-module-batch10-live-source-drift-capsule.html"
},
{
"id": "paper_module.batch9_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 9 Source Engines Bundle",
"relation": "depends on",
"url": "paper-module-batch9-macro-engines-capsule.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.batch10_governance_compilers_capsule",
"reader_packet": {
"abstract": "Set 10 Governance And Compilers Bundle imports governance, compiler, launch, finance, dependency, DAG, table, reference, and recent-change source bodies as public source-open evidence without granting live ledger, public sharing, launch, market, or source-file changes.",
"authority_ceiling": "Fixture-bound public source-body import, source-faithful public refactor evidence, deterministic exercise evidence, and metadata-only result records only; no live work log truth, live work log truth, source-file changes, publishing-scope decision, launch-scope decision, external model access, market advice, whole-system equivalence, neutral benchmark claim, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch10_governance_compilers_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch10-governance-compilers-capsule-diagram-1",
"html_url": "docs/paper-module-batch10-governance-compilers-capsule.html#paper-module-batch10-governance-compilers-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Input[\"Public input dir or exported bundle\"] Manifest[\"Validate source_module_manifest.json require 10 copied modules\"] Matrices[\"Build 11 mechanism matrices each on bundle-constructed input\"] Integrity[\"_build_integrity_matrix one row per mechanism recompute each negative case\"] Echo[\"fixture_verdict_echo_risk? any uncomputed negative case\"] Result[\"Result record: source evidence, computed outputs, scope limit body_in_receipt: false\"] Input --> Manifest --> Matrices --> Integrity --> Echo Echo -->|no risk| Result Echo -->|risk found| Result",
"source_format": "mermaid",
"source_hash": "sha256:853eec4078a46b49c0c38d2709de2a1240d38f24a7197f6db7692a95f95ed2a8",
"source_line_end": 108,
"source_line_start": 97
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 3,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.batch10_live_source_drift_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 10 Live Source Drift Bundle",
"relation": "depends on",
"url": "paper-module-batch10-live-source-drift-capsule.html"
},
{
"id": "paper_module.batch9_macro_engines_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 9 Source Engines Bundle",
"relation": "depends on",
"url": "paper-module-batch9-macro-engines-capsule.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch10_governance_compilers_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Governance & Compiler Mechanisms Bundle",
"relation": "explains",
"url": "component-batch10-governance-compilers-capsule.html"
},
{
"id": "mechanism.batch10_governance_compilers_capsule.validates_public_governance_compilers_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public governance compilers bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch10-governance-compilers-capsule-validates-public-governance-compilers-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch10_governance_compilers_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 201,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch10_governance_compilers_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Open Policy Agent",
"section_id": "policy-as-code-and-supply-chain-assurance",
"url": "https://www.openpolicyagent.org/docs/latest"
},
{
"kind": "external_reference",
"label": "validating admission policies",
"section_id": "policy-as-code-and-supply-chain-assurance",
"url": "https://kubernetes.io/docs/concepts/policy"
},
{
"kind": "external_reference",
"label": "SLSA",
"section_id": "policy-as-code-and-supply-chain-assurance",
"url": "https://slsa.dev/spec/"
},
{
"kind": "external_reference",
"label": "OpenSSF Scorecard",
"section_id": "policy-as-code-and-supply-chain-assurance",
"url": "https://openssf.org/scorecard/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "Eleven governance and compiler mechanisms are copied into a public bundle and re-run against bundle-built inputs, with a result record for each that separates a computed failure from a fixture-read one. This component takes eleven governance and compiler mechanisms that run in the wider system, copies their source bodies into a public bundle, and re-runs a small source-faithful port of each one against inputs it constructs itself. It writes a result record for every mechanism: what source it came from, what the positive run produced, what each planted failure case produced, and whether the bundle computed that failure itself or only read it from a fixture.",
"level": 4,
"order": 1,
"section_id": "four-governance-compilers-run-from-copied-bodies",
"source_line_end": 12,
"source_line_start": 3,
"title": "Four governance compilers run from copied bodies"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The problem is trust: a cold reader cannot confirm the system's own claims, so this re-runs the logic in the open and puts the computed value beside the expected one. The problem it solves is trust. When the system says a mutation gate blocks a write, or a finance horizon rejects a bad date, a cold reader has no way to confirm that from the outside. This component re-runs the logic in the open and records the computed value next to the expected one.",
"level": 4,
"order": 2,
"section_id": "one-runner-exercises-all-compiler-fixtures",
"source_line_end": 19,
"source_line_start": 13,
"title": "One runner exercises all compiler fixtures"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It exercises the batch10governancecompilerscapsule component and grants no launch, public sharing, live-ledger, or source-file changes. It exercises batch10governancecompilerscapsule in src/microcosmcore/components/batch10governancecompilerscapsule.py. It is not launch-scope decision, public sharing decision, live ledger truth, or source-file changes permission.",
"level": 4,
"order": 3,
"section_id": "compiler-evidence-stops-before-live-governance",
"source_line_end": 26,
"source_line_start": 20,
"title": "Compiler evidence stops before live governance"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 28,
"source_line_start": 27,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A fixture that names its own expected error code proves nothing, because the declaration can be right while the ported logic is broken or absent. A fixture that declares \"this case should return error code X\" proves nothing on its own. The declaration can be right while the ported logic is broken, or absent. A loose bundle would let a case grade itself green by assertion.",
"level": 4,
"order": 5,
"section_id": "why-a-declared-error-code-proves-nothing",
"source_line_end": 34,
"source_line_start": 29,
"title": "Why a declared error code proves nothing"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Each case's on-disk label and error code are treated as input, recomputed in code and compared, and a case with no computed evidence raises fixtureverdictechorisk and blocks the run.** This component refuses that. Every negative case carries a label and a stable error code on disk, but the bundle treats those as input, not verdict. For each case it recomputes the outcome in code and compares the computed value to the expectation. The result record marks each case `fixture_role: negativecaselabelnotverdict_authority` and stamps the verdict computedbybatch10governancecompilerscapsuleintegrity_matrix. A case with no computed evidence raises fixtureverdictecho_risk and blocks the run.",
"level": 4,
"order": 6,
"section_id": "labels-are-input-not-verdict",
"source_line_end": 44,
"source_line_start": 35,
"title": "Labels are input, not verdict"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 46,
"source_line_start": 45,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run and runbatch10governancecompilersbundle both feed evaluate, which builds eleven mechanism matrices before buildintegritymatrix writes one row each. run and runbatch10governancecompilersbundle are the two entry points; both hand evaluate to a shared runner. evaluate builds eleven mechanism matrices, then buildintegrity_matrix writes one row per mechanism.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-over-one-evaluator",
"source_line_end": 52,
"source_line_start": 47,
"title": "Two entry points over one evaluator"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "Every governance mechanism uses its own matrix function to construct controlled input and assert the copied logic's behavior. Each mechanism has its own matrix function that constructs a controlled input and asserts the ported logic behaves.",
"level": 4,
"order": 9,
"section_id": "each-compiler-has-a-controlled-matrix",
"source_line_end": 72,
"source_line_start": 53,
"title": "Each compiler has a controlled matrix"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourceevidence reads the manifest and records each mechanism as a verbatim copy or a source-faithful refactor, blocking the under-bound lane-width repair rather than presenting it as an import. Source evidence is resolved per mechanism by source_evidence. It reads the bundle's sourcemodulemanifest.json, looks up each mechanism's named source reference, and records what it finds. A verbatim copy carries its digest status and body-copied flag. A body that cannot ship as-is is recorded as a source-faithful public refactor with the original source digest, not a copy. publicationmanifestselectorcontractverifier is a refactor because its source carried a private home-path example. weightedlanewid\u2026",
"level": 4,
"order": 10,
"section_id": "how-source-evidence-records-each-source",
"source_line_end": 84,
"source_line_start": 73,
"title": "How `_source_evidence` records each source"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "computenegativecaseprobe recomputes each case from its probeinput, and a missing computation, or fewer than ten copied source modules, turns into a blocking finding. Negative cases run through computenegativecaseprobe. For each case it reads probe_input from the fixture, computes the observed value, and compares it to the binding's expected value. A row counts as proven only when the computed value matches. buildintegritymatrix sets fixtureverdictechorisk on any mechanism whose cases are missing or uncomputed, and _evaluate turns any such count into a blocking finding. The run also requires exactly ten copied source modules; a thinned bundle raises BATCH10SOURCEMODULECOUNTINVALID and blocks\u2026",
"level": 4,
"order": 11,
"section_id": "recomputing-every-negative-case",
"source_line_end": 109,
"source_line_start": 85,
"title": "Recomputing every negative case"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The eleven planted cases, each recomputed rather than trusted: mutationstatusintentblockswrites: a diagnostic message on an evolution route sets prohibitfilewrites to true. observeapplymalformedartifactrefused: an artifact with a non-list operations field yields no operations. reviewermissingboundary_detected: a missing boundary doc drops the safety persona to warn. releasetogglenogofail_closed: a no-go toggle with one blocker is not operator-review-ready. publicationhardexcluderejected: private/rawseed.md against a private/ hard-exclude is rejected. receiptreusestableinputchanged: a changed stable digest counts one changed input. financeinvalidhorizon_rejected: the horizon not-a-date is re\u2026",
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 135,
"source_line_start": 110,
"title": "Negative cases"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "prior-art-grounding",
"source_line_end": 137,
"source_line_start": 136,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The design borrows the gate-and-compiler shape from Open Policy Agent, Kubernetes admission policies, and SLSA with OpenSSF Scorecard, staying fixture-bound evidence over copied or refactored source. The component follows policy-as-code, admission-control, and supply-chain assurance patterns: compile rules into deterministic checks, reject unsupported actions before they mutate state, and keep provenance for the decision. Honest anchors are Open Policy Agent, which separates policy decisions from enforcement and evaluates structured input against machine-readable rules; Kubernetes validating admission policies, which block, warn, or audit API requests before admission; and SLSA with OpenSSF\u2026",
"level": 4,
"order": 14,
"section_id": "policy-as-code-and-supply-chain-assurance",
"source_line_end": 153,
"source_line_start": 138,
"title": "Policy-as-code and supply-chain assurance"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from the microcosm-substrate/ public root:",
"level": 2,
"order": 15,
"section_id": "validation-receipt-path",
"source_line_end": 173,
"source_line_start": 154,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The first command writes the result record, the second validates source and rows, and a pass means computed evidence and metadata-only records, not launch or whole-system correctness. The first command writes the integrity-matrix result record and sign-off JSON. The second validates the copied and refactored source, the source evidence, the positive and negative rows, and the scope limit fields. A pass means the mechanisms had computed source evidence, computed positive and negative rows, and metadata-only records; it does not mean launch or whole-system correctness.",
"level": 4,
"order": 16,
"section_id": "what-the-two-commands-write-and-prove",
"source_line_end": 181,
"source_line_start": 174,
"title": "What the two commands write and prove"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "claim-ceiling",
"source_line_end": 183,
"source_line_start": 182,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Over the public fixtures the eleven mechanisms carry source evidence, computed rows, and metadata-only records, and every negative case is recomputed so no fixture passes itself by declaration. The strongest claim: over the public fixtures and the refreshed source-open bundle, the eleven governance and compiler mechanisms have copied or declared source evidence, computed positive and negative exercise rows, and result records that carry references, digests, and counts but no copied bodies. Every negative case is recomputed by the bundle, so a fixture cannot pass itself by declaring an error code.",
"level": 4,
"order": 18,
"section_id": "the-fixture-bound-proof-boundary",
"source_line_end": 192,
"source_line_start": 184,
"title": "The fixture-bound proof boundary"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Past that line it establishes no live ledger truth, mutation, public sharing, launch, provider, benchmark, or production authority, only fixture-bound public source-open evidence. The proof boundary stops there. This component does not establish live work log truth, live work log truth, source-file changes, public sharing approval, launch-scope decision, external model access, neutral benchmark evidence, whole-system equivalence, investment-related actions, or deployment posture. The scope limit is fixture-bound public source-open evidence only.",
"level": 4,
"order": 19,
"section_id": "what-stays-out-of-reach",
"source_line_end": 200,
"source_line_start": 193,
"title": "What stays out of reach"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch10_governance_compilers_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 10 Governance And Compilers Bundle imports governance, compiler, launch, finance, dependency, DAG, table, reference, and recent-change source bodies as public source-open evidence without granting live ledger, public sharing, launch, market, or source-file changes."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch10_governance_compilers_capsule.py",
"path": "src/microcosm_core/organs/batch10_governance_compilers_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for Set 10 governance/compiler source-open imports, deterministic mechanism exercises, integrity-matrix source evidence, negative cases, exact-copy digest checks, and scope limits.",
"symbols": [
"_build_integrity_matrix",
"_source_evidence",
"_evaluate",
"run",
"run_batch10_governance_compilers_bundle",
"result_card",
"EXPECTED_MECHANISMS",
"EXPECTED_NEGATIVE_CASES"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch10_governance_compilers_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch10_governance_compilers_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch10_governance_compilers_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 10 Governance And Compilers Bundle binds the accepted batch10_governance_compilers_capsule component to a refreshed source-open bundle. It exercises governed-mutation intent, observe/apply compilation, public-proof review, launch blocker triage, public sharing path contracts, result record reuse, no-lookahead horizons, session-wave execution, claim-conflict wait tax, role-aware DAG blocking, frontend table shaping, reference grouping, recent-change coalescing, and the deferred Set-9 lane-width repair while preserving copied source digests, negative cases, and scope limits.",
"summary_line": "Set 10 Governance And Compilers Bundle imports governance, compiler, launch, finance, dependency, DAG, table, reference, and recent-change source bodies as public source-open evidence without granting live ledger, public sharing, launch, market, or source-file changes.",
"title": "Set 10 Governance And Compilers Bundle"
},
{
"anchor": "paper-module-batch11-saturation-engines-capsule",
"authority_ceiling": "Fixture-bound public source-body import, source-faithful public port evidence, computed negative-probe evidence, and metadata-only result records only; no live work log truth, navigation authority, complete secret detection, live market data, investment-related actions, raw transcript authority, video capture, source-file changes, publishing-scope decision, launch-scope decision, external model access, whole-system equivalence, or whole-system correctness.",
"body_html": "
Four saturation engines run from copied source
\n
Thirteen unrelated pieces of internal logic are copied into a public bundle and re-run on synthetic inputs so a reader can watch each behave. This component takes thirteen unrelated pieces of internal logic, copies their source bodies into a public bundle, and re-runs each one against small synthetic inputs so a reader can watch the logic behave instead of taking a claim on trust. The targets share nothing except that treatment: a run ranker, a cluster-insight deriver, a delta gate, a navigation grader, a supersession rollup, a browse sectioner, an evidence selector, a secret scanner, two market-board engines, a graph wayfinder, a session-diagnostic lens, and a demo-coverage auditor.
\n
Each engine answers a distinct bounded question
\n
One narrow question: does each imported mechanism actually compute the guard it claims, on inputs designed to make it fail? The single question it answers is narrow. For each mechanism, does the imported source actually compute the guard it claims to, on inputs designed to make it fail? It is a saturation pass because the targets are swept up together so one reviewer can audit a broad slice of the codebase from one place.
\n
The result reports saturation without acting on it
\n
The records are metadata-only: digests, computed outputs, probe values, and verdicts, while copied source bodies stay in the bundle tree. It emits metadata-only result records: source-module digests, per-mechanism computed outputs, computed negative-case probe values, and pass or blocked verdicts. Copied source bodies stay in the bundle's source_modules tree and are never inlined into a record.
\n
Purpose
\n
Why a fixture filename proves nothing
\n
A fixture filename is only a label, and a runner that counts it without exercising the real function is green and blind at once. A test that trusts its own fixture filenames proves nothing. A file named run_affinity_stale_terminal_rejected is a label. If the runner counts that label as a rejection without exercising the real function, the suite is green and blind at the same time. That is the failure this component is built to avoid.
\n
Two agreeing computations per negative case
\n
A case is verified only when the mechanism's own exercise and a separate probe over its probe_input agree, so an undemonstrated failure is flagged. So it never lets a label stand in for a result. For each negative case it recomputes the guard from the fixture's own probe_input and refuses to mark the case verified unless two independent computations agree: the mechanism's own exercise and a separate probe over the fixture input. A fixture that asserts a failure it cannot demonstrate is flagged, not counted.
\n
How it works
\n
Two entry points, one fixed pipeline
\n
Both entry points forward to _run_batch11_crown_jewel_organ, which drives the whole pipeline in one fixed order.run and run_batch11_saturation_engines_bundle are the two entry points. Both forward to _run_batch11_crown_jewel_organ, which drives the pipeline in a fixed order.
\n
Manifest and digest checks before any logic
\n
validate_source_manifest checks each copied module's digest, line count, and anchors first, so thirteen mechanisms rest on twelve modules before any logic runs. First it loads the required input batch11_saturation_engines_capsule_probe_manifest.json. Then validate_source_manifest checks each copied module against its recorded digest, line count, and required source anchors, so a drifted or partial copy is caught before any logic runs. EXPECTED_MODULE_IDS names the twelve copied modules; the market-board engines share one copied quant_presentation_mart source, which is why thirteen mechanisms rest on twelve modules.
\n
Thirteen exercises in fixed order in _evaluate
\n
_evaluate runs each copied function in the fixed EXPECTED_MECHANISMS order, and any reorder, block, or wrong module count becomes a finding._evaluate then exercises all thirteen mechanisms in the order fixed by EXPECTED_MECHANISMS. Each exercise runs the copied function and returns a small computed output. _run_affinity_matrix, for example, calls the copied recommend_run_by_affinity over two candidate runs and records whether a stale terminal run was refused. Any reordering, blocked exercise, or wrong module count becomes a finding: BATCH11_MECHANISM_ORDER_INVALID, BATCH11_MECHANISM_BLOCKED, or BATCH11_SOURCE_MODULE_COUNT_INVALID.
\n
Echo-risk pairing in _build_integrity_matrix
\n
Each row pairs the mechanism's output with an independent probe over probe_input, and any disagreement sets echo risk that blocks the whole run._build_integrity_matrix is where the negative cases are kept honest. For each mechanism it reads two values. One is the mechanism's own computed output. The other comes from _compute_negative_case_probe, which recomputes the case from the fixture's probe_input using the same copied function, then digests that input into fixture_probe_input_digest. A row's negative result counts as computed only when both values are true. When they disagree, the row's fixture_verdict_echo_risk is set. A non-zero fixture_verdict_echo_risk_count raises BATCH11_FIXTURE_VERDICT_ECHO_RISK and blocks the whole run. _source_evidence attaches the per-mechanism digest and anchor rows so each verdict is traceable to a specific copied module.
\n
Forbidden-class scan and result record body scan
\n
After validation the runner scans fixtures against forbidden classes and passes the payload through a body scan that blocks any leaked copied body. After evaluation the runner calls validate_negative_cases, then scans the fixture files with scan_paths against the forbidden classes loaded from core/private_state_forbidden_classes.json. Finally it writes the result, board, and validation records, and passes the assembled payload through scan_receipt_payload_for_bodies. If a copied body leaked into a record, that scan flips the status to blocked with CROWN_JEWEL_RECEIPT_BODY_SCAN_BLOCKED. result_card projects the passing payload into a compact card; main is the command-line dispatch for the run and validate-bundle actions.
\n
Tier and binding disposition per row
\n
Each row is tagged by verification tier and by a binding disposition recording how the mechanism relates to the rest of the system. Each row is tagged by tier and binding disposition. The six mechanisms in TIER_B_MECHANISMS are marked B_verified_by_controller; the rest are A_adversarially_verified_in_cap. MECHANISM_BINDING_DISPOSITIONS records how each mechanism relates to the rest of the system: a new import, an already-bound gate the bundle only re-checks, an under-bound path it extends, or a mechanism absent from the public component plane at import time.
\n
Function
Role
_run_batch11_crown_jewel_organ
Orders the pipeline and writes the metadata-only records
_evaluate
Runs the thirteen exercises and raises order, blocked, and count findings
_build_integrity_matrix
Pairs each mechanism output with an independent probe and sets echo risk
_compute_negative_case_probe
Recomputes each negative case from the fixture's own probe_input
_source_evidence
Attaches per-mechanism copied-source digest and anchor rows
result_card
Projects a passing result into a compact reader card
\nDiagram of the mechanism (7 steps).Diagram source & refs
Source refs
digest and anchor checks
validate_source_manifest
flowchart TD input["probe manifest + copied source modules"] manifest["validate_source_manifest digest and anchor checks"] evaluate["_evaluate 13 exercises in fixed order"] matrix["_build_integrity_matrix mechanism value AND fixture probe"] probe["_compute_negative_case_probe recompute from probe_input"] scan["scan_paths + result record body scan"] record["metadata-only result record"] input --> manifest manifest --> evaluate evaluate --> matrix probe --> matrix matrix --> scan scan --> record
\n
Negative cases
\n
Thirteen named cases as claims to satisfy
\n
EXPECTED_NEGATIVE_CASES declares one named case per mechanism, each an error-coded claim the copied code must actually satisfy on breaking inputs.EXPECTED_NEGATIVE_CASES declares thirteen named cases, one per mechanism, each with an error code. They read as claims the code must actually satisfy. A stale terminal run is refused even when made sticky and feed-rich (run_affinity_stale_terminal_rejected). A zero-energy calculator bucket produces no fake dominance share (calculator_zero_bucket_no_fake_dominance). A new gap blocks the delta gate (std_python_new_gap_blocks). A wrong route is graded down (exogenous_nav_wrong_route_graded_down). An unresolved hard failure blocks the portability rollup (portability_unresolved_hard_fail_blocks). A synthetic key shape blocks the projection scan (projection_secret_token_blocks). An unreachable navigation target returns a blocker rather than a fabricated path (frontend_nav_unreachable_target). A demo take missing its anchors scores below a complete one (demo_take_missing_anchor_penalized).
\n
Synthetic secret-scan fixture, never live material
\n
The probe recomputes each outcome from the fixture input, and the secret-scan case uses synthetic test strings, never live material, that must be blocked. For each case the probe recomputes the outcome from the fixture input. The secret-scan example is deliberate: the fixture carries a synthetic key shape and a private ledger path, both test strings, never live material, and the probe confirms both are blocked.
\n
Prior Art Grounding
\n
Characterization testing and the test-oracle problem
\n
The lineage is characterization testing plus the test-oracle problem: a stored expectation is no oracle unless an independent probe from the fixture input agrees. The method is characterization testing: pin a piece of existing code by exercising it against fixtures and recording what it actually does, rather than asserting a value someone typed by hand. The negative-case discipline is a version of the test-oracle problem. A stored expectation is not an oracle if it can pass without the code running, so the probe computes an independent expectation from the fixture's own input and requires the two to agree. There is no external specification these engines implement, so the honest lineage is that engineering practice, not a standard. The links in earlier drafts to overload and backpressure literature described a different kind of saturation and are dropped here.
\n
Validation Result record Path
\n
Fixture files are inputs, not verdicts
\n
Each fixture is a public probe_input the component recomputes into recorded probe, fixture, and mechanism values before counting a case verified. Negative-case fixture files are inputs, not verdicts. Each file carries a public probe_input; the component computes the corresponding probe and records fixture_probe_input_digest, fixture_computed_value, and mechanism_computed_value in the integrity matrix before counting a case as verified.
\n
Run these from the microcosm-substrate/ public root:
\n
What a pass certifies
\n
A pass means digests and anchors matched, all thirteen exercises ran in order, every case reproduced with zero echo risk, and no body leaked. A pass means every copied module matched its digest and anchors, all thirteen exercises ran in order, every negative case was reproduced from its own input with zero echo risk, and no copied body reached a record. It is fixture evidence about this standalone component, nothing wider.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The copied source bodies compute their guards on breaking fixtures, records carry only refs and verdicts, and no fixture label can pass on its own. The strongest claim the evidence supports is this: the copied Set-11 source bodies compute the guards they claim to, on public fixtures built to break them, and the result records carry only refs, digests, counts, and computed verdicts. Each negative case is reproduced from its own input by an independent probe, so a fixture label cannot pass on its own.
\n
The scope limit holds everything false
\n
AUTHORITY_CEILING sets provider, model, mutation, public sharing, navigation, secret-detection, market, and video flags all false, so a green run is bounded evidence about copied logic only. The proof boundary stops there. The scope limit, fixed in the AUTHORITY_CEILING constant, sets external model access, model dispatch, source-file changes, public sharing, launch, work-ledger authority, navigation authority, complete secret detection, market advice, raw-session truth, and video capture all to false. This component is not live work log truth, not navigation authority, not complete secret detection, not live market data, not investment-related actions, not raw transcript authority, not video capture, not source-file changes, and not launch or publishing-scope decision. A green run is bounded evidence about copied logic, not a statement about the whole system.
\n",
"components": [
{
"anchor": "component-batch11_saturation_engines_capsule",
"component_id": "batch11_saturation_engines_capsule",
"public_label": "Saturation Engines Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 4,
"principle": 4
},
"link_count": 15,
"links": [
{
"id": "batch11_saturation_engines_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Saturation Engines Bundle",
"relation": "explains",
"url": "component-batch11-saturation-engines-capsule.html"
},
{
"id": "mechanism.batch11_saturation_engines_capsule.validates_public_saturation_engines_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public saturation engines bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch11-saturation-engines-capsule-validates-public-saturation-engines-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.batch8_compliance_pipeline_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Compliance Pipeline Bundle",
"relation": "depends on",
"url": "paper-module-batch8-compliance-pipeline-capsule.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch8_compliance_pipeline_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Compliance Pipeline Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[59:paper_module.batch8_compliance_pipeline_capsule].depends_on",
"url": "paper-module-batch8-compliance-pipeline-capsule.html"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[19:paper_module.macro_projection_import_protocol].depends_on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"paper_module_id": "paper_module.batch11_saturation_engines_capsule",
"reader_packet": {
"abstract": "Set 11 Saturation Engines Bundle imports saturation, diagnostic, wayfinding, market-board, secret-scan, and demo-take source bodies as public source-open evidence without granting live runtime, launch, market, or navigation authority.",
"authority_ceiling": "Fixture-bound public source-body import, source-faithful public port evidence, computed negative-probe evidence, and metadata-only result records only; no live work log truth, navigation authority, complete secret detection, live market data, investment-related actions, raw transcript authority, video capture, source-file changes, publishing-scope decision, launch-scope decision, external model access, whole-system equivalence, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch11_saturation_engines_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch11-saturation-engines-capsule-diagram-1",
"html_url": "docs/paper-module-batch11-saturation-engines-capsule.html#paper-module-batch11-saturation-engines-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD input[\"probe manifest + copied source modules\"] manifest[\"validate_source_manifest digest and anchor checks\"] evaluate[\"_evaluate 13 exercises in fixed order\"] matrix[\"_build_integrity_matrix mechanism value AND fixture probe\"] probe[\"_compute_negative_case_probe recompute from probe_input\"] scan[\"scan_paths + result record body scan\"] record[\"metadata-only result record\"] input --> manifest manifest --> evaluate evaluate --> matrix probe --> matrix matrix --> scan scan --> record",
"source_format": "mermaid",
"source_hash": "sha256:9849c0d96c971ace33b9c6696d936eeae37108405e795f5051402ab5f9ac6ec2",
"source_line_end": 76,
"source_line_start": 60
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 4,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.batch8_compliance_pipeline_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Compliance Pipeline Bundle",
"relation": "depends on",
"url": "paper-module-batch8-compliance-pipeline-capsule.html"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch11_saturation_engines_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Saturation Engines Bundle",
"relation": "explains",
"url": "component-batch11-saturation-engines-capsule.html"
},
{
"id": "mechanism.batch11_saturation_engines_capsule.validates_public_saturation_engines_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public saturation engines bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch11-saturation-engines-capsule-validates-public-saturation-engines-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch11_saturation_engines_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 132,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch11_saturation_engines_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "Thirteen unrelated pieces of internal logic are copied into a public bundle and re-run on synthetic inputs so a reader can watch each behave. This component takes thirteen unrelated pieces of internal logic, copies their source bodies into a public bundle, and re-runs each one against small synthetic inputs so a reader can watch the logic behave instead of taking a claim on trust. The targets share nothing except that treatment: a run ranker, a cluster-insight deriver, a delta gate, a navigation grader, a supersession rollup, a browse sectioner, an evidence selector, a secret scanner, two market-board engines, a graph wayfinder, a session-diagnostic lens, and a demo-coverage auditor.",
"level": 4,
"order": 1,
"section_id": "four-saturation-engines-run-from-copied-source",
"source_line_end": 6,
"source_line_start": 3,
"title": "Four saturation engines run from copied source"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "One narrow question: does each imported mechanism actually compute the guard it claims, on inputs designed to make it fail? The single question it answers is narrow. For each mechanism, does the imported source actually compute the guard it claims to, on inputs designed to make it fail? It is a saturation pass because the targets are swept up together so one reviewer can audit a broad slice of the codebase from one place.",
"level": 4,
"order": 2,
"section_id": "each-engine-answers-a-distinct-bounded-question",
"source_line_end": 10,
"source_line_start": 7,
"title": "Each engine answers a distinct bounded question"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The records are metadata-only: digests, computed outputs, probe values, and verdicts, while copied source bodies stay in the bundle tree. It emits metadata-only result records: source-module digests, per-mechanism computed outputs, computed negative-case probe values, and pass or blocked verdicts. Copied source bodies stay in the bundle's source_modules tree and are never inlined into a record.",
"level": 4,
"order": 3,
"section_id": "the-result-reports-saturation-without-acting-on-it",
"source_line_end": 14,
"source_line_start": 11,
"title": "The result reports saturation without acting on it"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A fixture filename is only a label, and a runner that counts it without exercising the real function is green and blind at once. A test that trusts its own fixture filenames proves nothing. A file named runaffinitystaleterminalrejected is a label. If the runner counts that label as a rejection without exercising the real function, the suite is green and blind at the same time. That is the failure this component is built to avoid.",
"level": 4,
"order": 5,
"section_id": "why-a-fixture-filename-proves-nothing",
"source_line_end": 20,
"source_line_start": 17,
"title": "Why a fixture filename proves nothing"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A case is verified only when the mechanism's own exercise and a separate probe over its probeinput agree, so an undemonstrated failure is flagged. So it never lets a label stand in for a result. For each negative case it recomputes the guard from the fixture's own probeinput and refuses to mark the case verified unless two independent computations agree: the mechanism's own exercise and a separate probe over the fixture input. A fixture that asserts a failure it cannot demonstrate is flagged, not counted.",
"level": 4,
"order": 6,
"section_id": "two-agreeing-computations-per-negative-case",
"source_line_end": 24,
"source_line_start": 21,
"title": "Two agreeing computations per negative case"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Both entry points forward to runbatch11crownjewelorgan, which drives the whole pipeline in one fixed order. run and runbatch11saturationenginesbundle are the two entry points. Both forward to runbatch11crownjewelorgan, which drives the pipeline in a fixed order.",
"level": 4,
"order": 8,
"section_id": "two-entry-points-one-fixed-pipeline",
"source_line_end": 30,
"source_line_start": 27,
"title": "Two entry points, one fixed pipeline"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validatesourcemanifest checks each copied module's digest, line count, and anchors first, so thirteen mechanisms rest on twelve modules before any logic runs. First it loads the required input batch11saturationenginescapsuleprobemanifest.json. Then validatesourcemanifest checks each copied module against its recorded digest, line count, and required source anchors, so a drifted or partial copy is caught before any logic runs. EXPECTEDMODULEIDS names the twelve copied modules; the market-board engines share one copied quantpresentation_mart source, which is why thirteen mechanisms rest on twelve modules.",
"level": 4,
"order": 9,
"section_id": "manifest-and-digest-checks-before-any-logic",
"source_line_end": 34,
"source_line_start": 31,
"title": "Manifest and digest checks before any logic"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluate runs each copied function in the fixed EXPECTEDMECHANISMS order, and any reorder, block, or wrong module count becomes a finding. evaluate then exercises all thirteen mechanisms in the order fixed by EXPECTEDMECHANISMS. Each exercise runs the copied function and returns a small computed output. runaffinitymatrix, for example, calls the copied recommendrunbyaffinity over two candidate runs and records whether a stale terminal run was refused. Any reordering, blocked exercise, or wrong module count becomes a finding: BATCH11MECHANISMORDERINVALID, BATCH11MECHANISMBLOCKED, or BATCH11SOURCEMODULECOUNTINVALID.",
"level": 4,
"order": 10,
"section_id": "thirteen-exercises-in-fixed-order-in-evaluate",
"source_line_end": 38,
"source_line_start": 35,
"title": "Thirteen exercises in fixed order in `_evaluate`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Each row pairs the mechanism's output with an independent probe over probeinput, and any disagreement sets echo risk that blocks the whole run. buildintegritymatrix is where the negative cases are kept honest. For each mechanism it reads two values. One is the mechanism's own computed output. The other comes from computenegativecaseprobe, which recomputes the case from the fixture's probeinput using the same copied function, then digests that input into fixtureprobeinputdigest. A row's negative result counts as computed only when both values are true. When they disagree, the row's fixtureverdictechorisk is set. A non-zero fixtureverdictechoriskcount raises BATCH11FIXTUREVERDICTECHORISK and\u2026",
"level": 4,
"order": 11,
"section_id": "echo-risk-pairing-in-build-integrity-matrix",
"source_line_end": 42,
"source_line_start": 39,
"title": "Echo-risk pairing in `_build_integrity_matrix`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "After validation the runner scans fixtures against forbidden classes and passes the payload through a body scan that blocks any leaked copied body. After evaluation the runner calls validatenegativecases, then scans the fixture files with scanpaths against the forbidden classes loaded from core/privatestateforbiddenclasses.json. Finally it writes the result, board, and validation records, and passes the assembled payload through scanreceiptpayloadforbodies. If a copied body leaked into a record, that scan flips the status to blocked with CROWNJEWELRECEIPTBODYSCANBLOCKED. resultcard projects the passing payload into a compact card; main is the command-line dispatch for the run and validate-b\u2026",
"level": 4,
"order": 12,
"section_id": "forbidden-class-scan-and-receipt-body-scan",
"source_line_end": 46,
"source_line_start": 43,
"title": "Forbidden-class scan and result record body scan"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "Each row is tagged by verification tier and by a binding disposition recording how the mechanism relates to the rest of the system. Each row is tagged by tier and binding disposition. The six mechanisms in TIERBMECHANISMS are marked Bverifiedbycontroller; the rest are Aadversariallyverifiedincap. MECHANISMBINDING_DISPOSITIONS records how each mechanism relates to the rest of the system: a new import, an already-bound gate the bundle only re-checks, an under-bound path it extends, or a mechanism absent from the public component plane at import time.",
"level": 4,
"order": 13,
"section_id": "tier-and-binding-disposition-per-row",
"source_line_end": 77,
"source_line_start": 47,
"title": "Tier and binding disposition per row"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 79,
"source_line_start": 78,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVECASES declares one named case per mechanism, each an error-coded claim the copied code must actually satisfy on breaking inputs. EXPECTEDNEGATIVECASES declares thirteen named cases, one per mechanism, each with an error code. They read as claims the code must actually satisfy. A stale terminal run is refused even when made sticky and feed-rich (runaffinitystaleterminalrejected). A zero-energy calculator bucket produces no fake dominance share (calculatorzerobucketnofakedominance). A new gap blocks the delta gate (stdpythonnewgapblocks). A wrong route is graded down (exogenousnavwrongroutegradeddown). An unresolved hard failure blocks the portability rollup (portabilityunreso\u2026",
"level": 4,
"order": 15,
"section_id": "thirteen-named-cases-as-claims-to-satisfy",
"source_line_end": 83,
"source_line_start": 80,
"title": "Thirteen named cases as claims to satisfy"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The probe recomputes each outcome from the fixture input, and the secret-scan case uses synthetic test strings, never live material, that must be blocked. For each case the probe recomputes the outcome from the fixture input. The secret-scan example is deliberate: the fixture carries a synthetic key shape and a private ledger path, both test strings, never live material, and the probe confirms both are blocked.",
"level": 4,
"order": 16,
"section_id": "synthetic-secret-scan-fixture-never-live-material",
"source_line_end": 87,
"source_line_start": 84,
"title": "Synthetic secret-scan fixture, never live material"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 89,
"source_line_start": 88,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The lineage is characterization testing plus the test-oracle problem: a stored expectation is no oracle unless an independent probe from the fixture input agrees. The method is characterization testing: pin a piece of existing code by exercising it against fixtures and recording what it actually does, rather than asserting a value someone typed by hand. The negative-case discipline is a version of the test-oracle problem. A stored expectation is not an oracle if it can pass without the code running, so the probe computes an independent expectation from the fixture's own input and requires the two to agree. There is no external specification these engines implement, so the honest lineage is\u2026",
"level": 4,
"order": 18,
"section_id": "characterization-testing-and-the-test-oracle-problem",
"source_line_end": 93,
"source_line_start": 90,
"title": "Characterization testing and the test-oracle problem"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 95,
"source_line_start": 94,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Each fixture is a public probeinput the component recomputes into recorded probe, fixture, and mechanism values before counting a case verified. Negative-case fixture files are inputs, not verdicts. Each file carries a public probeinput; the component computes the corresponding probe and records fixtureprobeinputdigest, fixturecomputedvalue, and mechanismcomputed_value in the integrity matrix before counting a case as verified. Run these from the microcosm-substrate/ public root:",
"level": 4,
"order": 20,
"section_id": "fixture-files-are-inputs-not-verdicts",
"source_line_end": 117,
"source_line_start": 96,
"title": "Fixture files are inputs, not verdicts"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means digests and anchors matched, all thirteen exercises ran in order, every case reproduced with zero echo risk, and no body leaked. A pass means every copied module matched its digest and anchors, all thirteen exercises ran in order, every negative case was reproduced from its own input with zero echo risk, and no copied body reached a record. It is fixture evidence about this standalone component, nothing wider.",
"level": 4,
"order": 21,
"section_id": "what-a-pass-certifies",
"source_line_end": 121,
"source_line_start": 118,
"title": "What a pass certifies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 123,
"source_line_start": 122,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The copied source bodies compute their guards on breaking fixtures, records carry only refs and verdicts, and no fixture label can pass on its own. The strongest claim the evidence supports is this: the copied Set-11 source bodies compute the guards they claim to, on public fixtures built to break them, and the result records carry only refs, digests, counts, and computed verdicts. Each negative case is reproduced from its own input by an independent probe, so a fixture label cannot pass on its own.",
"level": 4,
"order": 23,
"section_id": "the-strongest-supported-claim",
"source_line_end": 127,
"source_line_start": 124,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "AUTHORITYCEILING sets provider, model, mutation, public sharing, navigation, secret-detection, market, and video flags all false, so a green run is bounded evidence about copied logic only. The proof boundary stops there. The scope limit, fixed in the AUTHORITYCEILING constant, sets external model access, model dispatch, source-file changes, public sharing, launch, work-ledger authority, navigation authority, complete secret detection, market advice, raw-session truth, and video capture all to false. This component is not live work log truth, not navigation authority, not complete secret detection, not live market data, not investment-related actions, not raw transcript authority, not video\u2026",
"level": 4,
"order": 24,
"section_id": "the-authority-ceiling-holds-everything-false",
"source_line_end": 131,
"source_line_start": 128,
"title": "The scope limit holds everything false"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch11_saturation_engines_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 11 Saturation Engines Bundle imports saturation, diagnostic, wayfinding, market-board, secret-scan, and demo-take source bodies as public source-open evidence without granting live runtime, launch, market, or navigation authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch11_saturation_engines_capsule.py",
"path": "src/microcosm_core/organs/batch11_saturation_engines_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for Set 11 saturation-engine source imports, deterministic source-faithful exercises, computed negative probes, exact-copy digest checks, and scope limits.",
"symbols": [
"_build_integrity_matrix",
"_source_evidence",
"_evaluate",
"run",
"run_batch11_saturation_engines_bundle",
"result_card",
"EXPECTED_MECHANISMS",
"EXPECTED_NEGATIVE_CASES"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch11_saturation_engines_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch11_saturation_engines_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch11_saturation_engines_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 11 Saturation Engines Bundle binds the accepted batch11_saturation_engines_capsule component to a refreshed source-open bundle. It exercises run-affinity scoring, calculator cluster insight derivation, std_python delta gating, exogenous navigation grading, portability supersession rollup, shard browse priority, holographic evidence selection, projection secret scanning, stockgrid flow normalization, source-regime board bucketing, frontend wayfinding, agent-session diagnostic lenses, and demo-take coverage auditing while preserving copied source digests, computed negative probes, and scope limits.",
"summary_line": "Set 11 Saturation Engines Bundle imports saturation, diagnostic, wayfinding, market-board, secret-scan, and demo-take source bodies as public source-open evidence without granting live runtime, launch, market, or navigation authority.",
"title": "Set 11 Saturation Engines Bundle"
},
{
"anchor": "paper-module-batch4-proof-authority-runtime",
"authority_ceiling": "Fixture-bound public source-body import, static proof-placeholder checks, dry-run authority-boundary evidence, source-anchor evidence, negative-case evidence, and metadata-only result records only; no theorem success, Erdos #257 solution, official benchmark claims, live sandbox enforcement, live Codex orchestration, external model access, source-file changes, publishing-scope decision, launch-scope decision, whole-system equivalence, or whole-system correctness.",
"body_html": "
Proof-authority claims are recomputed from source
\n
A single component bundles seven easily overclaimed source files into one inspectable unit: proof search, an Erdos #257 Lean kernel, reasoning-authority fences, completion, bitemporal logging, task-policy wrapping, and context-yield.batch4_proof_authority_runtime is one Microcosm component that gathers a group of source files that are easy to overclaim and checks them as a single inspectable unit. The files cover proof search, a copied Lean certificate kernel for Erdos #257, reasoning-authority fences, completion planning, bitemporal claim logging, macOS task-policy wrapping, and context-yield attribution.
\n
Two runtime paths share one authority boundary
\n
It never runs those systems: it reads their copied public source, checks each mechanism against required anchors, recomputes one negative case per mechanism, and writes a metadata-only result record. The component does not run any of those systems. It reads their copied public source, checks each declared mechanism against named required anchors, recomputes a negative case per mechanism from that source, and writes a result record. The record carries source references, digests, anchor counts, and bounded outcomes. It does not carry source bodies or live private state.
\n
A passing fixture does not establish a theorem
\n
The only positive claim is narrow: copied source is present, its required anchors are in place, and the copied Lean kernel is free of proof placeholders, with everything stronger refused openly. Its positive claim is narrow. The copied source is present, its required anchors are in place, and the copied Lean kernel is free of proof placeholders. Everything stronger is refused in the open.
\n
Purpose
\n
Each mechanism invites an overclaim
\n
Each bundled mechanism looks like more than it is: proof search like solving open problems, the Lean kernel like a solution, a grant like a live sandbox, a wrapper like orchestration. These mechanisms sit next to claims a reader will want to make on their behalf. A proof-search benchmark looks like solving open problems. A copied CertificateKernel.lean for Erdos #257 looks like a solution. A reasoning grant looks like a live sandbox. A task-policy wrapper looks like live orchestration.
\n
One narrow question per mechanism
\n
The component asks only whether each mechanism can be shown as copied, anchored public source without inheriting an authority it lacks, making the import auditable. A loose import surface would let any of those read as more than it is. The component answers one narrow question per mechanism: can this be shown to a reader as copied, anchored, public source, without inheriting an authority it does not have. It makes the import auditable and forces every stronger statement into a visible limit.
\n
How it works
\n
Two entry shapes over one gate
\n
run takes a fixture directory and run_batch4_bundle takes the exported bundle, but both forward to run_crown_jewel_organ with _evaluate, so they share one gate. There are two entry shapes. run consumes a fixture input directory. run_batch4_bundle consumes the exported bundle directory. Both forward to run_crown_jewel_organ with _evaluate as the evaluator, so they share one gate. main exposes them as the run and validate-bundle subcommands.
_module_rows builds a module-id to row map from the exported source_module_manifest.json. _evaluate checks that all seventeen ids in EXPECTED_MODULE_IDS are present; a gap emits BATCH4_SOURCE_MODULES_MISSING.
It checks the manifest's mechanism set equals the twelve ids in EXPECTED_MECHANISMS exactly. Any missing or extra id emits BATCH4_MECHANISM_SET_MISMATCH.
_mechanism_status checks, per mechanism, that each named source module is present and that every string in that module's required_anchors appears in the copied file text. A missing module or anchor marks the mechanism blocked.
_synthetic_runtime_exercises reads the copied module texts and derives three bundles: a proof bundle, an authority bundle, and a runtime bundle. Each bundle is a set of anchor-presence facts computed from the copied source, not a live run.
_erdos_static_scan runs a regular expression over the copied CertificateKernel.lean and rejects the tokens sorry, admit, and axiom. It counts declarations and samples a few names. A banned token blocks the scan. This is a token scan, not a Lean proof check.
_lean_lake_probe optionally runs lake env lean over the copied kernel under the source Lake project. A zero-exit run records live_compile_pass; a missing lean/lake binary or an unavailable Mathlib dependency records unavailable; a nonzero exit records blocked. Every branch fixes proof_authority_delta to none, so a compile pass is an availability signal, not formal-result correctness.
_semantic_negative_case_proofs recomputes the twelve negative cases and reports how many were rejected.
\n
Negative verdicts re-derived from source
\n
_computed_negative_case_verdict ignores the fixture's declared verdict and re-derives each one from source evidence, an overclaim-shape check, and a runtime probe, rejecting only when the source supports it. The negative-case step is the part worth reading closely. _computed_negative_case_verdict does not trust the fixture to declare its own verdict. It re-derives each verdict from source-observed evidence, an overclaim-shape check on the mechanism's declared ceiling, and a runtime probe, then marks the row declared_fixture_ignored. A case counts as rejected only when the source itself supports the rejection.
\n
Pass gate and the omission record
\n
_evaluate passes only with no findings and a validated manifest, while result_card attaches AUTHORITY_CEILING and an omission record naming what is left out: source bodies, private runtime state, and model-output data._evaluate returns pass only when there are no findings and the source manifest validated. result_card projects the outcome into a compact card and attaches AUTHORITY_CEILING plus an omission record naming what is left out of the result: full copied source bodies, raw private runtime state, and live provider or session payloads.
\n
Function
Role
run, run_batch4_bundle, main
Entry shapes over a fixture directory or the exported bundle
_load_manifest
Parse batch4_probe_manifest.json or reject it as a finding
_module_rows
Build the copied-module map from source_module_manifest.json
_mechanism_status
Check per-mechanism module presence and required anchors
_synthetic_runtime_exercises
Derive proof, authority, and runtime bundles from copied text
_erdos_static_scan
Reject sorry, admit, axiom in the copied Lean kernel
_lean_lake_probe
Optional local lake env lean availability signal
_evaluate
Orchestrate the gate and collect findings
_semantic_negative_case_proofs
Recompute each negative case from source
result_card
Attach AUTHORITY_CEILING and the omission record
\nDiagram of the mechanism (8 steps).Diagram source & refs
The fixture's twelve negative cases, one per mechanism with a stable error code, are the public failure floor stopping the source import from reading as an unbounded proof or runtime claim. The fixture declares twelve negative cases, one per mechanism, each with a stable error code held in EXPECTED_NEGATIVE_CASES. They are the public failure floor that stops a source-import surface from reading as an unbounded proof or runtime claim:
A case is marked rejected only when the source-observed verdict supports it, never because the fixture said so. Each case is marked rejected only when the source-observed verdict supports it, not because the fixture said so.
\n
Prior Art Grounding
\n
Evidence before authority, from proof-carrying code
\n
It separates the authority to act from the evidence permitting it, borrowing the evidence-before-authority ordering of proof-carrying code and capability security, yielding fixture-bound evidence rather than a verified authorization system. The component keeps the authority to act separate from the evidence that an action is permitted. That ordering is the idea behind proof-carrying code (Necula, 1997) and capability-based security, where a request arrives with evidence of its own legitimacy rather than relying on ambient trust. This module borrows the evidence-before-authority ordering over copied fixtures. The result is fixture-bound evidence, not a verified authorization system.
The focused test covers fixture, bundle, digest, anchor, mutated-Lean-body rejection, and negative-case recomputation, while the corpus check confirms the bundle, Mermaid, Atlas, and Markdown projections stay consistent. The focused test exercises the fixture and bundle expectations, the source digest and anchor checks, the mutated-Lean-body rejection, and the negative-case recomputation. A pass is reader-verifiable evidence only.
\n
Scope boundary
\n
Scope limit
\n
What the module may claim
\n
The module may claim only fixture-bound import, digest and anchor checks over the seventeen copied modules, a static Lean placeholder scan, an optional compile-availability signal, and metadata-only records with twelve recomputed negative cases. This module may claim fixture-bound import of copied public source, digest and required-anchor checks over the seventeen copied modules, a static placeholder-token scan over the copied Lean kernel, an optional local compile availability signal, and metadata-only result records with twelve recomputed negative cases. The proof boundary is exactly that copied-source checking: AUTHORITY_CEILING fixes proof_authority_delta to none and holds every action flag, including provider_dispatch, source_mutation_authorized, publication_authorized, and release_authorized, at false.
\n
What stays refused
\n
It claims no formal-result correctness, no Erdos #257 solution, no benchmark claims, no live enforcement or orchestration, and no mutation or launch, and the ceiling can rise only from changed evidence, never this Markdown. It does not claim formal-result correctness, a solution of Erdos #257, official benchmark claims, live sandbox enforcement, live Codex orchestration, external model access, source-file changes, public sharing or launch-scope decision, whole-system equivalence, or whole-system correctness. The static scan is bounded evidence checking, and digest equality does not establish semantic equivalence to any private original.
\n",
"components": [
{
"anchor": "component-batch4_proof_authority_runtime",
"component_id": "batch4_proof_authority_runtime",
"public_label": "Proof / Control / Runtime Import Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 5,
"principle": 4
},
"link_count": 16,
"links": [
{
"id": "batch4_proof_authority_runtime",
"kind": "organ",
"kind_label": "component",
"label": "Proof / Control / Runtime Import Bundle",
"relation": "explains",
"url": "component-batch4-proof-authority-runtime.html"
},
{
"id": "mechanism.batch4_proof_authority_runtime.validates_public_proof_authority_runtime_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public proof authority runtime bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch4-proof-authority-runtime-validates-public-proof-authority-runtime-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.formal_math_readiness_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Readiness Gate",
"relation": "depends on",
"url": "paper-module-formal-math-readiness-gate.html"
},
{
"id": "paper_module.proof_derived_governed_mutation_authorization",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof-Derived Governed Mutation Authorization",
"relation": "depends on",
"url": "paper-module-proof-derived-governed-mutation-authorization.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.batch4_proof_authority_runtime",
"reader_packet": {
"abstract": "Set 4 Proof, Authority, and Runtime Bundle imports proof-search, reasoning-authority, completion, Codex runtime, bitemporal, taskpolicy, and context-yield source bodies as public source-open evidence without claiming proof success, benchmark claims, live runtime control, or launch-scope decision.",
"authority_ceiling": "Fixture-bound public source-body import, static proof-placeholder checks, dry-run authority-boundary evidence, source-anchor evidence, negative-case evidence, and metadata-only result records only; no theorem success, Erdos #257 solution, official benchmark claims, live sandbox enforcement, live Codex orchestration, external model access, source-file changes, publishing-scope decision, launch-scope decision, whole-system equivalence, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch4_proof_authority_runtime.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch4-proof-authority-runtime-diagram-1",
"html_url": "docs/paper-module-batch4-proof-authority-runtime.html#paper-module-batch4-proof-authority-runtime-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD probe[\"batch4_probe_manifest.json 12 mechanisms + 12 negative cases\"] manifest[\"source_module_manifest.json 17 copied source modules\"] evaluate[\"_evaluate run / validate-bundle\"] anchors[\"_mechanism_status module present + required anchors\"] scan[\"_erdos_static_scan reject sorry / admit / axiom\"] probe2[\"_lean_lake_probe compile availability, delta = none\"] negatives[\"_semantic_negative_case_proofs verdict recomputed from source\"] record[\"metadata-only result record refs, digests, anchor counts, ceiling\"] probe --> evaluate manifest --> evaluate evaluate --> anchors evaluate --> scan scan --> probe2 evaluate --> negatives anchors --> record scan --> record negatives --> record",
"source_format": "mermaid",
"source_hash": "sha256:8e70a8e7a990e9291241216f622a772a9c86e7bae94585d47db687c48f4d3856",
"source_line_end": 83,
"source_line_start": 63
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 5,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.formal_math_readiness_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Formal Math Readiness Gate",
"relation": "depends on",
"url": "paper-module-formal-math-readiness-gate.html"
},
{
"id": "paper_module.proof_derived_governed_mutation_authorization",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Proof-Derived Governed Mutation Authorization",
"relation": "depends on",
"url": "paper-module-proof-derived-governed-mutation-authorization.html"
},
{
"id": "paper_module.mission_transaction_work_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Mission Transaction Work Spine",
"relation": "depends on",
"url": "paper-module-mission-transaction-work-spine.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch4_proof_authority_runtime",
"kind": "organ",
"kind_label": "component",
"label": "Proof / Control / Runtime Import Bundle",
"relation": "explains",
"url": "component-batch4-proof-authority-runtime.html"
},
{
"id": "mechanism.batch4_proof_authority_runtime.validates_public_proof_authority_runtime_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public proof authority runtime bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch4-proof-authority-runtime-validates-public-proof-authority-runtime-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch4_proof_authority_runtime",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 138,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch4_proof_authority_runtime.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Necula, 1997",
"section_id": "evidence-before-authority-from-proof-carrying-code",
"url": "https://en.wikipedia.org/wiki/Proof-carrying_code"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "A single component bundles seven easily overclaimed source files into one inspectable unit: proof search, an Erdos #257 Lean kernel, reasoning-authority fences, completion, bitemporal logging, task-policy wrapping, and context-yield. batch4proofauthority_runtime is one Microcosm component that gathers a group of source files that are easy to overclaim and checks them as a single inspectable unit. The files cover proof search, a copied Lean certificate kernel for Erdos #257, reasoning-authority fences, completion planning, bitemporal claim logging, macOS task-policy wrapping, and context-yield attribution.",
"level": 4,
"order": 1,
"section_id": "proof-authority-claims-are-recomputed-from-source",
"source_line_end": 6,
"source_line_start": 3,
"title": "Proof-authority claims are recomputed from source"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It never runs those systems: it reads their copied public source, checks each mechanism against required anchors, recomputes one negative case per mechanism, and writes a metadata-only result record. The component does not run any of those systems. It reads their copied public source, checks each declared mechanism against named required anchors, recomputes a negative case per mechanism from that source, and writes a result record. The record carries source references, digests, anchor counts, and bounded outcomes. It does not carry source bodies or live private state.",
"level": 4,
"order": 2,
"section_id": "two-runtime-paths-share-one-authority-boundary",
"source_line_end": 10,
"source_line_start": 7,
"title": "Two runtime paths share one authority boundary"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The only positive claim is narrow: copied source is present, its required anchors are in place, and the copied Lean kernel is free of proof placeholders, with everything stronger refused openly. Its positive claim is narrow. The copied source is present, its required anchors are in place, and the copied Lean kernel is free of proof placeholders. Everything stronger is refused in the open.",
"level": 4,
"order": 3,
"section_id": "a-passing-fixture-does-not-prove-a-theorem",
"source_line_end": 14,
"source_line_start": 11,
"title": "A passing fixture does not establish a theorem"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Each bundled mechanism looks like more than it is: proof search like solving open problems, the Lean kernel like a solution, a grant like a live sandbox, a wrapper like orchestration. These mechanisms sit next to claims a reader will want to make on their behalf. A proof-search benchmark looks like solving open problems. A copied CertificateKernel.lean for Erdos #257 looks like a solution. A reasoning grant looks like a live sandbox. A task-policy wrapper looks like live orchestration.",
"level": 4,
"order": 5,
"section_id": "each-mechanism-invites-an-overclaim",
"source_line_end": 20,
"source_line_start": 17,
"title": "Each mechanism invites an overclaim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component asks only whether each mechanism can be shown as copied, anchored public source without inheriting an authority it lacks, making the import auditable. A loose import surface would let any of those read as more than it is. The component answers one narrow question per mechanism: can this be shown to a reader as copied, anchored, public source, without inheriting an authority it does not have. It makes the import auditable and forces every stronger statement into a visible limit.",
"level": 4,
"order": 6,
"section_id": "one-narrow-question-per-mechanism",
"source_line_end": 24,
"source_line_start": 21,
"title": "One narrow question per mechanism"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "run takes a fixture directory and runbatch4bundle takes the exported bundle, but both forward to runcrownjewelorgan with evaluate, so they share one gate. There are two entry shapes. run consumes a fixture input directory. runbatch4bundle consumes the exported bundle directory. Both forward to runcrownjewelorgan with evaluate as the evaluator, so they share one gate. main exposes them as the run and validate-bundle subcommands. _evaluate drives the checks in order: loadmanifest reads batch4probemanifest.json. Invalid JSON returns BATCH4PROBEMANIFESTINVALID; a non-object returns BATCH4PROBEMANIFESTNOT_OBJECT. modulerows builds a module-id to row map from the exported sourcemodulemanifest.jso\u2026",
"level": 4,
"order": 8,
"section_id": "two-entry-shapes-over-one-gate",
"source_line_end": 41,
"source_line_start": 27,
"title": "Two entry shapes over one gate"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "computednegativecaseverdict ignores the fixture's declared verdict and re-derives each one from source evidence, an overclaim-shape check, and a runtime probe, rejecting only when the source supports it. The negative-case step is the part worth reading closely. computednegativecaseverdict does not trust the fixture to declare its own verdict. It re-derives each verdict from source-observed evidence, an overclaim-shape check on the mechanism's declared ceiling, and a runtime probe, then marks the row declaredfixture_ignored. A case counts as rejected only when the source itself supports the rejection.",
"level": 4,
"order": 9,
"section_id": "negative-verdicts-re-derived-from-source",
"source_line_end": 45,
"source_line_start": 42,
"title": "Negative verdicts re-derived from source"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "evaluate passes only with no findings and a validated manifest, while resultcard attaches AUTHORITYCEILING and an omission record naming what is left out: source bodies, private runtime state, and model-output data. evaluate returns pass only when there are no findings and the source manifest validated. resultcard projects the outcome into a compact card and attaches scope_limit plus an omission record naming what is left out of the result: full copied source bodies, raw private runtime state, and live provider or session payloads.",
"level": 4,
"order": 10,
"section_id": "pass-gate-and-the-omission-record",
"source_line_end": 84,
"source_line_start": 46,
"title": "Pass gate and the omission record"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 11,
"section_id": "negative-cases",
"source_line_end": 86,
"source_line_start": 85,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The fixture's twelve negative cases, one per mechanism with a stable error code, are the public failure floor stopping the source import from reading as an unbounded proof or runtime claim. The fixture declares twelve negative cases, one per mechanism, each with a stable error code held in EXPECTEDNEGATIVECASES. They are the public failure floor that stops a source-import surface from reading as an unbounded proof or runtime claim: weakskeletonsynthesisfailure: BATCH4WEAKSKELETONMUSTNOTSILENT_PASS foundrylowrepairquarantine: BATCH4FOUNDRYREPAIREDCOUNTBELOWTHRESHOLD verisofttruthleak: BATCH4VERISOFTTRUTHLEAKREJECTED verisoftprefixanswerleakage: BATCH4VERISOFTPREFIXANSWERLEAKAGEREJECTED erdos\u2026",
"level": 4,
"order": 12,
"section_id": "twelve-stable-error-codes-one-per-mechanism",
"source_line_end": 103,
"source_line_start": 87,
"title": "Twelve stable error codes, one per mechanism"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A case is marked rejected only when the source-observed verdict supports it, never because the fixture said so. Each case is marked rejected only when the source-observed verdict supports it, not because the fixture said so.",
"level": 4,
"order": 13,
"section_id": "rejection-requires-source-support",
"source_line_end": 108,
"source_line_start": 104,
"title": "Rejection requires source support"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "prior-art-grounding",
"source_line_end": 110,
"source_line_start": 109,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It separates the authority to act from the evidence permitting it, borrowing the evidence-before-authority ordering of proof-carrying code and capability security, yielding fixture-bound evidence rather than a verified authorization system. The component keeps the authority to act separate from the evidence that an action is permitted. That ordering is the idea behind proof-carrying code (Necula, 1997) and capability-based security, where a request arrives with evidence of its own legitimacy rather than relying on ambient trust. This module borrows the evidence-before-authority ordering over copied fixtures. The result is fixture-bound evidence, not a verified authorization system.",
"level": 4,
"order": 15,
"section_id": "evidence-before-authority-from-proof-carrying-code",
"source_line_end": 114,
"source_line_start": 111,
"title": "Evidence before authority, from proof-carrying code"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from the microcosm-substrate/ public root:",
"level": 2,
"order": 16,
"section_id": "validation-receipt-path",
"source_line_end": 123,
"source_line_start": 115,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The focused test covers fixture, bundle, digest, anchor, mutated-Lean-body rejection, and negative-case recomputation, while the corpus check confirms the bundle, Mermaid, Atlas, and Markdown projections stay consistent. The focused test exercises the fixture and bundle expectations, the source digest and anchor checks, the mutated-Lean-body rejection, and the negative-case recomputation. A pass is reader-verifiable evidence only.",
"level": 4,
"order": 17,
"section_id": "what-the-two-checks-confirm",
"source_line_end": 127,
"source_line_start": 124,
"title": "What the two checks confirm"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "claim-ceiling",
"source_line_end": 129,
"source_line_start": 128,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The module may claim only fixture-bound import, digest and anchor checks over the seventeen copied modules, a static Lean placeholder scan, an optional compile-availability signal, and metadata-only records with twelve recomputed negative cases. This module may claim fixture-bound import of copied public source, digest and required-anchor checks over the seventeen copied modules, a static placeholder-token scan over the copied Lean kernel, an optional local compile availability signal, and metadata-only result records with twelve recomputed negative cases. The proof boundary is exactly that copied-source checking: AUTHORITYCEILING fixes proofauthoritydelta to none and holds every action fla\u2026",
"level": 4,
"order": 19,
"section_id": "what-the-module-may-claim",
"source_line_end": 133,
"source_line_start": 130,
"title": "What the module may claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It claims no formal-result correctness, no Erdos #257 solution, no benchmark claims, no live enforcement or orchestration, and no mutation or launch, and the ceiling can rise only from changed evidence, never this Markdown. It does not claim formal-result correctness, a solution of Erdos #257, official benchmark claims, live sandbox enforcement, live Codex orchestration, external model access, source-file changes, public sharing or launch-scope decision, whole-system equivalence, or whole-system correctness. The static scan is bounded evidence checking, and digest equality does not establish semantic equivalence to any private original.",
"level": 4,
"order": 20,
"section_id": "what-stays-refused",
"source_line_end": 137,
"source_line_start": 134,
"title": "What stays refused"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch4_proof_authority_runtime.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 4 Proof, Authority, and Runtime Bundle imports proof-search, reasoning-authority, completion, Codex runtime, bitemporal, taskpolicy, and context-yield source bodies as public source-open evidence without claiming proof success, benchmark claims, live runtime control, or launch-scope decision."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch4_proof_authority_runtime.py",
"path": "src/microcosm_core/organs/batch4_proof_authority_runtime.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for Set 4 proof/authority/runtime source imports, static proof-placeholder checks, bounded synthetic exercises, exact-copy digest checks, negative cases, and scope limits.",
"symbols": [
"_load_manifest",
"_module_rows",
"_mechanism_status",
"_erdos_static_scan",
"_synthetic_runtime_exercises",
"_evaluate",
"run",
"run_batch4_bundle"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch4_proof_authority_runtime.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch4_proof_authority_runtime.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch4_proof_authority_runtime.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 4 Proof, Authority, and Runtime Bundle binds the accepted batch4_proof_authority_runtime component to a refreshed source-open bundle. It exercises Lean strategy-control and prover-skill witnesses, VeriSoftBench harness and calibration rows, Erdos #257 certificate-kernel static checks, Lean packet integrity, reasoning grant and plan authority fences, forward-integration policy, completion executor deferral, Codex driver and idle heartbeat diagnostics, metabolism claim logs, taskpolicy passthrough, and context-yield attribution while preserving copied source digests, bounded negative cases, and scope limits.",
"summary_line": "Set 4 Proof, Authority, and Runtime Bundle imports proof-search, reasoning-authority, completion, Codex runtime, bitemporal, taskpolicy, and context-yield source bodies as public source-open evidence without claiming proof success, benchmark claims, live runtime control, or launch-scope decision.",
"title": "Set 4 Proof, Authority, and Runtime Bundle"
},
{
"anchor": "paper-module-batch6-unsurfaced-primitives-capsule",
"authority_ceiling": "Fixture-bound public source-body import, copied-module digest/anchor evidence, synthetic source-exercise evidence, and metadata-only result records only; no live operator memory, prompt-shelf capture authority, live market data, provider/browser state, media launch, source-file changes, publishing-scope decision, launch-scope decision, whole-system equivalence, or whole-system correctness.",
"body_html": "
Unsurfaced primitives are exercised from copied code
\n
A scout found eleven small, load-bearing functions never shown publicly; this component copies them into the public clone and checks the copied bodies still behave as the prose claims. A scout found eleven small functions scattered across the wider system that were real and load-bearing but had never been shown as public evidence. This component copies those eleven into the public clone and checks that the copied bodies still behave the way the prose claims. It is the kind of code that quietly decides whether a bigger feature is correct: a finance unit-scale check, a clock that fires each market event once per session, a function that subtracts paused time from a recorded video offset.
\n
Each primitive gets a positive and negative path
\n
The check is by execution: each copied body runs on fixed synthetic input and is compared to its exact expected value, with a paired wrong input proving the guard. The check is by execution, not by description. For each primitive the component imports the copied body and runs it on a fixed public synthetic input, then compares the result against the exact value the real code should produce. Each primitive also carries one paired negative case, a deliberately wrong input that the code must reject or correct, so the result record shows both the working path and the guard.
\n
Public fixtures do not become live authority
\n
The result record holds only pass or fail and summary numbers, never copied source, transcripts, payloads, or market data, with every finding stamped body_in_receipt: false. What lands in the result record is the pass or fail of each exercise plus a few summary numbers. Copied source text, operator transcripts, model-output data, and live market data never enter it. Every finding is stamped body_in_receipt: false.
\n
Purpose
\n
Why a digest is not enough
\n
A file hash proves only that bytes match, not that the logic is right, so this component closes the gap by running each copied body and asserting its output. Copying a function into a public bundle and checking its file hash proves only that the bytes match. It says nothing about whether the logic is right. A digest-only bundle is loose in exactly the place that matters: it can carry a body that no longer computes what its name says. This component closes that gap by running the copied body and asserting its output.
\n
Eleven primitives, one uniform shape
\n
The eleven primitives span ranking, distillation, handoff scoring, window merging, market graphs, finance assurance, status judgment, concurrency, clock scheduling, provider scoping, and demo remapping, each checked in the same shape. The eleven primitives span keyphrase ranking, schema-loose distillation, operator handoff scoring, observed-turn window merging, market situation graphs, finance numeric assurance, fail-closed status judgment, clone-local concurrency, market-clock scheduling, provider-recovery scoping, and demo-take temporal remapping. Each is checked in the same shape so a reader can see the working answer and the rejected answer side by side.
\n
How it works
\n
Two entrypoints, one shared driver
\n
run and run_batch6_bundle both pass SPEC and the _evaluate and evaluate_negative_case callbacks to run_crown_jewel_organ, with the bundle entrypoint setting input_mode to read copied source.run and run_batch6_bundle both hand SPEC, the input directory, and two callbacks to run_crown_jewel_organ: _evaluate and evaluate_negative_case. The bundle entrypoint sets input_mode to the exported-bundle mode so the same evaluator reads copied source rather than the fixture manifest.
\n
Manifest load and roster checks
\n
_evaluate loads the probe manifest, keys the 26 copied modules by module_id, and confirms both EXPECTED_MODULE_IDS and the exact 11 EXPECTED_MECHANISMS are present._evaluate runs the pipeline in order. It loads batch6_probe_manifest.json through _load_manifest, which rejects a non-JSON file with BATCH6_PROBE_MANIFEST_INVALID and a non-object with BATCH6_PROBE_MANIFEST_NOT_OBJECT. It resolves the copied modules with _module_rows, which reads the source-module manifest and keys each row by module_id. It confirms every id in EXPECTED_MODULE_IDS (26 rows) is present, then confirms the probe manifest names exactly the 11 mechanisms in EXPECTED_MECHANISMS, emitting BATCH6_MECHANISM_SET_MISMATCH on any missing or extra id.
\n
Anchors gate the source exercise
\n
_mechanism_status blocks any mechanism whose copied text is missing a module or required_anchor, and only when nothing is missing does _run_source_exercise import and run the real body. For each mechanism _mechanism_status reads the copied target text and checks that the module's required_anchors all appear in it. A missing module or a missing anchor sets that mechanism to blocked. Then, only if no expected module is missing, _run_source_exercise imports the real copied body and runs it.
\n
Isolated import of copied bodies
\n
_load_copied_module loads each file under a private name inside _temporary_sys_path and _temporary_import_namespaces, so copied bodies run from the bundle without colliding with the running package._load_copied_module does the import. It resolves the target path, then loads the file under a private module name inside _temporary_sys_path and _temporary_import_namespaces, so the copied idea_microcosm, system, and tools bodies load from the bundle without colliding with the running package. The exercise calls the real function and compares the answer.
\n
Four exercises make the idea concrete.
\n
demo_take_temporal_join removes paused seconds
\n
demo_take_temporal_join feeds video_t_seconds a 120-second offset paused from 10s to 25s and asserts exactly 105.0, with an open-pause branch returning 15.0.demo_take_temporal_join calls video_t_seconds. It feeds a 120-second wall offset with one pause at 10s and resume at 25s and asserts the result is exactly 105.0, the 15 paused seconds removed. A second call with a pause that has not yet resumed checks the open-pause branch returns 15.0.
\n
finance_numeric_assurance fails closed on bad numbers
\n
build_finance_numeric_assurance recomputes declared numbers, so a usd_millions row with mismatched flow and an out-of-range probability turns display_state to blocked and fails closed.finance_numeric_assurance calls build_finance_numeric_assurance, which recomputes declared numbers instead of trusting them. The exercise hands it a flow row tagged usd_millions whose flow and flow_usd disagree by orders of magnitude, plus a probability declared as 70.2. The result record's display_state becomes blocked and its check_ids include stockgrid_flow_unit_scale_mismatch and probability_bounds. A mislabelled unit or an out-of-range probability fails closed.
\n
operator_handoff_linkage scores capture against input
\n
operator_handoff_linkage scores a TypeBCapture against a TypeAUserInput by containment and overlap, putting a related pair at or above 0.8 and an unrelated one below 0.3.operator_handoff_linkage calls score_pair on a TypeBCapture (an agent's earlier suggestion) against a TypeAUserInput (what the operator later typed), using containment, token overlap, and anchor matching. A related pair scores at or above the 0.8 floor with containment true; an unrelated "summarize the weather forecast" input scores below 0.3.
metabolism_market_clock sets due_fire_points to 15:31 UTC and fires the hourly points while suppressing the already-fired open event, enforcing once-per-session idempotence.metabolism_market_clock calls due_fire_points. It sets the clock to 15:31 UTC with the open event already fired earlier that day, then asserts the hourly points fire while the already-fired open event stays suppressed. The guard is idempotence: an event that fired once must not fire again in the same session.
\n
The other seven checks in brief
\n
The remaining seven follow the same shape, from rake_ranked_phrases and distill_diagnostic through market-graph, status-judgment, and _paths_overlap checks, each pairing a working answer with a rejected one. The other seven follow the same pattern. rake_ranked_phrases returns ranked phrases for real text and an empty list for stopword-only input. distill_diagnostic keeps assistant text and operator tail as separate roles with body_persisted false. merge_observed_turn_window appends a new observed tail and leaves a re-sent duplicate window unchanged. build_market_situation_graph with strict validate_market_situation_graph requires counterevidence_edges and a not-investment-advice guard on every situation. judge_status_request allows a transition with evidence, downgrades one missing it, blocks a forbidden upgrade, and blocks under a malformed policy. _paths_overlap reports a parent directory and a child path as overlapping and two siblings as not.
\nDiagram of the mechanism (7 steps).Diagram source & refs
EXPECTED_NEGATIVE_CASES holds one case per mechanism, each routed through _observed_negative_case and passing only when the paired exercise passed and its guard fired.EXPECTED_NEGATIVE_CASES holds one case per mechanism, and evaluate_negative_case routes each through _observed_negative_case, which reads the matching runtime exercise output. A case only passes when the paired exercise passed and its guard fired. The eleven are: stopword_only_keyphrase_empty, schema_loose_voice_conflation, handoff_unrelated_below_floor, observed_turn_duplicate_rerender, market_situation_missing_counterevidence, finance_unit_scale_mismatch, status_policy_poisoned, concurrency_parent_child_conflict, market_clock_duplicate_fire_suppressed, provider_timeout_scope_narrow, and demo_take_pause_remap. Each carries a stable expected error code, for example BATCH6_FINANCE_UNIT_SCALE_MISMATCH_BLOCKED for the finance case.
\n
Prior Art Grounding
\n
Established provenance and policy patterns
\n
The primitives draw on established patterns, borrowing provenance and policy-engine design for the market and status pieces and event-time-versus-record-time separation for the demo-take remap. The copied primitives draw on established patterns rather than novel theory. The market and status pieces borrow from provenance and policy-engine design; the demo-take remap follows the practice of keeping event time separate from record time.
\n
W3C PROV, for reconciling derived artifacts back to entities, activities, and responsible agents.
Open Policy Agent, which separates policy evaluation from application code through a general-purpose engine.
Martin Fowler's bitemporal history, as a prior pattern for preserving event time separately from record time.
\n
Bounded to source-open public fixtures
\n
The component keeps to source-open public fixtures and exposes no live operator memory, market data, provider state, or publishing-scope decision. The component keeps to source-open public fixtures and does not expose live operator memory, live market data, provider state, or publishing-scope decision.
\n
Validation Result record Path
\n
Run from the microcosm-substrate/ public root:
\n
What a pass proves
\n
A pass means all 26 modules carried their anchors, all 11 exercises returned expected values, all 11 negative cases fired their guard, and no record leaked a body or private state. A pass means all 26 copied modules carried their required anchors, all 11 source exercises returned the expected values, all 11 negative cases fired their guard, and no result record leaked a copied body or private state. The corpus and projection checks prove only that the generated instance for this module is fresh.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The strongest claim is that eleven copied bodies were imported, anchor-checked, exercised to their exact expected outputs, and paired with negative cases, all under metadata-only records. The strongest claim here is narrow. Eleven copied bodies were imported into the public clone, checked against their required source anchors, exercised on public synthetic input to their exact expected outputs, and paired with negative cases, all under metadata-only result records. That is the proof boundary and the scope limit for this page.
\n
What it refuses past that
\n
It is not operator memory, capture authority, trading decisions, provider recovery, media launch, source-file changes, or launch-scope decision, and AUTHORITY_CEILING holds every action flag at false. It refuses everything past that. It is not live operator memory, not capture authority, not trading decisions, not live provider recovery, not demo media launch, not source-file changes, not public sharing or launch-scope decision, not equivalence to the private root, and not whole-system correctness. AUTHORITY_CEILING in the source sets launch_authorized, provider_dispatch, runtime_execution, source_mutation_authorized, publication_authorized, and release_authorized all to false.
\n",
"components": [
{
"anchor": "component-batch6_unsurfaced_primitives_capsule",
"component_id": "batch6_unsurfaced_primitives_capsule",
"public_label": "Unsurfaced Source Primitives Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 4,
"principle": 4
},
"link_count": 15,
"links": [
{
"id": "batch6_unsurfaced_primitives_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Unsurfaced Source Primitives Bundle",
"relation": "explains",
"url": "component-batch6-unsurfaced-primitives-capsule.html"
},
{
"id": "mechanism.batch6_unsurfaced_primitives_capsule.validates_public_unsurfaced_primitives_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public unsurfaced primitives bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch6-unsurfaced-primitives-capsule-validates-public-unsurfaced-primitives-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.batch12_market_dashboard_read_model_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 12 Market Dashboard Read-Model Bundle",
"relation": "depends on",
"url": "paper-module-batch12-market-dashboard-read-model-capsule.html"
},
{
"id": "paper_module.batch7_demo_take_console_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Demo Take Console Bundle",
"relation": "depends on",
"url": "paper-module-batch7-demo-take-console-capsule.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.batch6_unsurfaced_primitives_capsule",
"reader_packet": {
"abstract": "Set 6 Unsurfaced Primitives Bundle imports provenance, operator-handoff, market, finance, provider-recovery, and demo-take source primitives as public source-open evidence without granting live operator memory, market, provider, media, public sharing, or launch-scope decision.",
"authority_ceiling": "Fixture-bound public source-body import, copied-module digest/anchor evidence, synthetic source-exercise evidence, and metadata-only result records only; no live operator memory, prompt-shelf capture authority, live market data, provider/browser state, media launch, source-file changes, publishing-scope decision, launch-scope decision, whole-system equivalence, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch6_unsurfaced_primitives_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch6-unsurfaced-primitives-capsule-diagram-1",
"html_url": "docs/paper-module-batch6-unsurfaced-primitives-capsule.html#paper-module-batch6-unsurfaced-primitives-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Probe[\"batch6_probe_manifest.json\"] Manifest[\"source module manifest 26 copied bodies\"] Load[\"_load_manifest / _module_rows\"] Anchor[\"_mechanism_status module + anchor present?\"] Exercise[\"_run_source_exercise import copied body, run it\"] Neg[\"evaluate_negative_case one wrong input per mechanism\"] Record[\"result record pass/fail + summary numbers body_in_receipt false\"] Probe --> Load Manifest --> Load Load --> Anchor Anchor --> Exercise Exercise --> Neg Neg --> Record",
"source_format": "mermaid",
"source_hash": "sha256:ec4c2cdce4878a44e23ecbcf8ca7fa9c5372bb2d2764ccb79ce6aa6bb9b728fd",
"source_line_end": 80,
"source_line_start": 65
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 4,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
},
{
"id": "paper_module.agent_route_observability_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Agent Route Observability Runtime",
"relation": "depends on",
"url": "paper-module-agent-route-observability-runtime.html"
},
{
"id": "paper_module.batch12_market_dashboard_read_model_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 12 Market Dashboard Read-Model Bundle",
"relation": "depends on",
"url": "paper-module-batch12-market-dashboard-read-model-capsule.html"
},
{
"id": "paper_module.batch7_demo_take_console_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 7 Demo Take Console Bundle",
"relation": "depends on",
"url": "paper-module-batch7-demo-take-console-capsule.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch6_unsurfaced_primitives_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Unsurfaced Source Primitives Bundle",
"relation": "explains",
"url": "component-batch6-unsurfaced-primitives-capsule.html"
},
{
"id": "mechanism.batch6_unsurfaced_primitives_capsule.validates_public_unsurfaced_primitives_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public unsurfaced primitives bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch6-unsurfaced-primitives-capsule-validates-public-unsurfaced-primitives-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch6_unsurfaced_primitives_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 137,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch6_unsurfaced_primitives_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "PROV",
"section_id": "established-provenance-and-policy-patterns",
"url": "https://www.w3.org/TR/prov-overview/"
},
{
"kind": "external_reference",
"label": "Open Policy Agent",
"section_id": "established-provenance-and-policy-patterns",
"url": "https://www.openpolicyagent.org/docs/latest"
},
{
"kind": "external_reference",
"label": "bitemporal history",
"section_id": "established-provenance-and-policy-patterns",
"url": "https://martinfowler.com/articles/bitemporal-history.html"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "A scout found eleven small, load-bearing functions never shown publicly; this component copies them into the public clone and checks the copied bodies still behave as the prose claims. A scout found eleven small functions scattered across the wider system that were real and load-bearing but had never been shown as public evidence. This component copies those eleven into the public clone and checks that the copied bodies still behave the way the prose claims. It is the kind of code that quietly decides whether a bigger feature is correct: a finance unit-scale check, a clock that fires each market event once per session, a function that subtracts paused time from a recorded video offset.",
"level": 4,
"order": 1,
"section_id": "unsurfaced-primitives-are-exercised-from-copied-code",
"source_line_end": 6,
"source_line_start": 3,
"title": "Unsurfaced primitives are exercised from copied code"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The check is by execution: each copied body runs on fixed synthetic input and is compared to its exact expected value, with a paired wrong input proving the guard. The check is by execution, not by description. For each primitive the component imports the copied body and runs it on a fixed public synthetic input, then compares the result against the exact value the real code should produce. Each primitive also carries one paired negative case, a deliberately wrong input that the code must reject or correct, so the result record shows both the working path and the guard.",
"level": 4,
"order": 2,
"section_id": "each-primitive-gets-a-positive-and-negative-path",
"source_line_end": 10,
"source_line_start": 7,
"title": "Each primitive gets a positive and negative path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The result record holds only pass or fail and summary numbers, never copied source, transcripts, payloads, or market data, with every finding stamped bodyinreceipt: false. What lands in the result record is the pass or fail of each exercise plus a few summary numbers. Copied source text, operator transcripts, model-output data, and live market data never enter it. Every finding is stamped bodyinreceipt: false.",
"level": 4,
"order": 3,
"section_id": "public-fixtures-do-not-become-live-authority",
"source_line_end": 14,
"source_line_start": 11,
"title": "Public fixtures do not become live authority"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A file hash proves only that bytes match, not that the logic is right, so this component closes the gap by running each copied body and asserting its output. Copying a function into a public bundle and checking its file hash proves only that the bytes match. It says nothing about whether the logic is right. A digest-only bundle is loose in exactly the place that matters: it can carry a body that no longer computes what its name says. This component closes that gap by running the copied body and asserting its output.",
"level": 4,
"order": 5,
"section_id": "why-a-digest-is-not-enough",
"source_line_end": 20,
"source_line_start": 17,
"title": "Why a digest is not enough"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The eleven primitives span ranking, distillation, handoff scoring, window merging, market graphs, finance assurance, status judgment, concurrency, clock scheduling, provider scoping, and demo remapping, each checked in the same shape. The eleven primitives span keyphrase ranking, schema-loose distillation, operator handoff scoring, observed-turn window merging, market situation graphs, finance numeric assurance, fail-closed status judgment, clone-local concurrency, market-clock scheduling, provider-recovery scoping, and demo-take temporal remapping. Each is checked in the same shape so a reader can see the working answer and the rejected answer side by side.",
"level": 4,
"order": 6,
"section_id": "eleven-primitives-one-uniform-shape",
"source_line_end": 24,
"source_line_start": 21,
"title": "Eleven primitives, one uniform shape"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run and runbatch6bundle both pass SPEC and the evaluate and evaluatenegativecase callbacks to runcrownjewelorgan, with the bundle entrypoint setting inputmode to read copied source. run and runbatch6bundle both hand SPEC, the input directory, and two callbacks to runcrownjewelorgan: evaluate and evaluatenegativecase. The bundle entrypoint sets inputmode to the exported-bundle mode so the same evaluator reads copied source rather than the fixture manifest.",
"level": 4,
"order": 8,
"section_id": "two-entrypoints-one-shared-driver",
"source_line_end": 30,
"source_line_start": 27,
"title": "Two entrypoints, one shared driver"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluate loads the probe manifest, keys the 26 copied modules by moduleid, and confirms both EXPECTEDMODULEIDS and the exact 11 EXPECTEDMECHANISMS are present. evaluate runs the pipeline in order. It loads batch6probemanifest.json through loadmanifest, which rejects a non-JSON file with BATCH6PROBEMANIFESTINVALID and a non-object with BATCH6PROBEMANIFESTNOTOBJECT. It resolves the copied modules with modulerows, which reads the source-module manifest and keys each row by moduleid. It confirms every id in EXPECTEDMODULEIDS (26 rows) is present, then confirms the probe manifest names exactly the 11 mechanisms in EXPECTEDMECHANISMS, emitting BATCH6MECHANISMSET_MISMATCH on any missing or extra i\u2026",
"level": 4,
"order": 9,
"section_id": "manifest-load-and-roster-checks",
"source_line_end": 34,
"source_line_start": 31,
"title": "Manifest load and roster checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "mechanismstatus blocks any mechanism whose copied text is missing a module or requiredanchor, and only when nothing is missing does runsourceexercise import and run the real body. For each mechanism mechanismstatus reads the copied target text and checks that the module's requiredanchors all appear in it. A missing module or a missing anchor sets that mechanism to blocked. Then, only if no expected module is missing, runsource_exercise imports the real copied body and runs it.",
"level": 4,
"order": 10,
"section_id": "anchors-gate-the-source-exercise",
"source_line_end": 38,
"source_line_start": 35,
"title": "Anchors gate the source exercise"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "loadcopiedmodule loads each file under a private name inside temporarysyspath and temporaryimportnamespaces, so copied bodies run from the bundle without colliding with the running package. loadcopiedmodule does the import. It resolves the target path, then loads the file under a private module name inside temporarysyspath and temporaryimportnamespaces, so the copied ideamicrocosm, system, and tools bodies load from the bundle without colliding with the running package. The exercise calls the real function and compares the answer. Four exercises make the idea concrete.",
"level": 4,
"order": 11,
"section_id": "isolated-import-of-copied-bodies",
"source_line_end": 44,
"source_line_start": 39,
"title": "Isolated import of copied bodies"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "demotaketemporaljoin feeds videotseconds a 120-second offset paused from 10s to 25s and asserts exactly 105.0, with an open-pause branch returning 15.0. demotaketemporaljoin calls videotseconds. It feeds a 120-second wall offset with one pause at 10s and resume at 25s and asserts the result is exactly 105.0, the 15 paused seconds removed. A second call with a pause that has not yet resumed checks the open-pause branch returns 15.0.",
"level": 4,
"order": 12,
"section_id": "demo-take-temporal-join-removes-paused-seconds",
"source_line_end": 48,
"source_line_start": 45,
"title": "`demo_take_temporal_join` removes paused seconds"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildfinancenumericassurance recomputes declared numbers, so a usdmillions row with mismatched flow and an out-of-range probability turns displaystate to blocked and fails closed. financenumericassurance calls buildfinancenumericassurance, which recomputes declared numbers instead of trusting them. The exercise hands it a flow row tagged usdmillions whose flow and flowusd disagree by orders of magnitude, plus a probability declared as 70.2. The result record's displaystate becomes blocked and its checkids include stockgridflowunitscalemismatch and probability_bounds. A mislabelled unit or an out-of-range probability fails closed.",
"level": 4,
"order": 13,
"section_id": "finance-numeric-assurance-fails-closed-on-bad-numbers",
"source_line_end": 52,
"source_line_start": 49,
"title": "`finance_numeric_assurance` fails closed on bad numbers"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "operatorhandofflinkage scores a TypeBCapture against a TypeAUserInput by containment and overlap, putting a related pair at or above 0.8 and an unrelated one below 0.3. operatorhandofflinkage calls score_pair on a TypeBCapture (an agent's earlier suggestion) against a TypeAUserInput (what the operator later typed), using containment, token overlap, and anchor matching. A related pair scores at or above the 0.8 floor with containment true; an unrelated \"summarize the weather forecast\" input scores below 0.3.",
"level": 4,
"order": 14,
"section_id": "operator-handoff-linkage-scores-capture-against-input",
"source_line_end": 56,
"source_line_start": 53,
"title": "`operator_handoff_linkage` scores capture against input"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "metabolismmarketclock sets duefirepoints to 15:31 UTC and fires the hourly points while suppressing the already-fired open event, enforcing once-per-session idempotence. metabolismmarketclock calls duefirepoints. It sets the clock to 15:31 UTC with the open event already fired earlier that day, then asserts the hourly points fire while the already-fired open event stays suppressed. The guard is idempotence: an event that fired once must not fire again in the same session.",
"level": 4,
"order": 15,
"section_id": "metabolism-market-clock-enforces-per-session-idempotence",
"source_line_end": 60,
"source_line_start": 57,
"title": "`metabolism_market_clock` enforces per-session idempotence"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "The remaining seven follow the same shape, from rakerankedphrases and distilldiagnostic through market-graph, status-judgment, and pathsoverlap checks, each pairing a working answer with a rejected one. The other seven follow the same pattern. rakerankedphrases returns ranked phrases for real text and an empty list for stopword-only input. distilldiagnostic keeps assistant text and operator tail as separate roles with bodypersisted false. mergeobservedturnwindow appends a new observed tail and leaves a re-sent duplicate window unchanged. buildmarketsituationgraph with strict validatemarketsituationgraph requires counterevidenceedges and a not-investment-advice guard on every situation. judg\u2026",
"level": 4,
"order": 16,
"section_id": "the-other-seven-checks-in-brief",
"source_line_end": 81,
"source_line_start": 61,
"title": "The other seven checks in brief"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "negative-cases",
"source_line_end": 83,
"source_line_start": 82,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVECASES holds one case per mechanism, each routed through observednegativecase and passing only when the paired exercise passed and its guard fired. EXPECTEDNEGATIVECASES holds one case per mechanism, and evaluatenegativecase routes each through observednegativecase, which reads the matching runtime exercise output. A case only passes when the paired exercise passed and its guard fired. The eleven are: stopwordonlykeyphraseempty, schemaloosevoiceconflation, handoffunrelatedbelowfloor, observedturnduplicatererender, marketsituationmissingcounterevidence, financeunitscalemismatch, statuspolicypoisoned, concurrencyparentchildconflict, marketclockduplicatefiresuppressed, providert\u2026",
"level": 4,
"order": 18,
"section_id": "one-guarded-case-per-mechanism",
"source_line_end": 87,
"source_line_start": 84,
"title": "One guarded case per mechanism"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "prior-art-grounding",
"source_line_end": 89,
"source_line_start": 88,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The primitives draw on established patterns, borrowing provenance and policy-engine design for the market and status pieces and event-time-versus-record-time separation for the demo-take remap. The copied primitives draw on established patterns rather than novel theory. The market and status pieces borrow from provenance and policy-engine design; the demo-take remap follows the practice of keeping event time separate from record time. W3C PROV, for reconciling derived artifacts back to entities, activities, and responsible agents. Open Policy Agent, which separates policy evaluation from application code through a general-purpose engine. Martin Fowler's bitemporal history, as a prior patter\u2026",
"level": 4,
"order": 20,
"section_id": "established-provenance-and-policy-patterns",
"source_line_end": 97,
"source_line_start": 90,
"title": "Established provenance and policy patterns"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The component keeps to source-open public fixtures and exposes no live operator memory, market data, provider state, or public sharing decision. The component keeps to source-open public fixtures and does not expose live operator memory, live market data, provider state, or public sharing decision.",
"level": 4,
"order": 21,
"section_id": "bounded-to-source-open-public-fixtures",
"source_line_end": 102,
"source_line_start": 98,
"title": "Bounded to source-open public fixtures"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from the microcosm-substrate/ public root:",
"level": 2,
"order": 22,
"section_id": "validation-receipt-path",
"source_line_end": 122,
"source_line_start": 103,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means all 26 modules carried their anchors, all 11 exercises returned expected values, all 11 negative cases fired their guard, and no record leaked a body or private state. A pass means all 26 copied modules carried their required anchors, all 11 source exercises returned the expected values, all 11 negative cases fired their guard, and no result record leaked a copied body or private state. The corpus and projection checks prove only that the generated instance for this module is fresh.",
"level": 4,
"order": 23,
"section_id": "what-a-pass-proves",
"source_line_end": 126,
"source_line_start": 123,
"title": "What a pass proves"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 24,
"section_id": "claim-ceiling",
"source_line_end": 128,
"source_line_start": 127,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim is that eleven copied bodies were imported, anchor-checked, exercised to their exact expected outputs, and paired with negative cases, all under metadata-only records. The strongest claim here is narrow. Eleven copied bodies were imported into the public clone, checked against their required source anchors, exercised on public synthetic input to their exact expected outputs, and paired with negative cases, all under metadata-only result records. That is the proof boundary and the scope limit for this page.",
"level": 4,
"order": 25,
"section_id": "the-strongest-supported-claim",
"source_line_end": 132,
"source_line_start": 129,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It is not operator memory, capture authority, trading decisions, provider recovery, media launch, source-file changes, or launch-scope decision, and AUTHORITYCEILING holds every action flag at false. It refuses everything past that. It is not live operator memory, not capture authority, not trading decisions, not live provider recovery, not demo media launch, not source-file changes, not public sharing or launch-scope decision, not equivalence to the private root, and not whole-system correctness. AUTHORITYCEILING in the source sets launchauthorized, providerdispatch, runtimeexecution, sourcemutationauthorized, publicationauthorized, and release_authorized all to false.",
"level": 4,
"order": 26,
"section_id": "what-it-refuses-past-that",
"source_line_end": 136,
"source_line_start": 133,
"title": "What it refuses past that"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch6_unsurfaced_primitives_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 6 Unsurfaced Primitives Bundle imports provenance, operator-handoff, market, finance, provider-recovery, and demo-take source primitives as public source-open evidence without granting live operator memory, market, provider, media, public sharing, or launch-scope decision."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch6_unsurfaced_primitives_capsule.py",
"path": "src/microcosm_core/organs/batch6_unsurfaced_primitives_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for Set 6 source-open primitive imports, copied-module digest and anchor checks, public synthetic source exercises, negative cases, and scope limits.",
"symbols": [
"_load_manifest",
"_module_rows",
"_load_copied_module",
"_mechanism_status",
"_run_source_exercise",
"_evaluate",
"run",
"run_batch6_bundle"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch6_unsurfaced_primitives_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch6_unsurfaced_primitives_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch6_unsurfaced_primitives_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 6 Unsurfaced Primitives Bundle binds the accepted batch6_unsurfaced_primitives_capsule component to source-open primitive exercises. It checks source note keyphrase scoring, schema-loose distillation, operator handoff linkage, observed-turn window merging, market situation graphs, finance numeric assurance, fail-closed policy judgment, clone-local concurrency, market-clock scheduling, provider-recovery scoping, and demo-take temporal remapping while preserving public fixture inputs, exact-source digest expectations, negative cases, and scope limits.",
"summary_line": "Set 6 Unsurfaced Primitives Bundle imports provenance, operator-handoff, market, finance, provider-recovery, and demo-take source primitives as public source-open evidence without granting live operator memory, market, provider, media, public sharing, or launch-scope decision.",
"title": "Set 6 Unsurfaced Primitives Bundle"
},
{
"anchor": "paper-module-engine-room-public-projection-leak-gate",
"authority_ceiling": "Public projection leak-gate fixture and rendered-root scan result records only; no general security scanner, prompt-injection defense, sandbox, information-flow proof, launch-scope decision, whole-system equivalence, source-file changes, or whole-system correctness.",
"body_html": "
Before rendered tree
\n
Before a rendered tree reaches a public reader, one narrow question stands: does it carry account secret shapes, a non-public paths, a transport symbol, or an escaping symlink that should not leave the workspace? Before a rendered tree of files is exposed to a public reader, someone has to answer one narrow question: does this tree contain anything that should not leave the workspace? account secret-shaped strings, a private home path, a browser transport symbol, or a symlink pointing outside the tree are all ways for private material to ride along with an otherwise public projection.
\n
Component answers question
\n
This component answers that question over a directory of rendered files with a green or red verdict: a data-loss-prevention gate for one tree, not a general scanner. This component answers that one question over a directory of rendered files and returns a green or red verdict. It is a data-loss-prevention gate for one tree, not a general scanner.
\n
Design point is
\n
The design point is what it does with a find: each match is recorded by category, path, line, and a SHA-256 hash while the matched value is dropped, so the report never becomes a second copy of the leak. The design point is what it does with what it finds. A secret scanner that prints the secret into its own report has made a second copy of the leak. This gate never does that. Every match is recorded by category, path, line number, and a SHA-256 hash of the matched text, and the matched value itself is dropped. A reviewer can confirm that a leak was found and where, without the result record becoming the thing that leaks.
\n
Purpose
\n
The quiet co-travel failure mode
\n
The failure a loose gate allows is quiet co-travel: a rendered file carries a key or non-public paths nobody noticed, and a naive report copies that value into a second guarded artifact. The failure mode a loose gate allows is quiet co-travel. A publish step renders a directory, a human skims it, and one file carries a key or a non-public paths that nobody noticed. The problem is not that leaks are undetectable. The problem is that ad hoc review misses them and that a naive report copies the sensitive value out of the file and into a second artifact that then also has to be guarded.
\n
Fixing the position of the check
\n
This gate reads rendered files and path names against fixed regular expressions, treats an escaping symlink as a hard blocker, folds in an optional gitleaks run, and emits a hash-only publishable record. This gate fixes the position of that check. It reads the rendered files and their path names against a fixed set of regular expressions, treats a symlink that escapes the root as a hard blocker, and folds an optional gitleaks run into the same result. The verdict is deterministic for the same input, the counts are honest, and the result record is safe to publish because it holds hashes and categories, not the matched text.
\n
How it works
\n
scan_projection resolves and walks the root
\n
scan_projection resolves the supplied root, raises ValueError if it is missing or not a directory, normalises policy-exception paths, and walks the tree in sorted path order. The entry point is scan_projection(root, ...). It resolves the supplied root and raises ValueError if the root is missing or is not a directory. It then normalises the caller's policy-exception paths against DEFAULT_POLICY_EXCEPTION_PATHS through _normalise_policy_paths, and walks the tree in sorted path order.
\n
Three per-entry checks in order
\n
Each entry passes three checks in order: skip-and-count for named directories and .pyc files, record a symlink whose target escapes the root, and scan every other path two ways. For each entry the walk applies three checks in order. Directories named in SCAN_SKIP_DIR_NAMES (node_modules, .git, .vite, dist, __pycache__, .pytest_cache) and files with a .pyc suffix are skipped and counted. A symlink whose resolved target does not sit under the root is recorded as a symlink escape, keeping only the relative path and a hash of the escaped target. Every other path is scanned two ways.
\n
Path and content pattern scans
\n
_scan_path runs the relative path against PATH_PATTERNS and _scan_file runs UTF-8 contents against CONTENT_PATTERNS for account secrets, non-public paths, and host-bound transport, skipping any file it cannot read as text._scan_path runs the relative path string against PATH_PATTERNS: private history and voice file shapes, a private notes tree shape, and browser transport file names. _scan_file reads each readable file as UTF-8 and runs its contents against CONTENT_PATTERNS: private key blocks and vendor key shapes under account secrets (for example openai_key_shape, github_token_shape, aws_access_key_shape), private home and profile paths under private_path, and browser transport symbols and debug ports under host_bound_transport. A file that cannot be read as text is skipped, not failed.
\n
_hit stores a hash, never the value
\n
Every match becomes a _hit record carrying category, pattern name, path, optional line, source kind, policy-exception flag, and match_sha256, and this is the single place the leak-safety property is enforced. Every match becomes a record through _hit. That record carries the category, the pattern name, the relative path, an optional line number, the source kind (content or path), whether the path is a policy exception, and match_sha256. The matched value is never stored. This is the single place the leak-safety property is enforced.
\n
Splitting hits and computing the verdict
\n
scan_projection retains policy-exception hits as non-blocking evidence, treats every other hit as blocking, folds in an optional redacted gitleaks run, and sets the verdict red on any blocker.scan_projection then splits the hits. Any hit whose path is in the policy exception set is retained as evidence but does not block. Every other hit is a blocking hit. If the caller asked for it, run_gitleaks runs the external scanner over the root with --redact and records only its status, finding count, return code, and hashed stdout and stderr. _overall_status computes the verdict: red if there is any blocking hit or symlink escape, or if the gitleaks status is red, error, or unavailable_fail_closed; otherwise green. The result sets public_release_allowed_by_scan to true only when the status is green.
\n
Fixture-driven evaluation
\n
evaluate_case materialises one JSON case, runs scan_projection, and checks the observed status, while evaluate_fixture_dir passes only when every case meets its expectation. Two functions drive the check from fixtures. evaluate_case materialises one JSON case into a temporary root, runs scan_projection, and compares the observed status to the case's expected status. evaluate_fixture_dir runs every case in a directory and returns status: pass only when all cases meet their expectation. main exposes both a scan and an evaluate-fixtures subcommand.
\n
Function
Role
scan_projection
Walks the root, scans paths and files, computes the verdict
_scan_path / _scan_file
Apply PATH_PATTERNS and CONTENT_PATTERNS
_hit
Builds a hash-only record and drops the matched value
_overall_status
Derives green or red from hits, symlinks, and gitleaks
run_gitleaks
Optional external scan, recorded as status and hashes only
evaluate_case / evaluate_fixture_dir
Run fixture cases and check expected status
\nDiagram of the mechanism (12 steps).Diagram source & refs
flowchart TD Root["Rendered root walk in sorted order"] Root --> Skip{"Skip dir or .pyc?"} Skip -- "yes" --> Count["Counted, not scanned"] Skip -- "no" --> Link{"Symlink escapes root?"} Link -- "yes" --> Escape["Symlink escape target hash only"] Link -- "no" --> Scan["_scan_path and _scan_file _hit stores hash only"] Scan --> Policy{"Path in policy exception set?"} Policy -- "yes" --> Kept["Retained, non-blocking"] Policy -- "no" --> Block["Blocking hit"] Escape --> Verdict{"Blocking hit, escape, or gitleaks red?"} Block --> Verdict Verdict -- "yes" --> Red["red verdict"] Verdict -- "no" --> Green["green verdict"]
\n
Negative cases
\n
Two green cases, three red
\n
Five cases split two green and three red, and policy_exception_hash_only plants a boundary-document example inside a policy-exception path and still expects green with the hit retained but non-blocking. The fixture set has five cases: two that must stay green and three that must go red. clean_projection and policy_exception_hash_only are the positive cases; the second plants a boundary-document example inside a policy-exception path and expects a green verdict with the hit retained but non-blocking.
\n
Three blockers, each hash-only
\n
The three negative cases force a non-public paths, a vendor key shape, and a path-name pattern block, and the tests assert none of those values survive into the emitted record. The three negative cases each force a distinct blocker. planted_private_path places a private home path in a file body and expects red with the match stored as a hash. planted_key_shape plants a vendor key shape and expects red without the raw key appearing anywhere in the result. path_pattern_blocked carries a non-public paths shape in a file name so the block comes from _scan_path, not file contents. The focused tests confirm each of these, and they assert directly that the non-public paths, the key shape, and the symlink target do not survive into the emitted record.
\n
Prior Art Grounding
\n
Data-loss-prevention and secret-scanning lineage
\n
The component follows data-loss-prevention and secret-scanning practice, anchored by NIST DLP, GitHub secret scanning, and Gitleaks, contributing the hash-only result record and policy-exception split over one projection tree. The component follows data-loss-prevention and secret-scanning practice: scan artifacts before public sharing, detect account secret-shaped strings and non-public paths markers, keep enough evidence for triage, and never copy the sensitive payload into the report. Relevant anchors are NIST's Data Loss Prevention work, GitHub secret scanning, and Gitleaks, which this gate can call as an optional second pass. The contribution here is the hash-only result record and the policy-exception split, narrowed to one rendered projection tree.
\n
Validation Result record Path
\n
Reader-verifiable result record is
\n
The reader-verifiable result record is the focused test suite plus the corpus parity check. The reader-verifiable result record is the focused test suite plus the corpus parity check:
A pass means the five fixture cases behave as expected and the hash-only handling holds; it approves no launch and proves no whole-system safety. A pass means the five fixture cases behave as expected and the hash-only handling holds. It does not approve launch or prove whole-system safety.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
Given a directory, the gate finds account secret shapes, non-public paths shapes, and escaping symlinks with fixed pattern sets and reports them as hashes, so the result record is itself safe to publish. The strongest claim the evidence supports: given a directory, this gate finds account secret shapes, non-public paths shapes, and escaping symlinks with the fixed pattern sets, and reports them as hashes so the result record is itself safe to publish. That is the whole proof boundary.
\n
What the gate cannot promise
\n
It is no general security scanner or information-flow proof: regex detection can miss encoded or novel secrets, and a green verdict means only that this one gate found no blocker in this one scan. It refuses more. It is not a general security scanner, not prompt-injection defense, not sandboxing, and not an information-flow proof. Regex and path-shape detection can miss encoded, split, or novel secrets, and can flag benign examples that are not routed through the policy-exception list. A green verdict, and public_release_allowed_by_scan: true, means only that this one gate found no blocker in this one scan. gitleaks_status: unavailable is not an external pass. launch-scope decision, source-open authority, and whole-system equivalence stay outside this gate; that is the scope limit.
\n",
"components": [],
"doctrine_feed": {
"counts": {
"axiom": 5,
"concept": 1,
"mechanism": 1,
"paper_module": 1,
"principle": 6
},
"link_count": 14,
"links": [
{
"id": "mechanism.engine_room_public_projection_leak_gate.validates_public_projection_leak_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public projection leak gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-public-projection-leak-gate-validates-public-projection-leak-gate"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[31:paper_module.engine_room_demo].depends_on",
"url": "paper-module-engine-room-demo.html"
},
{
"id": "paper_module.engine_room_egress_self_compliance_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Egress Self-Compliance Gate",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[86:paper_module.engine_room_egress_self_compliance_gate].depends_on",
"url": "paper-module-engine-room-egress-self-compliance-gate.html"
}
],
"paper_module_id": "paper_module.engine_room_public_projection_leak_gate",
"reader_packet": {
"abstract": "The Engine Room public projection leak gate validates rendered public projection roots for account secret-shaped strings, non-public paths shapes, symlink escapes, policy-exception handling, and optional gitleaks status while keeping findings hash-only.",
"authority_ceiling": "Public projection leak-gate fixture and rendered-root scan result records only; no general security scanner, prompt-injection defense, sandbox, information-flow proof, launch-scope decision, whole-system equivalence, source-file changes, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.engine_room_public_projection_leak_gate.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-engine-room-public-projection-leak-gate-diagram-1",
"html_url": "docs/paper-module-engine-room-public-projection-leak-gate.html#paper-module-engine-room-public-projection-leak-gate-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Root[\"Rendered root walk in sorted order\"] Root --> Skip{\"Skip dir or .pyc?\"} Skip -- \"yes\" --> Count[\"Counted, not scanned\"] Skip -- \"no\" --> Link{\"Symlink escapes root?\"} Link -- \"yes\" --> Escape[\"Symlink escape target hash only\"] Link -- \"no\" --> Scan[\"_scan_path and _scan_file _hit stores hash only\"] Scan --> Policy{\"Path in policy exception set?\"} Policy -- \"yes\" --> Kept[\"Retained, non-blocking\"] Policy -- \"no\" --> Block[\"Blocking hit\"] Escape --> Verdict{\"Blocking hit, escape, or gitleaks red?\"} Block --> Verdict Verdict -- \"yes\" --> Red[\"red verdict\"] Verdict -- \"no\" --> Green[\"green verdict\"]",
"source_format": "mermaid",
"source_hash": "sha256:49d72977d7c83fd3f8b03b4dad92ae52e56830745c6921cee6af71ce8f54b311",
"source_line_end": 153,
"source_line_start": 138
}
],
"grouped_relation_feeds": [
{
"count": 5,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"relation": "depends on"
},
{
"count": 1,
"links": [
{
"id": "mechanism.engine_room_public_projection_leak_gate.validates_public_projection_leak_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public projection leak gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-public-projection-leak-gate-validates-public-projection-leak-gate"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.engine_room_public_projection_leak_gate",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 248,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_public_projection_leak_gate.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Data Loss Prevention",
"section_id": "data-loss-prevention-and-secret-scanning-lineage",
"url": "https://www.nist.gov/publications/data-loss-prevention"
},
{
"kind": "external_reference",
"label": "GitHub secret scanning",
"section_id": "data-loss-prevention-and-secret-scanning-lineage",
"url": "https://docs.github.com/en/code-security/concepts/secret-security/about-secret-scanning"
},
{
"kind": "external_reference",
"label": "Gitleaks",
"section_id": "data-loss-prevention-and-secret-scanning-lineage",
"url": "https://github.com/gitleaks/gitleaks"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Before a rendered tree reaches a public reader, one narrow question stands: does it carry account secret shapes, a non-public paths, a transport symbol, or an escaping symlink that should not leave the workspace?** Before a rendered tree of files is exposed to a public reader, someone has to answer one narrow question: does this tree contain anything that should not leave the workspace? account secret-shaped strings, a private home path, a browser transport symbol, or a symlink pointing outside the tree are all ways for private material to ride along with an otherwise public projection.",
"level": 4,
"order": 1,
"section_id": "before-rendered-tree",
"source_line_end": 13,
"source_line_start": 3,
"title": "Before rendered tree"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component answers that question over a directory of rendered files with a green or red verdict: a data-loss-prevention gate for one tree, not a general scanner.** This component answers that one question over a directory of rendered files and returns a green or red verdict. It is a data-loss-prevention gate for one tree, not a general scanner.",
"level": 4,
"order": 2,
"section_id": "component-answers-question",
"source_line_end": 21,
"source_line_start": 14,
"title": "Component answers question"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The design point is what it does with a find: each match is recorded by category, path, line, and a SHA-256 hash while the matched value is dropped, so the report never becomes a second copy of the leak.** The design point is what it does with what it finds. A secret scanner that prints the secret into its own report has made a second copy of the leak. This gate never does that. Every match is recorded by category, path, line number, and a SHA-256 hash of the matched text, and the matched value itself is dropped. A reviewer can confirm that a leak was found and where, without the result record becoming the thing that leaks.",
"level": 4,
"order": 3,
"section_id": "design-point-is",
"source_line_end": 32,
"source_line_start": 22,
"title": "Design point is"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 34,
"source_line_start": 33,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The failure a loose gate allows is quiet co-travel: a rendered file carries a key or non-public paths nobody noticed, and a naive report copies that value into a second guarded artifact.** The failure mode a loose gate allows is quiet co-travel. A publish step renders a directory, a human skims it, and one file carries a key or a non-public paths that nobody noticed. The problem is not that leaks are undetectable. The problem is that ad hoc review misses them and that a naive report copies the sensitive value out of the file and into a second artifact that then also has to be guarded.",
"level": 4,
"order": 5,
"section_id": "the-quiet-co-travel-failure-mode",
"source_line_end": 45,
"source_line_start": 35,
"title": "The quiet co-travel failure mode"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This gate reads rendered files and path names against fixed regular expressions, treats an escaping symlink as a hard blocker, folds in an optional gitleaks run, and emits a hash-only publishable record.** This gate fixes the position of that check. It reads the rendered files and their path names against a fixed set of regular expressions, treats a symlink that escapes the root as a hard blocker, and folds an optional gitleaks run into the same result. The verdict is deterministic for the same input, the counts are honest, and the result record is safe to publish because it holds hashes and categories, not the matched text.",
"level": 4,
"order": 6,
"section_id": "fixing-the-position-of-the-check",
"source_line_end": 56,
"source_line_start": 46,
"title": "Fixing the position of the check"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 58,
"source_line_start": 57,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "scanprojection resolves the supplied root, raises ValueError if it is missing or not a directory, normalises policy-exception paths, and walks the tree in sorted path order. The entry point is scanprojection(root, ...). It resolves the supplied root and raises ValueError if the root is missing or is not a directory. It then normalises the caller's policy-exception paths against DEFAULTPOLICYEXCEPTIONPATHS through normalisepolicypaths, and walks the tree in sorted path order.",
"level": 4,
"order": 8,
"section_id": "scan-projection-resolves-and-walks-the-root",
"source_line_end": 68,
"source_line_start": 59,
"title": "`scan_projection` resolves and walks the root"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Each entry passes three checks in order: skip-and-count for named directories and .pyc files, record a symlink whose target escapes the root, and scan every other path two ways.** For each entry the walk applies three checks in order. Directories named in SCANSKIPDIRNAMES (nodemodules, .git, .vite, dist, pycache, .pytest_cache) and files with a .pyc suffix are skipped and counted. A symlink whose resolved target does not sit under the root is recorded as a symlink escape, keeping only the relative path and a hash of the escaped target. Every other path is scanned two ways.",
"level": 4,
"order": 9,
"section_id": "three-per-entry-checks-in-order",
"source_line_end": 79,
"source_line_start": 69,
"title": "Three per-entry checks in order"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "scanpath runs the relative path against PATHPATTERNS and scan_file runs UTF-8 contents against CONTENT_PATTERNS for account secrets, non-public paths, and host-bound transport, skipping any file it cannot read as text. scan_path runs the relative path string against PATH_PATTERNS: private history and voice file shapes, a private notes tree shape, and browser transport file names. scanfile reads each readable file as UTF-8 and runs its contents against CONTENT_PATTERNS: private key blocks and vendor key shapes under account secrets (for example openaikeyshape, githubtokenshape, awsaccesskey_shape), private home and profile paths under private_path, and browser transport symbols and debug por\u2026",
"level": 4,
"order": 10,
"section_id": "path-and-content-pattern-scans",
"source_line_end": 93,
"source_line_start": 80,
"title": "Path and content pattern scans"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Every match becomes a hit record carrying category, pattern name, path, optional line, source kind, policy-exception flag, and match_sha256, and this is the single place the leak-safety property is enforced.** Every match becomes a record through _hit. That record carries the category, the pattern name, the relative path, an optional line number, the source kind (content or path), whether the path is a policy exception, and match_sha256. The matched value is never stored. This is the single place the leak-safety property is enforced.",
"level": 4,
"order": 11,
"section_id": "hit-stores-a-hash-never-the-value",
"source_line_end": 103,
"source_line_start": 94,
"title": "`_hit` stores a hash, never the value"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "scanprojection retains policy-exception hits as non-blocking evidence, treats every other hit as blocking, folds in an optional redacted gitleaks run, and sets the verdict red on any blocker. scanprojection then splits the hits. Any hit whose path is in the policy exception set is retained as evidence but does not block. Every other hit is a blocking hit. If the caller asked for it, run_gitleaks runs the external scanner over the root with --redact and records only its status, finding count, return code, and hashed stdout and stderr. overallstatus computes the verdict: red if there is any blocking hit or symlink escape, or if the gitleaks status is red, error, or unavailablefailclosed; othe\u2026",
"level": 4,
"order": 12,
"section_id": "splitting-hits-and-computing-the-verdict",
"source_line_end": 117,
"source_line_start": 104,
"title": "Splitting hits and computing the verdict"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "evaluatecase materialises one JSON case, runs scan_projection, and checks the observed status, while evaluatefixturedir passes only when every case meets its expectation.** Two functions drive the check from fixtures. evaluate_case materialises one JSON case into a temporary root, runs scan_projection, and compares the observed status to the case's expected status. evaluatefixturedir runs every case in a directory and returns status: pass only when all cases meet their expectation. main exposes both a scan and an evaluate-fixtures subcommand.",
"level": 4,
"order": 13,
"section_id": "fixture-driven-evaluation",
"source_line_end": 154,
"source_line_start": 118,
"title": "Fixture-driven evaluation"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 156,
"source_line_start": 155,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Five cases split two green and three red, and policyexceptionhashonly plants a boundary-document example inside a policy-exception path and still expects green with the hit retained but non-blocking.** The fixture set has five cases: two that must stay green and three that must go red. clean_projection and policyexceptionhash_only are the positive cases; the second plants a boundary-document example inside a policy-exception path and expects a green verdict with the hit retained but non-blocking.",
"level": 4,
"order": 15,
"section_id": "two-green-cases-three-red",
"source_line_end": 166,
"source_line_start": 157,
"title": "Two green cases, three red"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The three negative cases force a non-public paths, a vendor key shape, and a path-name pattern block, and the tests assert none of those values survive into the emitted record.** The three negative cases each force a distinct blocker. plantedprivatepath places a private home path in a file body and expects red with the match stored as a hash. plantedkeyshape plants a vendor key shape and expects red without the raw key appearing anywhere in the result. pathpatternblocked carries a non-public paths shape in a file name so the block comes from scanpath, not file contents. The focused tests confirm each of these, and they assert directly that the non-public paths, the key shape, and the syml\u2026",
"level": 4,
"order": 16,
"section_id": "three-blockers-each-hash-only",
"source_line_end": 179,
"source_line_start": 167,
"title": "Three blockers, each hash-only"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 181,
"source_line_start": 180,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component follows data-loss-prevention and secret-scanning practice, anchored by NIST DLP, GitHub secret scanning, and Gitleaks, contributing the hash-only result record and policy-exception split over one projection tree.** The component follows data-loss-prevention and secret-scanning practice: scan artifacts before public sharing, detect account secret-shaped strings and non-public paths markers, keep enough evidence for triage, and never copy the sensitive payload into the report. Relevant anchors are NIST's Data Loss Prevention work, GitHub secret scanning, and Gitleaks, which this gate can call as an optional second pass. The contribution here is the hash-only result record and\u2026",
"level": 4,
"order": 18,
"section_id": "data-loss-prevention-and-secret-scanning-lineage",
"source_line_end": 196,
"source_line_start": 182,
"title": "Data-loss-prevention and secret-scanning lineage"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 198,
"source_line_start": 197,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "**The reader-verifiable result record is the focused test suite plus the corpus parity check.** The reader-verifiable result record is the focused test suite plus the corpus parity check: The public exercise runs the same fixture matrix directly:",
"level": 4,
"order": 20,
"section_id": "reader-verifiable-receipt-is",
"source_line_end": 215,
"source_line_start": 199,
"title": "Reader-verifiable result record is"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the five fixture cases behave as expected and the hash-only handling holds; it approves no launch and proves no whole-system safety.** A pass means the five fixture cases behave as expected and the hash-only handling holds. It does not approve launch or prove whole-system safety.",
"level": 4,
"order": 21,
"section_id": "what-a-pass-means-here",
"source_line_end": 222,
"source_line_start": 216,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 224,
"source_line_start": 223,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Given a directory, the gate finds account secret shapes, non-public paths shapes, and escaping symlinks with fixed pattern sets and reports them as hashes, so the result record is itself safe to publish.** The strongest claim the evidence supports: given a directory, this gate finds account secret shapes, non-public paths shapes, and escaping symlinks with the fixed pattern sets, and reports them as hashes so the result record is itself safe to publish. That is the whole proof boundary.",
"level": 4,
"order": 23,
"section_id": "the-strongest-supported-claim",
"source_line_end": 234,
"source_line_start": 225,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is no general security scanner or information-flow proof: regex detection can miss encoded or novel secrets, and a green verdict means only that this one gate found no blocker in this one scan.** It refuses more. It is not a general security scanner, not prompt-injection defense, not sandboxing, and not an information-flow proof. Regex and path-shape detection can miss encoded, split, or novel secrets, and can flag benign examples that are not routed through the policy-exception list. A green verdict, and publicreleaseallowedbyscan: true, means only that this one gate found no blocker in this one scan. gitleaks_status: unavailable is not an external pass. launch-scope decision, source-\u2026",
"level": 4,
"order": 24,
"section_id": "what-the-gate-cannot-promise",
"source_line_end": 247,
"source_line_start": 235,
"title": "What the gate cannot promise"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/engine_room_public_projection_leak_gate.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The Engine Room public projection leak gate validates rendered public projection roots for account secret-shaped strings, non-public paths shapes, symlink escapes, policy-exception handling, and optional gitleaks status while keeping findings hash-only."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "public_projection_leak_gate.py",
"path": "src/microcosm_core/engine_room/public_projection_leak_gate.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for public projection leak-gate scans, fixture evaluation, hash-only evidence, optional gitleaks result records, and scope limits.",
"symbols": [
"scan_projection",
"run_gitleaks",
"evaluate_fixture_dir",
"main"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/public_projection_leak_gate.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_public_projection_leak_gate.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_public_projection_leak_gate.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Engine Room Public Projection Leak Gate is a DLP-style projection boundary. It scans rendered public projection files and paths, checks symlink escapes, records policy-exception hits as hash-only evidence, reports optional gitleaks status, and validates two positive plus three negative fixture cases without copying sensitive payloads, approving launch, or claiming general security, prompt-injection, sandbox, or information-flow authority.",
"summary_line": "The Engine Room public projection leak gate validates rendered public projection roots for account secret-shaped strings, non-public paths shapes, symlink escapes, policy-exception handling, and optional gitleaks status while keeping findings hash-only.",
"title": "Engine Room Public Projection Leak Gate"
},
{
"anchor": "paper-module-batch5-authority-systems-capsule",
"authority_ceiling": "Fixture-bound public source-body import, copied-module digest and anchor evidence, synthetic source-exercise evidence, and metadata-only result records only; no live external model access, proof success, process signal authority, generated-state mutation authority, source-file changes, publishing-scope decision, launch-scope decision, whole-system equivalence, or whole-system correctness.",
"body_html": "
Authority-system checks run from copied bodies
\n
A reader can inspect eight unrelated authority and systems mechanisms at once and confirm each refuses its wrong answer, without running any of the real machinery. This component lets a reader inspect eight unrelated authority and systems mechanisms at once, and confirm that each one refuses the wrong answer, without running any of the real machinery. The eight are: post-execution result record validation, reasoning replay scope, verifier-gated proof repair, process orphan classification, generated-state fixpoint settlement, trace-tape compaction, code blast radius, and doctrine-graph compilation.
\n
Fixture cases exercise delegation and refusal
\n
It ships an exact copy of each source body, checks the copies against digests and anchors, then runs a synthetic re-derivation per mechanism that recomputes its own verdict instead of echoing a stored answer. It carries an exact copy of each original source body in examples/batch5_authority_systems_capsule/exported_batch5_authority_systems_capsule_bundle/source_modules/ and checks those copies against recorded digests and required text anchors. Alongside that, it runs a small synthetic re-derivation for each mechanism. Every re-derivation recomputes its own verdict from the fixture input. It does not echo a stored answer.
\n
The result carries evidence without inherited authority
\n
The run emits a metadata-only result record for each mechanism, each naming a pass or a refusal and the scope limit it stayed inside. The output is a set of metadata-only result records under receipts/first_wave/batch5_authority_systems_capsule, each stating a pass or a refusal and the scope limit the run stayed inside.
\n
Purpose
\n
One shared discipline across eight subjects
\n
The eight mechanisms share no subject, only the discipline of declining to claim more than they earned, where a loose version accepts a drifted result record, hands Lean a placeholder, or invents coverage. The eight mechanisms share nothing in subject. One checks post-execution result records. Another decides when a reasoning step needs re-running. Another gates a Lean proof attempt. Another classifies a stray process. What they share is a single discipline: each must decline to claim more than it earned. A loose version fails in the obvious ways. The result record check would accept a drifted result record. The proof gate would hand a placeholder to Lean. The orphan check would signal a live process. The blast-radius pass would invent coverage for a leaf that has no dependents.
\n
Synthetic exercises beside verifiable source
\n
Rather than replay the real tools, it verifies copied source against digests and anchors and runs one synthetic exercise per mechanism, passing a negative case only when the exercise itself reaches the refusal. The unusual choice here is that the component does not replay the real tools. It reads the copied source bodies, verifies them against digests and anchors, then runs one synthetic exercise per mechanism. A negative case passes only when the exercise itself reaches the refusal, not when a fixture asserts the refusal for it. That is the point of the design: you can read eight refusal behaviours in one place while the genuine source bodies stay verifiable next to them.
\n
How it works
\n
Two entrypoints share one evaluator
\n
run and run_batch5_bundle both drive run_crown_jewel_organ with _evaluate and evaluate_negative_case, differing only by BUNDLE_INPUT_MODE, and both write metadata-only records with body_in_receipt always False.run and run_batch5_bundle are the two entrypoints. Both call the shared run_crown_jewel_organ driver with SPEC, passing _evaluate as the evaluator and evaluate_negative_case as the negative-case evaluator. run_batch5_bundle sets BUNDLE_INPUT_MODE so the same evaluator runs against the exported source bundle instead of the plain fixture. Both write metadata-only result records and, when asked, an sign-off record whose body_in_receipt field is always False.
\n
The _evaluate pipeline order
\n
_evaluate loads the manifest, checks the ten source paths against SOURCE_REQUIRED_ANCHORS, runs the eight exercises, then recomputes the eight negative-case probes, passing only when no findings were raised._evaluate runs the pipeline in this order. It loads batch5_exercise_manifest.json. It compares the source manifest against SOURCE_REQUIRED_ANCHORS, the map of the ten required source paths to the text anchors each must contain, and emits BATCH5_SOURCE_MODULE_SET_INCOMPLETE if any are missing. It then runs the eight exercises, one per manifest block, and emits BATCH5_MECHANISM_EXERCISE_FAILED for any exercise that does not pass. Last, it recomputes the eight negative-case probes over sorted(EXPECTED_NEGATIVE_CASES). The run status is pass only when no findings were raised.
\n
Exercises recompute and never act
\n
Every mechanism exercise derives its verdict purely from a manifest block, touching no provider, no Lean run, no process signal, and no generated-state mutation. Each exercise reads a small block from the manifest and recomputes a verdict. None call a provider, run Lean, signal a process, or mutate generated state.
\n
Function
What it checks
_receipt_validator_exercise
Given a grant and two post-execution result records, recomputes drift codes for the second: provider_substitution, forbidden_context_sent, output_artifact_divergence, execution_claimed_without_runtime_grant. Passes only when the valid result record is accepted and the drifted one carries at least one code.
_replay_scope_exercise
Compares the context classes a step consumed against the classes that changed. Disjoint sets classify as no_replay, otherwise partial. Passes when the classification matches the fixture's expected value.
_proof_contract_gate_exercise
Scans a candidate proof string before any Lean call. A sorry token, a plan-only phrase (plan: or i will), or a proof that restates the declared theorem without exact each become a failure class and set the gate to rejected_before_lean. Records 0/8 historical banked attempts; no proof-success claim.
_orphan_reaper_exercise
Classifies a process. A live-session descendant becomes requires_owner_check, never safe_close_candidate, and signal_sent stays False.
_fixpoint_drainer_exercise
Walks residual signatures. A residual id that reappears under a moved source signature classifies as settlement_residual_source_moved; a repeated signature as settlement_residual_signature_repeated. mutation_authorized is always False.
_trace_tape_exercise
Joins span text. Under the byte budget it passes the text through. Over budget it truncates to a head budget and appends a #r0001 omitted_trace_bytes pointer plus an omission record with the omitted byte count. A breach with no pointer does not pass.
_blast_radius_exercise
Builds the reverse-dependency graph and takes the transitive closure of dependents for a target. Passes when the target has dependents and the named leaf reports an honestly empty bucket rather than invented coverage.
_doctrine_graph_exercise
Scans doctrine nodes. A node whose code path no longer exists is reported as authority_gap_deleted_code_path; a node marked tombstone is reported with its replacement id. Passes only when both a drift finding and a tombstone are present.
\n
Probes rerun exercises, echo risk tracked
\n
_compute_negative_case_probe reruns each exercise on the case's probe_input and rejects only on a recomputed refusal, while fixture_verdict_echo_risk_count flags any verdict not recomputed from its own input. The negative-case probes reuse those same functions. _compute_negative_case_probe runs the relevant exercise on the case's probe_input, then evaluate_negative_case marks the case rejected only when the recomputed verdict is the refusal. The run tracks a fixture_verdict_echo_risk_count: a case whose verdict was not recomputed from its own probe input counts as risk, so a fixture cannot pass a negative case by simply asserting the answer.
\nDiagram of the mechanism (6 steps).Diagram source & refs
flowchart TD Manifest["Copied source bundle + exercise manifest digests and anchors checked first"] --> Evaluate["_evaluate source-set check, then eight exercises"] Evaluate --> Exercises["Eight mechanism exercises each recomputes its own verdict"] Exercises --> Probes["Negative-case probes rerun exercises on probe_input"] Probes --> Records["metadata-only result records status pass only when no findings"] Records --> Ceiling["Scope limit no external model access, no mutation, no proof success, no launch"]
\n
Negative cases
\n
Eight cases with stable error codes
\n
EXPECTED_NEGATIVE_CASES declares eight cases, each carrying the stable error code the run must raise when its perturbation is caught.EXPECTED_NEGATIVE_CASES declares eight cases, each with a stable error code the run must raise when the perturbation is caught.
The design borrows the lineage, trace, and attestation shape of W3C PROV, OpenTelemetry, and SLSA provenance, while staying bounded to copied public source, synthetic inputs, and refusal checks. This bundle borrows from provenance interchange, trace instrumentation, and software supply-chain attestation practice. W3C PROV models the entities, activities, and agents that produce data so a reader can judge its reliability. OpenTelemetry is a vendor-neutral pattern for traces, metrics, and logs across composed systems. SLSA provenance treats artifact origin, builder identity, and build parameters as explicit attestable metadata. The lineage, trace, and attestation shape is borrowed; the exercise stays bounded to copied public source bodies, synthetic inputs, and refusal checks.
\n
Validation Result record Path
\n
Reader-verifiable commands, run from the microcosm-substrate/ public root:
The first command writes the synthetic result, the second validates copied modules, digests, anchors, and secret-exclusion, and the test file checks exercises, bundle, omission record, and the eight refusal cases. The first command writes the synthetic exercise result. The second validates the copied source modules, their manifest digests, anchor rows, and secret-exclusion posture while keeping source bodies out of the record. The test file checks the exercises, the exported bundle, the omission record, the body-scan boundary, and the eight refusal cases. A pass here means the bounded exercise ran and its refusals held, nothing more.
\n
Scope boundary
\n
Scope limit
\n
The narrow claim and what stays false
\n
The strongest claim is that a reader can inspect eight mechanisms as copied source plus a synthetic exercise and confirm each refusal, with no external model access, proof-success, process signal, mutation, or launch-scope decision. The strongest claim the evidence supports: a reader can inspect eight authority and systems mechanisms as copied public source plus a synthetic exercise, and confirm each refuses its wrong answer, from the commands above alone. The proof boundary is narrow. This page and its generated structured source record are reader evidence, source-linked only. The scope limit holds these facts. No live model or external model access. No Lean proof-success or benchmark claim. No process signal is sent. No generated-state mutation is authorized. No source-file changes, whole-system equivalence, public sharing, or launch-scope decision.
\n",
"components": [
{
"anchor": "component-batch5_authority_systems_capsule",
"component_id": "batch5_authority_systems_capsule",
"public_label": "Authority Systems Source Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 4
},
"link_count": 12,
"links": [
{
"id": "batch5_authority_systems_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Authority Systems Source Bundle",
"relation": "explains",
"url": "component-batch5-authority-systems-capsule.html"
},
{
"id": "mechanism.batch5_authority_systems_capsule.validates_public_authority_systems_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public authority systems bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch5-authority-systems-capsule-validates-public-authority-systems-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.batch5_authority_systems_capsule",
"reader_packet": {
"abstract": "Set 5 Authority and Systems Bundle imports post-execution authority, replay, proof-repair, process, generated-state, trace, blast-radius, and doctrine-graph source bodies as public source-open evidence without claiming live authority or launch-scope decision.",
"authority_ceiling": "Fixture-bound public source-body import, copied-module digest and anchor evidence, synthetic source-exercise evidence, and metadata-only result records only; no live external model access, proof success, process signal authority, generated-state mutation authority, source-file changes, publishing-scope decision, launch-scope decision, whole-system equivalence, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch5_authority_systems_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch5-authority-systems-capsule-diagram-1",
"html_url": "docs/paper-module-batch5-authority-systems-capsule.html#paper-module-batch5-authority-systems-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Manifest[\"Copied source bundle + exercise manifest digests and anchors checked first\"] --> Evaluate[\"_evaluate source-set check, then eight exercises\"] Evaluate --> Exercises[\"Eight mechanism exercises each recomputes its own verdict\"] Exercises --> Probes[\"Negative-case probes rerun exercises on probe_input\"] Probes --> Records[\"metadata-only result records status pass only when no findings\"] Records --> Ceiling[\"Scope limit no external model access, no mutation, no proof success, no launch\"]",
"source_format": "mermaid",
"source_hash": "sha256:e2659e5dc1fa6bf760819b27e851ef1897b488ecf1b928cf4869594c6aff7f77",
"source_line_end": 61,
"source_line_start": 54
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch5_authority_systems_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Authority Systems Source Bundle",
"relation": "explains",
"url": "component-batch5-authority-systems-capsule.html"
},
{
"id": "mechanism.batch5_authority_systems_capsule.validates_public_authority_systems_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public authority systems bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch5-authority-systems-capsule-validates-public-authority-systems-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch5_authority_systems_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 111,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch5_authority_systems_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "PROV",
"section_id": "provenance-tracing-and-supply-chain-attestation",
"url": "https://www.w3.org/TR/prov-overview/"
},
{
"kind": "external_reference",
"label": "OpenTelemetry",
"section_id": "provenance-tracing-and-supply-chain-attestation",
"url": "https://opentelemetry.io/docs/"
},
{
"kind": "external_reference",
"label": "SLSA provenance",
"section_id": "provenance-tracing-and-supply-chain-attestation",
"url": "https://slsa.dev/spec/v1.2/provenance"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "A reader can inspect eight unrelated authority and systems mechanisms at once and confirm each refuses its wrong answer, without running any of the real machinery. This component lets a reader inspect eight unrelated authority and systems mechanisms at once, and confirm that each one refuses the wrong answer, without running any of the real machinery. The eight are: post-execution result record validation, reasoning replay scope, verifier-gated proof repair, process orphan classification, generated-state fixpoint settlement, trace-tape compaction, code blast radius, and doctrine-graph compilation.",
"level": 4,
"order": 1,
"section_id": "authority-system-checks-run-from-copied-bodies",
"source_line_end": 6,
"source_line_start": 3,
"title": "Authority-system checks run from copied bodies"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It ships an exact copy of each source body, checks the copies against digests and anchors, then runs a synthetic re-derivation per mechanism that recomputes its own verdict instead of echoing a stored answer. It carries an exact copy of each original source body in examples/batch5authoritysystemscapsule/exportedbatch5authoritysystemscapsulebundle/source_modules/ and checks those copies against recorded digests and required text anchors. Alongside that, it runs a small synthetic re-derivation for each mechanism. Every re-derivation recomputes its own verdict from the fixture input. It does not echo a stored answer.",
"level": 4,
"order": 2,
"section_id": "fixture-cases-exercise-delegation-and-refusal",
"source_line_end": 10,
"source_line_start": 7,
"title": "Fixture cases exercise delegation and refusal"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The run emits a metadata-only result record for each mechanism, each naming a pass or a refusal and the scope limit it stayed inside. The output is a set of metadata-only result records under result records/firstwave/batch5authoritysystemscapsule, each stating a pass or a refusal and the scope limit the run stayed inside.",
"level": 4,
"order": 3,
"section_id": "the-result-carries-evidence-without-inherited-authority",
"source_line_end": 14,
"source_line_start": 11,
"title": "The result carries evidence without inherited authority"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The eight mechanisms share no subject, only the discipline of declining to claim more than they earned, where a loose version accepts a drifted result record, hands Lean a placeholder, or invents coverage. The eight mechanisms share nothing in subject. One checks post-execution result records. Another decides when a reasoning step needs re-running. Another gates a Lean proof attempt. Another classifies a stray process. What they share is a single discipline: each must decline to claim more than it earned. A loose version fails in the obvious ways. The result record check would accept a drifted result record. The proof gate would hand a placeholder to Lean. The orphan check would signal a li\u2026",
"level": 4,
"order": 5,
"section_id": "one-shared-discipline-across-eight-subjects",
"source_line_end": 20,
"source_line_start": 17,
"title": "One shared discipline across eight subjects"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Rather than replay the real tools, it verifies copied source against digests and anchors and runs one synthetic exercise per mechanism, passing a negative case only when the exercise itself reaches the refusal. The unusual choice here is that the component does not replay the real tools. It reads the copied source bodies, verifies them against digests and anchors, then runs one synthetic exercise per mechanism. A negative case passes only when the exercise itself reaches the refusal, not when a fixture asserts the refusal for it. That is the point of the design: you can read eight refusal behaviours in one place while the genuine source bodies stay verifiable next to them.",
"level": 4,
"order": 6,
"section_id": "synthetic-exercises-beside-verifiable-source",
"source_line_end": 24,
"source_line_start": 21,
"title": "Synthetic exercises beside verifiable source"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run and runbatch5bundle both drive runcrownjewelorgan with evaluate and evaluatenegativecase, differing only by BUNDLEINPUTMODE, and both write metadata-only records with bodyinreceipt always False. run and runbatch5bundle are the two entrypoints. Both call the shared runcrownjewelorgan driver with SPEC, passing evaluate as the evaluator and evaluatenegativecase as the negative-case evaluator. runbatch5bundle sets BUNDLEINPUTMODE so the same evaluator runs against the exported source bundle instead of the plain fixture. Both write metadata-only result records and, when asked, an sign-off record whose bodyinreceipt field is always False.",
"level": 4,
"order": 8,
"section_id": "two-entrypoints-share-one-evaluator",
"source_line_end": 30,
"source_line_start": 27,
"title": "Two entrypoints share one evaluator"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluate loads the manifest, checks the ten source paths against SOURCEREQUIREDANCHORS, runs the eight exercises, then recomputes the eight negative-case probes, passing only when no findings were raised. evaluate runs the pipeline in this order. It loads batch5exercisemanifest.json. It compares the source manifest against SOURCEREQUIREDANCHORS, the map of the ten required source paths to the text anchors each must contain, and emits BATCH5SOURCEMODULESETINCOMPLETE if any are missing. It then runs the eight exercises, one per manifest block, and emits BATCH5MECHANISMEXERCISEFAILED for any exercise that does not pass. Last, it recomputes the eight negative-case probes over sorted(EXPECTEDNEG\u2026",
"level": 4,
"order": 9,
"section_id": "the-evaluate-pipeline-order",
"source_line_end": 34,
"source_line_start": 31,
"title": "The `_evaluate` pipeline order"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "Every mechanism exercise derives its verdict purely from a manifest block, touching no provider, no Lean run, no process signal, and no generated-state mutation. Each exercise reads a small block from the manifest and recomputes a verdict. None call a provider, run Lean, signal a process, or mutate generated state.",
"level": 4,
"order": 10,
"section_id": "exercises-recompute-and-never-act",
"source_line_end": 49,
"source_line_start": 35,
"title": "Exercises recompute and never act"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "computenegativecaseprobe reruns each exercise on the case's probeinput and rejects only on a recomputed refusal, while fixtureverdictechoriskcount flags any verdict not recomputed from its own input. The negative-case probes reuse those same functions. computenegativecaseprobe runs the relevant exercise on the case's probeinput, then evaluatenegativecase marks the case rejected only when the recomputed verdict is the refusal. The run tracks a fixtureverdictechorisk_count: a case whose verdict was not recomputed from its own probe input counts as risk, so a fixture cannot pass a negative case by simply asserting the answer.",
"level": 4,
"order": 11,
"section_id": "probes-rerun-exercises-echo-risk-tracked",
"source_line_end": 62,
"source_line_start": 50,
"title": "Probes rerun exercises, echo risk tracked"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 64,
"source_line_start": 63,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "EXPECTEDNEGATIVECASES declares eight cases, each carrying the stable error code the run must raise when its perturbation is caught. EXPECTEDNEGATIVECASES declares eight cases, each with a stable error code the run must raise when the perturbation is caught. receiptprovidercontextdrift gives BATCH5RECEIPTDRIFTREJECTED. irrelevantcontextnoreplay gives BATCH5REPLAYIRRELEVANTCONTEXTNOREPLAY. proofcontractsorry gives BATCH5PROOFCONTRACTBADPROOF_REJECTED. livedescendantneversignal gives BATCH5ORPHANREAPERLIVEDESCENDANTNEVER_SIGNAL. nonconvergingresidualclassified gives BATCH5DRAINERNONCONVERGINGRESIDUALCLASSIFIED. traceomissionreceiptrequired gives BATCH5TRACEOMISSIONRECEIPT_REQUIRED. blastradius\u2026",
"level": 4,
"order": 13,
"section_id": "eight-cases-with-stable-error-codes",
"source_line_end": 78,
"source_line_start": 65,
"title": "Eight cases with stable error codes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "prior-art-grounding",
"source_line_end": 80,
"source_line_start": 79,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The design borrows the lineage, trace, and attestation shape of W3C PROV, OpenTelemetry, and SLSA provenance, while staying bounded to copied public source, synthetic inputs, and refusal checks. This bundle borrows from provenance interchange, trace instrumentation, and software supply-chain attestation practice. W3C PROV models the entities, activities, and agents that produce data so a reader can judge its reliability. OpenTelemetry is a vendor-neutral pattern for traces, metrics, and logs across composed systems. SLSA provenance treats artifact origin, builder identity, and build parameters as explicit attestable metadata. The lineage, trace, and attestation shape is borrowed; the exerci\u2026",
"level": 4,
"order": 15,
"section_id": "provenance-tracing-and-supply-chain-attestation",
"source_line_end": 84,
"source_line_start": 81,
"title": "Provenance, tracing, and supply-chain attestation"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Reader-verifiable commands, run from the microcosm-substrate/ public root:",
"level": 2,
"order": 16,
"section_id": "validation-receipt-path",
"source_line_end": 100,
"source_line_start": 85,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The first command writes the synthetic result, the second validates copied modules, digests, anchors, and secret-exclusion, and the test file checks exercises, bundle, omission record, and the eight refusal cases. The first command writes the synthetic exercise result. The second validates the copied source modules, their manifest digests, anchor rows, and secret-exclusion posture while keeping source bodies out of the record. The test file checks the exercises, the exported bundle, the omission record, the body-scan boundary, and the eight refusal cases. A pass here means the bounded exercise ran and its refusals held, nothing more.",
"level": 4,
"order": 17,
"section_id": "what-each-command-verifies",
"source_line_end": 104,
"source_line_start": 101,
"title": "What each command verifies"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "claim-ceiling",
"source_line_end": 106,
"source_line_start": 105,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim is that a reader can inspect eight mechanisms as copied source plus a synthetic exercise and confirm each refusal, with no external model access, proof-success, process signal, mutation, or launch-scope decision. The strongest claim the evidence supports: a reader can inspect eight authority and systems mechanisms as copied public source plus a synthetic exercise, and confirm each refuses its wrong answer, from the commands above alone. The proof boundary is narrow. This page and its generated structured source record are reader evidence, source-linked only. The scope limit holds these facts. No live model or external model access. No Lean proof-success or benchmark clai\u2026",
"level": 4,
"order": 19,
"section_id": "the-narrow-claim-and-what-stays-false",
"source_line_end": 110,
"source_line_start": 107,
"title": "The narrow claim and what stays false"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch5_authority_systems_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 5 Authority and Systems Bundle imports post-execution authority, replay, proof-repair, process, generated-state, trace, blast-radius, and doctrine-graph source bodies as public source-open evidence without claiming live authority or launch-scope decision."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch5_authority_systems_capsule.py",
"path": "src/microcosm_core/organs/batch5_authority_systems_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for Set 5 authority-system source imports, synthetic exercises, exact-copy digest checks, negative cases, and scope limits.",
"symbols": [
"_receipt_validator_exercise",
"_replay_scope_exercise",
"_proof_contract_gate_exercise",
"_orphan_reaper_exercise",
"_fixpoint_drainer_exercise",
"_trace_tape_exercise",
"_blast_radius_exercise",
"_doctrine_graph_exercise"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch5_authority_systems_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch5_authority_systems_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch5_authority_systems_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 5 Authority and Systems Bundle binds the legacy Markdown projection, Set 5 runnable source locus, exported copied-source bundle, source-module digests, synthetic negative exercises, metadata-only result records, and scope limits to a mechanism-backed JSON bundle. It covers post-execution result record validation, reasoning replay scope and lineage, verifier-gated Lean repair harnessing, process orphan classification, generated-state fixpoint settlement, trace-tape compaction, code blast radius, and doctrine graph compilation while preserving fixture boundaries.",
"summary_line": "Set 5 Authority and Systems Bundle imports post-execution authority, replay, proof-repair, process, generated-state, trace, blast-radius, and doctrine-graph source bodies as public source-open evidence without claiming live authority or launch-scope decision.",
"title": "Set 5 Authority and Systems Bundle"
},
{
"anchor": "paper-module-batch7-oracle-sibling-capsule",
"authority_ceiling": "Fixture-bound public Oracle source-body import, copied-module digest and anchor evidence, deterministic local exercise evidence, original pytest witness evidence, and metadata-only result records only; no Oracle reasoning authority, semantic truth authority, external model access, bridge-backed reasoning, private orchestration engine invocation, source-file changes, publishing-scope decision, launch-scope decision, whole-system equivalence, complete Oracle coverage, or whole-system correctness.",
"body_html": "
Oracle-sibling behavior is replayed from copied code
\n
The Oracle mostly runs on live feeds and a bridge-backed reasoning engine, so this component asks which parts are pure and inspectable without any feed, provider, or reasoning call. The Oracle is a sibling system that reasons about market evidence. Most of it depends on live data feeds and on a bridge-backed reasoning engine, so it cannot be shown to a public reader directly. This component answers a narrower question. Which parts of the Oracle are pure, deterministic, and inspectable, and can those parts be run and checked without touching any feed, provider, or reasoning call?
\n
The fixture separates advisory output from authority
\n
The answer is the grounding and bookkeeping layer: four network-free tools that map evidence to targets, hydrate saved artifacts, diff timed feeds, and plan chain repair. The answer is the Oracle's grounding and bookkeeping layer. Before the Oracle reasons, it builds a map of which evidence supports which prediction target, it hydrates a saved artifact from a recorded run, it diffs two timed snapshots of the same economic feed, and it plans how to repair a missing artifact chain. None of that needs the network or the reasoning engine. This component imports the exact source for those four tools, exercises each against small in-memory runs, and re-runs the Oracle's own tests as an independent witness.
\n
A passing oracle replay remains non-authoritative
\n
The output stays metadata-only: a result record of counts, booleans, and digests, nothing more. Every run emits a result record of counts, booleans, and digests. Source bodies and test output never enter it.
\n
Purpose
\n
Why replay against the real source
\n
A paraphrased or hand-tuned replay would be indistinguishable from theatre, so the component drives the real Oracle source against small readable fixtures and stops before any reasoning call. A loose replay would be easy to fake and hard to trust. If the bundle copied paraphrased logic, or ran the tools against inputs shaped to pass, a reader could not tell inspection from theatre. So the component imports the real Oracle source and drives it against fixtures small enough to read but specific enough to make the domain rules visible. It also stops at a hard line. The Oracle's repair planner can, if asked, call the bridge-backed GodModeEngine to fill a missing node by reasoning. This component walks the planner up to that point and no further. It builds the plan, copies one aliasable artifact, and records run_missing_quartet and GodModeEngine as excluded. The exclusion is part of the evidence, not a gap in it.
\n
How it works
\n
_evaluate orders the pipeline
\n
_evaluate runs the pytest witness then five engines in fixed order, and a single engine that misses pass turns the whole result to blocked._evaluate runs the pipeline in a fixed order. It first calls _run_original_pytest_witness, then the five engines listed below, then it collects findings from the source manifest check, the witness, and each engine. Any engine that does not reach pass becomes a structured finding, and one finding turns the whole result to blocked. run and run_batch7_oracle_sibling_bundle wrap _evaluate through the shared run_crown_jewel_organ runner, which handles input loading, the source-manifest gate, and result record writing.
\n
Function
What it does
_run_original_pytest_witness
Runs the Oracle's own focused tests as a subprocess, requires at least seven passed, records return code and stdout/stderr byte counts only
_subject_index_engine
Drives subject_index.run on a seeded run and checks the grounding map
_subject_snapshot_engine
Drives subject_snapshot.run and checks artifact provenance survives hydration
_truth_diff_macro_engine
Drives truth_diff_macro.run on two timed feeds and checks the series delta
_quartet_repair_engine
Drives build_quartet_repair_plan and materialize_missing_aliases, checks the plan and one written alias
_original_pytest_engine
Folds the witness status into an engine record
\n
Grounding a price prediction to admissible support
\n
Only stock or ETF evidence (the S_ or E_ prefix) can anchor a price, so XOM and XLE ground while economic TLT is carried forward as an unmet gap. The subject-index engine seeds three pieces of evidence through _seed_subject_run: a stock (XOM, ledger id S_CAFEBABE), an ETF (XLE, ledger id E_DEADBEEF), and an economic instrument (TLT, ledger id M_FEEDBEEF). The Oracle's rule is that evidence grounds a price prediction only when its subject is a valid target and its ledger id marks it as stock or ETF support, the S_ or E_ prefix. So XOM and XLE land in admissible_evidence_by_subject, while TLT stays in contextual_evidence_by_subject even though it is a valid target, because its support is context that cannot anchor a price. The engine then checks that TLT appears in missing_admissible_support_targets. The Oracle does not silently drop a target it cannot ground, it carries the gap forward.
\n
Provenance survives snapshot hydration
\n
A hydrated artifact must still carry its source_artifact_id and subject_run_id, so a caller knows which run a payload came from, with no body copied. The subject-snapshot engine writes one lab_director artifact and hydrates it by id. The result must carry source_artifact_id equal to lab_director and subject_run_id equal to subject, so a downstream caller knows which run a prediction payload came from. No artifact body is copied into the result record.
\n
Diffing two timed snapshots of a feed
\n
Diffing a 14:00 and 22:00 snapshot names the changed series by absolute delta (US10Y 4.10 to 4.35), the newly appeared CPI, and the dropped OIL. The truth-diff engine writes two snapshots of the same economic feed, a subject snapshot at 14:00 and a truth snapshot at 22:00. It confirms the diff names the changed series ranked by strongest absolute delta, here US10Y moving from 4.10 to 4.35, a newly appeared series CPI, and a dropped series OIL. This is the difference between a number changing and the system knowing which number changed, by how much, and whether the set of series itself shifted.
\n
Planning repair up to the reasoning boundary
\n
The repair planner reports BLOCKED, names the deepest missing node, and writes one alias, but stops short of ever constructing the reasoning-backed GodModeEngine. The quartet engine exercises the repair planner against a truth run missing most of its node chain. The plan must report readiness BLOCKED, name the deepest missing target oracle_cp2_emitter, and list prediction_reconciliation as aliasable. The engine materialises that one alias and reads the written file back to confirm its artifact_alias_of records oracle_truth_diff_equity. It asserts run_missing_quartet_excluded and that GodModeEngine was never constructed.
\nDiagram of the mechanism (8 steps).Diagram source & refs
flowchart TD input["Fixture input exercise manifest"] witness["_run_original_pytest_witness focused Oracle tests, >=7 passed"] index["_subject_index_engine admissible vs contextual grounding"] snap["_subject_snapshot_engine provenance survives hydration"] diff["_truth_diff_macro_engine changed / new / dropped series"] quartet["_quartet_repair_engine plan BLOCKED, one alias written"] stop["excluded: run_missing_quartet / private orchestration engine"] record["Result record counts, booleans, digests; no bodies"] input --> witness witness --> index index --> snap snap --> diff diff --> quartet quartet --> stop quartet --> record
\n
Negative cases
\n
Five malformed inputs, one error code each
\n
Five negative inputs are declared in EXPECTED_NEGATIVE_CASES, each mapped to a stable error code and run through evaluate_negative_case. The component declares five negative cases in EXPECTED_NEGATIVE_CASES, each mapped to a stable error code, and evaluated through evaluate_negative_case.
\n
What each rejection code guards
\n
Missing run, artifact-id, or truth directories each reject with a named code, the excluded quartet node is enforced, and observations record only the exception type, never its body. A subject-index run with no run directory rejects with BATCH7_ORACLE_SUBJECT_RUN_REQUIRED. A snapshot run with no artifact id rejects with BATCH7_ORACLE_ARTIFACT_ID_REQUIRED. A truth-diff run with no truth directory rejects with BATCH7_ORACLE_TRUTH_RUN_REQUIRED. The quartet planner must both refuse to run the missing node and confirm the manifest lists run_missing_quartet as excluded, tracked as BATCH7_ORACLE_RUN_MISSING_EXCLUDED. If the Oracle's own tests do not pass as a witness, the case BATCH7_ORACLE_ORIGINAL_PYTEST_WITNESS_REQUIRED fires. Each observation records only the exception type and whether the expected message fragment appeared, never the exception body.
\n
Prior Art Grounding
\n
Test-oracle and provenance lineage
\n
The design reuses two known ideas, the test oracle that decides whether an output is acceptable and W3C PROV data provenance, for deterministic replay, not for semantic truth. The component draws on two established practices. A test oracle is the part of a test that decides whether an observed output is acceptable, surveyed in the test oracle problem literature. Data provenance records where an artifact came from and how it was derived, modelled by the W3C PROV family of entities, activities, and agents. The subject index and snapshot tools apply the provenance idea, the truth-diff and quartet checks apply the oracle idea. This component reuses those patterns for deterministic replay. It does not turn a local check into semantic truth.
\n
Validation Result record Path
\n
Run from microcosm-substrate/:
\n
What a result record writes and proves
\n
The run writes result, board, and validation-result record files, and a pass means the five engines met their contracts and the record held no bodies, nothing more. The run writes batch7_oracle_sibling_capsule_result.json, batch7_oracle_sibling_capsule_board.json, and batch7_oracle_sibling_capsule_validation_receipt.json under the output directory, plus the sign-off file. The focused test suite is tests/test_batch7_oracle_sibling_capsule.py. A pass means the five engines met their contracts, the negative cases rejected as declared, and the result record held no bodies. It means nothing beyond that run.
\n
Scope boundary
\n
Scope limit
\n
The bounded claim and its ceiling
\n
The strongest claim is a bounded source-open replay over one fixture; it runs no Oracle reasoning, dispatches no provider, and excludes launch, mutation, or semantic truth. The strongest claim this component supports is a bounded, source-open replay: the real Oracle grounding tools ran against readable fixtures, produced the exact expected shapes, rejected five malformed inputs, and left no source or test bodies in the result record. That is the proof boundary and the scope limit. It does not run Oracle reasoning, dispatch any provider or model, invoke run_missing_quartet or GodModeEngine, touch a market feed or wallet, change source files, or prove complete Oracle coverage. Its AUTHORITY_CEILING sets release_authorized, publication_authorized, provider_dispatch, source_mutation_authorized, and semantic_truth_authority to false. A green run is evidence about one fixture, not about the whole system.
\n",
"components": [
{
"anchor": "component-batch7_oracle_sibling_capsule",
"component_id": "batch7_oracle_sibling_capsule",
"public_label": "Oracle Sibling Source Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 4
},
"link_count": 12,
"links": [
{
"id": "batch7_oracle_sibling_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Oracle Sibling Source Bundle",
"relation": "explains",
"url": "component-batch7-oracle-sibling-capsule.html"
},
{
"id": "mechanism.batch7_oracle_sibling_capsule.validates_public_oracle_sibling_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public oracle sibling bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch7-oracle-sibling-capsule-validates-public-oracle-sibling-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch8_tools_tail_primitives_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Tools-Tail Primitives Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[62:paper_module.batch8_tools_tail_primitives_capsule].depends_on",
"url": "paper-module-batch8-tools-tail-primitives-capsule.html"
}
],
"paper_module_id": "paper_module.batch7_oracle_sibling_capsule",
"reader_packet": {
"abstract": "Set 7 Oracle Sibling Bundle imports Oracle sibling source bodies and exercises deterministic subject-index, snapshot, truth-diff, quartet-plan, and original pytest witness boundaries.",
"authority_ceiling": "Fixture-bound public Oracle source-body import, copied-module digest and anchor evidence, deterministic local exercise evidence, original pytest witness evidence, and metadata-only result records only; no Oracle reasoning authority, semantic truth authority, external model access, bridge-backed reasoning, private orchestration engine invocation, source-file changes, publishing-scope decision, launch-scope decision, whole-system equivalence, complete Oracle coverage, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch7_oracle_sibling_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch7-oracle-sibling-capsule-diagram-1",
"html_url": "docs/paper-module-batch7-oracle-sibling-capsule.html#paper-module-batch7-oracle-sibling-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD input[\"Fixture input exercise manifest\"] witness[\"_run_original_pytest_witness focused Oracle tests, >=7 passed\"] index[\"_subject_index_engine admissible vs contextual grounding\"] snap[\"_subject_snapshot_engine provenance survives hydration\"] diff[\"_truth_diff_macro_engine changed / new / dropped series\"] quartet[\"_quartet_repair_engine plan BLOCKED, one alias written\"] stop[\"excluded: run_missing_quartet / private orchestration engine\"] record[\"Result record counts, booleans, digests; no bodies\"] input --> witness witness --> index index --> snap snap --> diff diff --> quartet quartet --> stop quartet --> record",
"source_format": "mermaid",
"source_hash": "sha256:698a62907828b58ad58dde695d337d8b43af9178bfd43eb930262d339df761ed",
"source_line_end": 69,
"source_line_start": 52
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch7_oracle_sibling_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Oracle Sibling Source Bundle",
"relation": "explains",
"url": "component-batch7-oracle-sibling-capsule.html"
},
{
"id": "mechanism.batch7_oracle_sibling_capsule.validates_public_oracle_sibling_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public oracle sibling bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch7-oracle-sibling-capsule-validates-public-oracle-sibling-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch7_oracle_sibling_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 110,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch7_oracle_sibling_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "test oracle problem",
"section_id": "test-oracle-and-provenance-lineage",
"url": "https://discovery.ucl.ac.uk/1471263/"
},
{
"kind": "external_reference",
"label": "PROV",
"section_id": "test-oracle-and-provenance-lineage",
"url": "https://www.w3.org/TR/prov-overview/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "The Oracle mostly runs on live feeds and a bridge-backed reasoning engine, so this component asks which parts are pure and inspectable without any feed, provider, or reasoning call. The Oracle is a sibling system that reasons about market evidence. Most of it depends on live data feeds and on a bridge-backed reasoning engine, so it cannot be shown to a public reader directly. This component answers a narrower question. Which parts of the Oracle are pure, deterministic, and inspectable, and can those parts be run and checked without touching any feed, provider, or reasoning call?",
"level": 4,
"order": 1,
"section_id": "oracle-sibling-behavior-is-replayed-from-copied-code",
"source_line_end": 6,
"source_line_start": 3,
"title": "Oracle-sibling behavior is replayed from copied code"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The answer is the grounding and bookkeeping layer: four network-free tools that map evidence to targets, hydrate saved artifacts, diff timed feeds, and plan chain repair. The answer is the Oracle's grounding and bookkeeping layer. Before the Oracle reasons, it builds a map of which evidence supports which prediction target, it hydrates a saved artifact from a recorded run, it diffs two timed snapshots of the same economic feed, and it plans how to repair a missing artifact chain. None of that needs the network or the reasoning engine. This component imports the exact source for those four tools, exercises each against small in-memory runs, and re-runs the Oracle's own tests as an independen\u2026",
"level": 4,
"order": 2,
"section_id": "the-fixture-separates-advisory-output-from-authority",
"source_line_end": 10,
"source_line_start": 7,
"title": "The fixture separates advisory output from authority"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The output stays metadata-only: a result record of counts, booleans, and digests, nothing more. Every run emits a result record of counts, booleans, and digests. Source bodies and test output never enter it.",
"level": 4,
"order": 3,
"section_id": "a-passing-oracle-replay-remains-non-authoritative",
"source_line_end": 14,
"source_line_start": 11,
"title": "A passing oracle replay remains non-authoritative"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A paraphrased or hand-tuned replay would be indistinguishable from theatre, so the component drives the real Oracle source against small readable fixtures and stops before any reasoning call. A loose replay would be easy to fake and hard to trust. If the bundle copied paraphrased logic, or ran the tools against inputs shaped to pass, a reader could not tell inspection from theatre. So the component imports the real Oracle source and drives it against fixtures small enough to read but specific enough to make the domain rules visible. It also stops at a hard line. The Oracle's repair planner can, if asked, call the bridge-backed private orchestration engine to fill a missing node by reasoning\u2026",
"level": 4,
"order": 5,
"section_id": "why-replay-against-the-real-source",
"source_line_end": 20,
"source_line_start": 17,
"title": "Why replay against the real source"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 6,
"section_id": "how-it-works",
"source_line_end": 22,
"source_line_start": 21,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "evaluate runs the pytest witness then five engines in fixed order, and a single engine that misses pass turns the whole result to blocked. evaluate runs the pipeline in a fixed order. It first calls runoriginalpytestwitness, then the five engines listed below, then it collects findings from the source manifest check, the witness, and each engine. Any engine that does not reach pass becomes a structured finding, and one finding turns the whole result to blocked. run and runbatch7oraclesiblingbundle wrap evaluate through the shared runcrownjewel_organ runner, which handles input loading, the source-manifest gate, and result record writing.",
"level": 4,
"order": 7,
"section_id": "evaluate-orders-the-pipeline",
"source_line_end": 35,
"source_line_start": 23,
"title": "`_evaluate` orders the pipeline"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Only stock or ETF evidence (the S or E prefix) can anchor a price, so XOM and XLE ground while economic TLT is carried forward as an unmet gap. The subject-index engine seeds three pieces of evidence through seedsubjectrun: a stock (XOM, ledger id SCAFEBABE), an ETF (XLE, ledger id EDEADBEEF), and an economic instrument (TLT, ledger id MFEEDBEEF). The Oracle's rule is that evidence grounds a price prediction only when its subject is a valid target and its ledger id marks it as stock or ETF support, the S or E prefix. So XOM and XLE land in admissibleevidencebysubject, while TLT stays in contextualevidencebysubject even though it is a valid target, because its support is context that cannot\u2026",
"level": 4,
"order": 8,
"section_id": "grounding-a-price-prediction-to-admissible-support",
"source_line_end": 39,
"source_line_start": 36,
"title": "Grounding a price prediction to admissible support"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A hydrated artifact must still carry its sourceartifactid and subjectrunid, so a caller knows which run a payload came from, with no body copied. The subject-snapshot engine writes one labdirector artifact and hydrates it by id. The result must carry sourceartifactid equal to labdirector and subjectrunid equal to subject, so a downstream caller knows which run a prediction payload came from. No artifact body is copied into the result record.",
"level": 4,
"order": 9,
"section_id": "provenance-survives-snapshot-hydration",
"source_line_end": 43,
"source_line_start": 40,
"title": "Provenance survives snapshot hydration"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Diffing a 14:00 and 22:00 snapshot names the changed series by absolute delta (US10Y 4.10 to 4.35), the newly appeared CPI, and the dropped OIL. The truth-diff engine writes two snapshots of the same economic feed, a subject snapshot at 14:00 and a truth snapshot at 22:00. It confirms the diff names the changed series ranked by strongest absolute delta, here US10Y moving from 4.10 to 4.35, a newly appeared series CPI, and a dropped series OIL. This is the difference between a number changing and the system knowing which number changed, by how much, and whether the set of series itself shifted.",
"level": 4,
"order": 10,
"section_id": "diffing-two-timed-snapshots-of-a-feed",
"source_line_end": 47,
"source_line_start": 44,
"title": "Diffing two timed snapshots of a feed"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "The repair planner reports BLOCKED, names the deepest missing node, and writes one alias, but stops short of ever constructing the reasoning-backed private orchestration engine. The quartet engine exercises the repair planner against a truth run missing most of its node chain. The plan must report readiness BLOCKED, name the deepest missing target oraclecp2emitter, and list predictionreconciliation as aliasable. The engine materialises that one alias and reads the written file back to confirm its artifactaliasof records oracletruthdiffequity. It asserts runmissingquartet_excluded and that private orchestration engine was never constructed.",
"level": 4,
"order": 11,
"section_id": "planning-repair-up-to-the-reasoning-boundary",
"source_line_end": 70,
"source_line_start": 48,
"title": "Planning repair up to the reasoning boundary"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 72,
"source_line_start": 71,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Five negative inputs are declared in EXPECTEDNEGATIVECASES, each mapped to a stable error code and run through evaluatenegativecase. The component declares five negative cases in EXPECTEDNEGATIVECASES, each mapped to a stable error code, and evaluated through evaluatenegativecase.",
"level": 4,
"order": 13,
"section_id": "five-malformed-inputs-one-error-code-each",
"source_line_end": 77,
"source_line_start": 73,
"title": "Five malformed inputs, one error code each"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Missing run, artifact-id, or truth directories each reject with a named code, the excluded quartet node is enforced, and observations record only the exception type, never its body. A subject-index run with no run directory rejects with BATCH7ORACLESUBJECTRUNREQUIRED. A snapshot run with no artifact id rejects with BATCH7ORACLEARTIFACTIDREQUIRED. A truth-diff run with no truth directory rejects with BATCH7ORACLETRUTHRUNREQUIRED. The quartet planner must both refuse to run the missing node and confirm the manifest lists runmissingquartet as excluded, tracked as BATCH7ORACLERUNMISSINGEXCLUDED. If the Oracle's own tests do not pass as a witness, the case BATCH7ORACLEORIGINALPYTESTWITNESS_REQUI\u2026",
"level": 4,
"order": 14,
"section_id": "what-each-rejection-code-guards",
"source_line_end": 81,
"source_line_start": 78,
"title": "What each rejection code guards"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 83,
"source_line_start": 82,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The design reuses two known ideas, the test oracle that decides whether an output is acceptable and W3C PROV data provenance, for deterministic replay, not for semantic truth. The component draws on two established practices. A test oracle is the part of a test that decides whether an observed output is acceptable, surveyed in the test oracle problem literature. Data provenance records where an artifact came from and how it was derived, modelled by the W3C PROV family of entities, activities, and agents. The subject index and snapshot tools apply the provenance idea, the truth-diff and quartet checks apply the oracle idea. This component reuses those patterns for deterministic replay. It do\u2026",
"level": 4,
"order": 16,
"section_id": "test-oracle-and-provenance-lineage",
"source_line_end": 87,
"source_line_start": 84,
"title": "Test-oracle and provenance lineage"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from microcosm-substrate/:",
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 99,
"source_line_start": 88,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The run writes result, board, and validation-result record files, and a pass means the five engines met their contracts and the record held no bodies, nothing more. The run writes batch7oraclesiblingcapsuleresult.json, batch7oraclesiblingcapsuleboard.json, and batch7oraclesiblingcapsulevalidationreceipt.json under the output directory, plus the sign-off file. The focused test suite is tests/testbatch7oraclesibling_capsule.py. A pass means the five engines met their contracts, the negative cases rejected as declared, and the result record held no bodies. It means nothing beyond that run.",
"level": 4,
"order": 18,
"section_id": "what-a-receipt-writes-and-proves",
"source_line_end": 103,
"source_line_start": 100,
"title": "What a result record writes and proves"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "claim-ceiling",
"source_line_end": 105,
"source_line_start": 104,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The strongest claim is a bounded source-open replay over one fixture; it runs no Oracle reasoning, dispatches no provider, and excludes launch, mutation, or semantic truth. The strongest claim this component supports is a bounded, source-open replay: the real Oracle grounding tools ran against readable fixtures, produced the exact expected shapes, rejected five malformed inputs, and left no source or test bodies in the result record. That is the proof boundary and the scope limit. It does not run Oracle reasoning, dispatch any provider or model, invoke runmissingquartet or private orchestration engine, touch a market feed or wallet, change source files, or prove complete Oracle coverage. It\u2026",
"level": 4,
"order": 20,
"section_id": "the-bounded-claim-and-its-ceiling",
"source_line_end": 109,
"source_line_start": 106,
"title": "The bounded claim and its ceiling"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch7_oracle_sibling_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 7 Oracle Sibling Bundle imports Oracle sibling source bodies and exercises deterministic subject-index, snapshot, truth-diff, quartet-plan, and original pytest witness boundaries."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch7_oracle_sibling_capsule.py",
"path": "src/microcosm_core/organs/batch7_oracle_sibling_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for Set 7 public Oracle sibling source imports, deterministic exercises, original pytest witnesses, exact-copy digest checks, negative cases, and scope limits.",
"symbols": [
"_subject_index_engine",
"_subject_snapshot_engine",
"_truth_diff_macro_engine",
"_quartet_repair_engine",
"_run_original_pytest_witness",
"_original_pytest_engine",
"_evaluate",
"run"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch7_oracle_sibling_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch7_oracle_sibling_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch7_oracle_sibling_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 7 Oracle Sibling Bundle binds the runnable Oracle sibling source locus, exported copied-source bundle, source-module digests, original pytest witnesses, deterministic subject-index/snapshot/truth-diff/quartet-plan exercises, negative cases, metadata-only result records, and scope limits to a mechanism-backed JSON bundle without claiming accepted-component authority or semantic truth authority.",
"summary_line": "Set 7 Oracle Sibling Bundle imports Oracle sibling source bodies and exercises deterministic subject-index, snapshot, truth-diff, quartet-plan, and original pytest witness boundaries.",
"title": "Set 7 Oracle Sibling Bundle"
},
{
"anchor": "paper-module-batch7-demo-take-console-capsule",
"authority_ceiling": "Fixture-bound public Swift source-body import, copied-module digest and anchor evidence, deterministic local exercise evidence, SwiftPM build-witness evidence, and metadata-only result records only; no app launch authority, screen capture authority, microphone capture authority, recording-session export, FFmpeg execution, WhisperKit/model dispatch, source-file changes, publishing-scope decision, launch-scope decision, whole-system equivalence, complete UI coverage, or whole-system correctness.",
"body_html": "
The console bundle replays one copied runtime
\n
This component reads the Demo Take Console's public Swift source and checks that its safety gates actually hold, without ever launching the macOS capture app. The Demo Take Console is a macOS app that records screen and microphone takes, drives FFmpeg through a Python helper, and hands the audio to an on-device WhisperKit transcriber. This component reads the app's public Swift source and checks that the safety gates in that source actually hold, without ever launching the app.
\n
Three checks bind inputs, output, and ceiling
\n
It copies eight Swift files, runs six source-token checks plus six paired deletion cases, and emits records of exit codes, counts, and booleans, never source bodies or logs. It copies eight Swift files into a public bundle, runs six checks that read those copies, then runs six paired negative cases that delete one guard and demand the matching check notice. It emits result records that carry exit codes, counts, and booleans, never the source bodies, recordings, or build logs.
\n
The demo result stays metadata-only and local
\n
A green run says only that the source contracts are present, not that capture, FFmpeg, or a model ever ran. A green run says the source contracts are present. It does not start capture, run FFmpeg, or dispatch a model.
\n
Purpose
\n
Do the source gates exist as source
\n
A capture app is trusted with screen, microphone, and disk, so the narrow question is whether the gates written into its Swift source exist as source. A capture app is trusted with a screen, a microphone, and disk. The question this component answers is narrow: before anyone trusts the app to record, do the gates written into its Swift source exist as source?
\n
Each check asserts a behaviour-carrying token
\n
Because a file can name clamp yet ship an unclamped meter, each check asserts a specific behaviour-carrying token in the copied source rather than shape alone. Checking shape alone is not enough. A file can contain the word "clamp" and still ship an unclamped meter. So each check asserts a specific token that carries the behaviour: the start gate requires a chosen recording source and free disk, the global hotkey requires Control-Option-Command, the audio meter returns min(max(rms * 8, 0), 1), the transcriber guards a missing audio file. The checks read the copied source, not the app's live state.
\n
How it works
\n
Two entrypoints and two callbacks
\n
run and run_batch7_demo_take_bundle both drive run_crown_jewel_organ with _evaluate for positive checks and evaluate_negative_case for mutation tests, over the fixture or the exported copy.run and run_batch7_demo_take_bundle both call run_crown_jewel_organ with two callbacks: _evaluate for the positive checks and evaluate_negative_case for the mutation tests. run reads the fixture input; the bundle variant reads the exported public copy. main wires the run and validate-bundle subcommands to the same runner and prints a result or a card.
\n
Six engines over eight copied modules
\n
_evaluate runs six engines, failing on any non-pass status or fewer than eight copied modules, with each engine reading its Swift file through _read and returning booleans plus a claim_ceiling._evaluate runs six engines in order and collects a finding for any that does not return status: "pass". It also fails if the copied source manifest holds fewer than eight modules, the count in SOURCE_REQUIRED_ANCHORS. Each engine reads its Swift file through _read, which resolves the copied path under the exported bundle, and returns a small dict of booleans plus a claim_ceiling string.
\n
Function
Reads
Passes when
_swiftpm_build_witness
app target
swift build --product DemoTakeConsoleApp exits zero
_recording_state_control_model
Models.swift
eleven RecordingState cases and the wall/video time keys are present
_capture_helper_bridge_contract
CaptureHelperClient.swift, HostEnvironment.swift
eleven helper commands, sorted-key config JSON, and the repo-bound script path are present
_recorder_store_capture_fsm
RecorderStore.swift
the start gate, start, pause/resume, and stop-to-review paths are present
_hotkey_audio_meter_contract
HotkeyMonitor.swift, AudioLevelMonitor.swift
the hotkey key and modifiers and the clamped meter return are present
_transcribe_payload_builder
DemoTakeTranscribe/main.swift
the WhisperKit decode config, the demo_take_transcript_v0 schema, the SRT builder, and the missing-audio guard are present
\n
The build witness runs swift build
\n
The one engine that runs code, _run_public_witness, executes swift build and returns only the exit code, byte counts, and a Build complete! flag, claiming compilation and nothing more. The build witness is the one engine that runs code. _cached_swiftpm_build_witness calls _run_public_witness, which runs swift build and returns only the exit code, the stdout and stderr byte counts, and whether the string Build complete! appeared. The build log body stays out of the record. The claim is that the copied sources compile as a SwiftPM target, nothing more.
\nDiagram of the mechanism (4 steps).Diagram source & refs
flowchart TD bundle["Copied public Swift bundle eight source files"] engines["Six engines build witness plus five source-token contracts"] negatives["Six negative cases delete one guard re-run that engine"] record["Result record exit codes, counts, booleans no source bodies or logs"] bundle --> engines bundle --> negatives engines --> record negatives --> record
\n
Negative cases
\n
Five mutations flip an engine to blocked
\n
Five negative cases share _mutated_source_negative, which copies the bundle, replaces one token, re-runs the matching engine, and expects it to flip to blocked with its guard flag false._semantic_runtime_exercises builds six negative results. Five share _mutated_source_negative: it copies the bundle to a temporary directory with _copy_public_bundle, replaces one token with _replace_copied_source_token, re-runs the matching engine, and expects the engine to flip to blocked with its guard flag now false. _observed_negative_case reads that flag per case.
\n
The five named source mutations
\n
The five mutations rewrite the helper path, remove the source blocker, drop a hotkey modifier, unclamp the meter, and delete the audio guard, and any check that still passes fails the case. The five source mutations are missing_helper_bridge (rewrites the demo_take_capture.py helper path), start_without_screen (removes the "Select at least one recording source" blocker), hotkey_wrong_modifier (drops .option from the modifiers), audio_meter_unclamped (replaces the clamp with a bare rms * 8), and transcribe_missing_audio (removes the audioNotFound guard). If a check still passes after its guard is gone, it was never testing the guard, and that case fails.
\n
The sixth case checks the build witness
\n
The sixth case, missing_swift_build_witness, runs swift build against an empty directory and expects a non-pass, since a success against nothing would mean the witness was asserted, not run. The sixth, missing_swift_build_witness, runs swift build against an empty temporary directory through _run_public_witness and expects a non-pass. A build that "succeeds" against nothing would mean the witness is asserted rather than run.
\n
Prior Art Grounding
\n
ScreenCaptureKit, FFmpeg, and WhisperKit lineage
\n
The app follows the desktop capture-console lineage anchored by Apple's ScreenCaptureKit, FFmpeg, and WhisperKit, while this component borrows only the contract shape and stays at source-copy and fixture checks. The app follows the common desktop capture-console lineage: OS capture APIs, a command-line media encoder, typed recording state, a global hotkey, a level meter, and a transcription handoff. The public anchors are Apple's ScreenCaptureKit for screen and audio streaming, FFmpeg for encoding, and WhisperKit for on-device speech recognition on Apple Silicon. This component borrows the contract shape and keeps the exercise at source-copy and fixture checks.
\n
Validation Result record Path
\n
Run from microcosm-substrate/:
\n
Output files and what a pass means
\n
The run writes result, board, and validation-result record JSON plus an sign-off file, and a pass means every engine returned pass, the six negative cases blocked, and no source body reached a record. The run writes batch7_demo_take_console_capsule_result.json, batch7_demo_take_console_capsule_board.json, and batch7_demo_take_console_capsule_validation_receipt.json under the output directory, plus the sign-off file. The validate-bundle action re-runs the same checks over exported_batch7_demo_take_console_capsule_bundle. A pass means every engine returned pass, the six negative cases blocked, and no source body reached a record.
\n
Scope boundary
\n
Scope limit
\n
The bounded strongest claim
\n
The strongest claim is that the copied sources compile as a SwiftPM target and the start gate, hotkey modifier, clamp, bridge contract, and transcribe guards are present, confirmed by six deletion tests. The strongest claim is bounded: the copied Demo Take Console sources compile as a SwiftPM target, and the recording-start gate, hotkey modifier, audio clamp, helper-bridge contract, and transcribe guards are present in that source, with six deletion tests confirming each check depends on its guard.
\n
What a green run refuses
\n
It launches nothing, grants no screen or microphone access, runs no FFmpeg or WhisperKit, and is neither launch-scope decision nor proof that every UI path is covered. It refuses the rest. It does not launch the app, grant screen or microphone access, start a recording session, run FFmpeg, or dispatch WhisperKit. It is not launch or publishing-scope decision and bounded evidence that every UI path is covered. A green fixture or bundle run is evidence for this source-copy replay only.
\n",
"components": [
{
"anchor": "component-batch7_demo_take_console_capsule",
"component_id": "batch7_demo_take_console_capsule",
"public_label": "Demo Take Console Source Bundle"
}
],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"organ": 1,
"paper_module": 1,
"principle": 4
},
"link_count": 12,
"links": [
{
"id": "batch7_demo_take_console_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Demo Take Console Source Bundle",
"relation": "explains",
"url": "component-batch7-demo-take-console-capsule.html"
},
{
"id": "mechanism.batch7_demo_take_console_capsule.validates_public_demo_take_console_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public demo take console bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch7-demo-take-console-capsule-validates-public-demo-take-console-capsule"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch6_unsurfaced_primitives_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 6 Unsurfaced Primitives Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[75:paper_module.batch6_unsurfaced_primitives_capsule].depends_on",
"url": "paper-module-batch6-unsurfaced-primitives-capsule.html"
},
{
"id": "paper_module.batch8_audio_level_rms_port",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Audio Level RMS Port",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[58:paper_module.batch8_audio_level_rms_port].depends_on",
"url": "paper-module-batch8-audio-level-rms-port.html"
}
],
"paper_module_id": "paper_module.batch7_demo_take_console_capsule",
"reader_packet": {
"abstract": "Set 7 Demo Take Console Bundle imports Swift capture-console source bodies and exercises SwiftPM build, recording-state, helper-bridge, recorder-store, hotkey/audio-meter, and transcribe-payload boundaries.",
"authority_ceiling": "Fixture-bound public Swift source-body import, copied-module digest and anchor evidence, deterministic local exercise evidence, SwiftPM build-witness evidence, and metadata-only result records only; no app launch authority, screen capture authority, microphone capture authority, recording-session export, FFmpeg execution, WhisperKit/model dispatch, source-file changes, publishing-scope decision, launch-scope decision, whole-system equivalence, complete UI coverage, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.batch7_demo_take_console_capsule.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-batch7-demo-take-console-capsule-diagram-1",
"html_url": "docs/paper-module-batch7-demo-take-console-capsule.html#paper-module-batch7-demo-take-console-capsule-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD bundle[\"Copied public Swift bundle eight source files\"] engines[\"Six engines build witness plus five source-token contracts\"] negatives[\"Six negative cases delete one guard re-run that engine\"] record[\"Result record exit codes, counts, booleans no source bodies or logs\"] bundle --> engines bundle --> negatives engines --> record negatives --> record",
"source_format": "mermaid",
"source_hash": "sha256:67dbebd36caa69235efe9ac19200dc1b0a13d92c6a213395e1b44778b14f13f7",
"source_line_end": 109,
"source_line_start": 99
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 2,
"links": [
{
"id": "batch7_demo_take_console_capsule",
"kind": "organ",
"kind_label": "component",
"label": "Demo Take Console Source Bundle",
"relation": "explains",
"url": "component-batch7-demo-take-console-capsule.html"
},
{
"id": "mechanism.batch7_demo_take_console_capsule.validates_public_demo_take_console_capsule",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public demo take console bundle",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-batch7-demo-take-console-capsule-validates-public-demo-take-console-capsule"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.batch7_demo_take_console_capsule",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 211,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch7_demo_take_console_capsule.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "ScreenCaptureKit",
"section_id": "screencapturekit-ffmpeg-and-whisperkit-lineage",
"url": "https://developer.apple.com/documentation/ScreenCaptureKit"
},
{
"kind": "external_reference",
"label": "FFmpeg",
"section_id": "screencapturekit-ffmpeg-and-whisperkit-lineage",
"url": "https://www.ffmpeg.org/ffmpeg.html"
},
{
"kind": "external_reference",
"label": "WhisperKit",
"section_id": "screencapturekit-ffmpeg-and-whisperkit-lineage",
"url": "https://github.com/argmaxinc/WhisperKit"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component reads the Demo Take Console's public Swift source and checks that its safety gates actually hold, without ever launching the macOS capture app.** The Demo Take Console is a macOS app that records screen and microphone takes, drives FFmpeg through a Python helper, and hands the audio to an on-device WhisperKit transcriber. This component reads the app's public Swift source and checks that the safety gates in that source actually hold, without ever launching the app.",
"level": 4,
"order": 1,
"section_id": "the-console-capsule-replays-one-copied-runtime",
"source_line_end": 13,
"source_line_start": 3,
"title": "The console bundle replays one copied runtime"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It copies eight Swift files, runs six source-token checks plus six paired deletion cases, and emits records of exit codes, counts, and booleans, never source bodies or logs.** It copies eight Swift files into a public bundle, runs six checks that read those copies, then runs six paired negative cases that delete one guard and demand the matching check notice. It emits result records that carry exit codes, counts, and booleans, never the source bodies, recordings, or build logs.",
"level": 4,
"order": 2,
"section_id": "three-checks-bind-inputs-output-and-ceiling",
"source_line_end": 24,
"source_line_start": 14,
"title": "Three checks bind inputs, output, and ceiling"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A green run says only that the source contracts are present, not that capture, FFmpeg, or a model ever ran.** A green run says the source contracts are present. It does not start capture, run FFmpeg, or dispatch a model.",
"level": 4,
"order": 3,
"section_id": "the-demo-result-stays-body-free-and-local",
"source_line_end": 31,
"source_line_start": 25,
"title": "The demo result stays metadata-only and local"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 33,
"source_line_start": 32,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A capture app is trusted with screen, microphone, and disk, so the narrow question is whether the gates written into its Swift source exist as source.** A capture app is trusted with a screen, a microphone, and disk. The question this component answers is narrow: before anyone trusts the app to record, do the gates written into its Swift source exist as source?",
"level": 4,
"order": 5,
"section_id": "do-the-source-gates-exist-as-source",
"source_line_end": 41,
"source_line_start": 34,
"title": "Do the source gates exist as source"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Because a file can name clamp yet ship an unclamped meter, each check asserts a specific behaviour-carrying token in the copied source rather than shape alone.** Checking shape alone is not enough. A file can contain the word \"clamp\" and still ship an unclamped meter. So each check asserts a specific token that carries the behaviour: the start gate requires a chosen recording source and free disk, the global hotkey requires Control-Option-Command, the audio meter returns min(max(rms * 8, 0), 1), the transcriber guards a missing audio file. The checks read the copied source, not the app's live state.",
"level": 4,
"order": 6,
"section_id": "each-check-asserts-a-behaviour-carrying-token",
"source_line_end": 53,
"source_line_start": 42,
"title": "Each check asserts a behaviour-carrying token"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 55,
"source_line_start": 54,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "run and runbatch7demotakebundle both drive runcrownjewelorgan with evaluate for positive checks and evaluatenegative_case for mutation tests, over the fixture or the exported copy.** run and runbatch7demotakebundle both call runcrownjewel_organ with two callbacks: evaluate for the positive checks and evaluatenegative_case for the mutation tests. run reads the fixture input; the bundle variant reads the exported public copy. main wires the run and validate-bundle subcommands to the same runner and prints a result or a card.",
"level": 4,
"order": 8,
"section_id": "two-entrypoints-and-two-callbacks",
"source_line_end": 66,
"source_line_start": 56,
"title": "Two entrypoints and two callbacks"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "**_evaluate runs six engines, failing on any non-pass status or fewer than eight copied modules, with each engine reading its Swift file through _read and returning booleans plus a claimceiling. _evaluate runs six engines in order and collects a finding for any that does not return status: \"pass\". It also fails if the copied source manifest holds fewer than eight modules, the count in SOURCEREQUIREDANCHORS. Each engine reads its Swift file through _read, which resolves the copied path under the exported bundle, and returns a small dict of booleans plus a claim_ceiling string.",
"level": 4,
"order": 9,
"section_id": "six-engines-over-eight-copied-modules",
"source_line_end": 87,
"source_line_start": 67,
"title": "Six engines over eight copied modules"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "The one engine that runs code, runpublicwitness, executes swift build and returns only the exit code, byte counts, and a Build complete! flag, claiming compilation and nothing more.** The build witness is the one engine that runs code. cachedswiftpmbuildwitness calls runpublic_witness, which runs swift build and returns only the exit code, the stdout and stderr byte counts, and whether the string Build complete! appeared. The build log body stays out of the record. The claim is that the copied sources compile as a SwiftPM target, nothing more.",
"level": 4,
"order": 10,
"section_id": "the-build-witness-runs-swift-build",
"source_line_end": 110,
"source_line_start": 88,
"title": "The build witness runs `swift build`"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 11,
"section_id": "negative-cases",
"source_line_end": 112,
"source_line_start": 111,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Five negative cases share mutatedsourcenegative, which copies the bundle, replaces one token, re-runs the matching engine, and expects it to flip to blocked with its guard flag false.** semanticruntime_exercises builds six negative results. Five share mutatedsource_negative: it copies the bundle to a temporary directory with copypublicbundle, replaces one token with replacecopiedsource_token, re-runs the matching engine, and expects the engine to flip to blocked with its guard flag now false. observednegative_case reads that flag per case.",
"level": 4,
"order": 12,
"section_id": "five-mutations-flip-an-engine-to-blocked",
"source_line_end": 123,
"source_line_start": 113,
"title": "Five mutations flip an engine to blocked"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The five mutations rewrite the helper path, remove the source blocker, drop a hotkey modifier, unclamp the meter, and delete the audio guard, and any check that still passes fails the case.** The five source mutations are missinghelperbridge (rewrites the demotakecapture.py helper path), startwithoutscreen (removes the \"Select at least one recording source\" blocker), hotkeywrongmodifier (drops .option from the modifiers), audiometerunclamped (replaces the clamp with a bare rms 8), and transcribemissing_audio (removes the audioNotFound guard). If a check still passes after its guard is gone, it was never testing the guard, and that case fails.",
"level": 4,
"order": 13,
"section_id": "the-five-named-source-mutations",
"source_line_end": 136,
"source_line_start": 124,
"title": "The five named source mutations"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The sixth case, missingswiftbuildwitness, runs swift build against an empty directory and expects a non-pass, since a success against nothing would mean the witness was asserted, not run.** The sixth, missingswiftbuild_witness, runs swift build against an empty temporary directory through runpublic_witness and expects a non-pass. A build that \"succeeds\" against nothing would mean the witness is asserted rather than run.",
"level": 4,
"order": 14,
"section_id": "the-sixth-case-checks-the-build-witness",
"source_line_end": 146,
"source_line_start": 137,
"title": "The sixth case checks the build witness"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "prior-art-grounding",
"source_line_end": 148,
"source_line_start": 147,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The app follows the desktop capture-console lineage anchored by Apple's ScreenCaptureKit, FFmpeg, and WhisperKit, while this component borrows only the contract shape and stays at source-copy and fixture checks.** The app follows the common desktop capture-console lineage: OS capture APIs, a command-line media encoder, typed recording state, a global hotkey, a level meter, and a transcription handoff. The public anchors are Apple's ScreenCaptureKit for screen and audio streaming, FFmpeg for encoding, and WhisperKit for on-device speech recognition on Apple Silicon. This component borrows the contract shape and keeps the exercise at source-copy and fixture checks.",
"level": 4,
"order": 16,
"section_id": "screencapturekit-ffmpeg-and-whisperkit-lineage",
"source_line_end": 162,
"source_line_start": 149,
"title": "ScreenCaptureKit, FFmpeg, and WhisperKit lineage"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run from microcosm-substrate/:",
"level": 2,
"order": 17,
"section_id": "validation-receipt-path",
"source_line_end": 174,
"source_line_start": 163,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The run writes result, board, and validation-result record JSON plus an sign-off file, and a pass means every engine returned pass, the six negative cases blocked, and no source body reached a record.** The run writes batch7demotakeconsolecapsule_result.json, batch7demotakeconsolecapsule_board.json, and batch7demotakeconsolecapsulevalidationreceipt.json under the output directory, plus the sign-off file. The validate-bundle action re-runs the same checks over exportedbatch7demotakeconsolecapsulebundle. A pass means every engine returned pass, the six negative cases blocked, and no source body reached a record.",
"level": 4,
"order": 18,
"section_id": "output-files-and-what-a-pass-means",
"source_line_end": 187,
"source_line_start": 175,
"title": "Output files and what a pass means"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "claim-ceiling",
"source_line_end": 189,
"source_line_start": 188,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is that the copied sources compile as a SwiftPM target and the start gate, hotkey modifier, clamp, bridge contract, and transcribe guards are present, confirmed by six deletion tests.** The strongest claim is bounded: the copied Demo Take Console sources compile as a SwiftPM target, and the recording-start gate, hotkey modifier, audio clamp, helper-bridge contract, and transcribe guards are present in that source, with six deletion tests confirming each check depends on its guard.",
"level": 4,
"order": 20,
"section_id": "the-bounded-strongest-claim",
"source_line_end": 199,
"source_line_start": 190,
"title": "The bounded strongest claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It launches nothing, grants no screen or microphone access, runs no FFmpeg or WhisperKit, and is neither launch-scope decision nor proof that every UI path is covered.** It refuses the rest. It does not launch the app, grant screen or microphone access, start a recording session, run FFmpeg, or dispatch WhisperKit. It is not launch or public sharing decision and bounded evidence that every UI path is covered. A green fixture or bundle run is evidence for this source-copy replay only.",
"level": 4,
"order": 21,
"section_id": "what-a-green-run-refuses",
"source_line_end": 210,
"source_line_start": 200,
"title": "What a green run refuses"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/batch7_demo_take_console_capsule.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Set 7 Demo Take Console Bundle imports Swift capture-console source bodies and exercises SwiftPM build, recording-state, helper-bridge, recorder-store, hotkey/audio-meter, and transcribe-payload boundaries."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "batch7_demo_take_console_capsule.py",
"path": "src/microcosm_core/organs/batch7_demo_take_console_capsule.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for Set 7 Demo Take Console public Swift source imports, deterministic exercises, exported-bundle checks, negative cases, and scope limits.",
"symbols": [
"_swiftpm_build_witness",
"_recording_state_control_model",
"_capture_helper_bridge_contract",
"_recorder_store_capture_fsm",
"_hotkey_audio_meter_contract",
"_transcribe_payload_builder",
"_evaluate",
"run"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/organs/batch7_demo_take_console_capsule.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch7_demo_take_console_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/batch7_demo_take_console_capsule.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Set 7 Demo Take Console Bundle binds the runnable Demo Take Console source locus, exported copied Swift source bundle, source-module digests, SwiftPM build-witness posture, deterministic recording-state/helper-bridge/recorder-store/hotkey-audio-meter/transcribe-payload exercises, negative cases, metadata-only result records, and scope limits to a mechanism-backed JSON bundle without claiming accepted-component authority, app launch authority, or recording authority.",
"summary_line": "Set 7 Demo Take Console Bundle imports Swift capture-console source bodies and exercises SwiftPM build, recording-state, helper-bridge, recorder-store, hotkey/audio-meter, and transcribe-payload boundaries.",
"title": "Set 7 Demo Take Console Bundle"
},
{
"anchor": "paper-module-engine-room-generated-projection-drift-gate",
"authority_ceiling": "Public owner-routed generated projection drift fixture and focused regression result records only; no semantic drift proof, full source registry validation, repair authority, launch-scope decision, whole-system equivalence, source-file changes, external model access, or whole-system correctness.",
"body_html": "
Generated files committed beside their sources can silently go stale, so given changed paths this gate reports which declared surfaces may now be out of date and whether each one's freshness check still passes. A repository that commits generated files next to the sources they were built from carries a standing risk: a generated file can quietly fall out of step with its source, and nothing fails until a reader trusts a stale copy. This component checks for that. Given a set of changed paths, it answers which declared generated surfaces might now be out of date, and whether each surface's own freshness check still passes.
\n
Module is source-faithful
\n
The module is a source-faithful refactor of two internal builders that emits one result record per surface: the fingerprints taken, whether a clean record was reused or the check ran, and any drift reasons. The public module is a source-faithful refactor of two internal builders, kept together in src/microcosm_core/engine_room/generated_projection_drift_gate.py. It emits one result record per checked surface: the fingerprints it took, whether it reused a prior clean record or ran the check, and, if the surface drifted, the exact reasons why.
\n
Owner routing is the design choice worth naming: each generated surface is a row that names its own check, so a changed path selects only the rows it could affect rather than triggering one global diff. The gate does not diff every generated file in one global sweep. Each surface is a row that names its own check, and a changed path selects only the rows it could affect.
\n
Purpose
\n
Why a global snapshot fails
\n
A single global-snapshot alternative fails in two directions: it reruns everything on every edit, and it cannot say who owns a difference or how to fix it. The loose alternative is a single global snapshot: hash the whole generated tree, compare against a stored snapshot, flag any difference. That fails in two directions. It reruns everything on every edit, and it cannot say who owns a difference or how to fix it.
\n
Splitting the tree into owners
\n
Each surface becomes a ProjectionOwner naming its artifacts, sources, and no-write check command, and that command, not the gate, decides freshness while the gate routes and records evidence. This component splits the tree into owners instead. A ProjectionOwner names its generated artifacts, its source authorities, and a no-write check command that is treated as the freshness authority for that surface. A changed path is matched against those patterns, so a small edit selects only the owners it could plausibly touch. The check command, not this gate, decides whether a surface is fresh. The gate routes to the right owner and records the evidence.
\n
How it works
\n
Runtime is pipeline
\n
The runtime is a pipeline. The runtime is a pipeline. It runs in this order for one check.
\n
Scope selects owners by path
\n
select_projection_owners keeps only owners whose artifact or source patterns match a changed path, returning the selected rows and a selection mode of scoped_paths, owner_filter, or all_owners.select_projection_owners takes the full owner list plus optional owner ids and changed paths. When changed paths are given, it keeps only owners whose artifact or source patterns match one of those paths, through owner_matches_path and projection_pattern_matches_path (exact, prefix, or glob match). It returns the selected rows and a selection record whose mode is scoped_paths, owner_filter, or all_owners.
\n
Fingerprint hashes sources and artifacts
\n
_owner_fingerprint content-addresses every source and artifact file with sha256, recording a combined source_hash, artifact_hash, path counts, and any pattern that matches no file as missing. For each selected owner, _owner_fingerprint content-addresses every source file and every artifact file with sha256. It records a combined source_hash, a combined artifact_hash, path counts, and any missing patterns. A pattern that matches no file on disk is counted as missing.
\n
Cache reuse is deliberately strict
\n
_source_hash_cache_hit reuses a prior clean record only when its status, source_hash, artifact_hash, and check command all match and no artifacts are missing, otherwise the check runs._source_hash_cache_hit decides whether a prior clean record can be reused. Reuse is deliberately strict. It requires a prior record with status clean, the same source_hash, the same artifact_hash, the same check command, and zero missing artifacts. If any of those differ, the cache misses and the check runs.
\n
Run the check or reuse cache
\n
A cache hit returns a synthetic source_hash_cache record, while a miss runs the owner's check command and captures its return code plus hashed, length-counted stdout and stderr without storing raw text. On a cache hit, _cached_clean_result returns a synthetic record marked source_hash_cache, returncode zero. On a miss, _run_command runs the owner's check command and captures its return code plus hashed, length-counted stdout and stderr (the raw text is not stored). The command runs through _run_builtin_command first, which handles the fixture-only builtin:pass, builtin:fail, and builtin:assert-file-equals verbs before falling through to a real subprocess.
\n
Classify drift reasons per owner
\n
_check_owner assembles drift reasons from a failing command, a missing artifact, or an incomplete lineage record, and any reason marks the owner drift while none marks it clean._check_owner assembles the drift reasons. A nonzero return code adds check_command_failed. A missing artifact adds artifact_missing, independent of the command result, so an absent generated file cannot be laundered by a green command. A required-but-incomplete lineage record adds fact_authority_lineage_missing_required or fact_authority_lineage_invalid. Any reason makes the owner status drift; none makes it clean.
\n
Aggregate the per-owner results
\n
check_projection_drift runs the per-owner check over the selected rows and rolls them into an overall status, drift_owner_count, and source_hash_cache hit and miss counts. This is the record a caller reads. The residual explanation remains available so the reader can inspect the concrete boundary without reconstructing omitted context.
\n
Function
Role
select_projection_owners
Scopes the owner set by owner id or changed path
check_projection_drift
Runs the per-owner checks and aggregates the result
evaluate_case
Materialises one fixture case in a scratch tree and checks it
evaluate_fixture_dir
Runs every fixture case in a directory and counts passes
main
Command-line entry: check a real tree or evaluate-fixtures
\n
The two CLI subcommands
\n
The check subcommand runs the gate against a live tree and exits nonzero on drift, while evaluate-fixtures materialises each bundled case in a temporary tree and checks observed against expected status. The check subcommand runs the gate against a live tree and exits nonzero when any owner drifted. The evaluate-fixtures subcommand runs the bundled cases; evaluate_case writes each case's declared files into a temporary tree, builds owners from the case JSON, and records whether the observed status matched the expected one.
\nDiagram of the mechanism (10 steps).Diagram source & refs
Source refs
Select projection owners
select_projection_owners
flowchart TD A["Changed path or owner id"] --> B["select_projection_owners"] B --> C["_owner_fingerprint: sha256 sources and artifacts"] C --> D{"_source_hash_cache_hit matches prior clean record?"} D -- "yes" --> E["reuse cached clean record"] D -- "no" --> F["_run_command runs the owner check"] E --> G["_check_owner assembles drift reasons"] F --> G G --> H{"any reason?"} H -- "yes" --> I["status drift with reasons"] H -- "no" --> J["status clean"]
\n
Negative cases
\n
Two positive and three drift cases
\n
Two positive fixtures and three drift fixtures cover a planted byte, a deleted artifact, and a missing lineage record, and a run over all five returns status: pass with every case passed. The fixture set has two positive cases, clean_owner and scoped_changed_path, and three that must be caught as drift. planted_byte_detected changes one byte in a generated file so the owner's builtin:assert-file-equals check returns nonzero, giving check_command_failed. missing_artifact deletes a declared artifact so the fingerprint reports a missing path, giving artifact_missing even though the check would otherwise pass. fact_authority_lineage_required_missing declares an owner that requires a lineage record but omits its guard, derivation, and residual fields, giving fact_authority_lineage_missing_required. A run over all five cases returns status: pass, case_count: 5, passed_case_count: 5.
\n
Prior Art Grounding
\n
Reproducible-build and regression-test lineage
\n
The gate follows reproducible-build and regression-test practice: declare source inputs, produce artifacts, compare content hashes, and rerun the owner check when source or artifact identity changes. The gate follows reproducible-build and regression-test practice: declare source inputs, produce generated artifacts, compare content hashes, and rerun the owner check when either source or artifact identity changes. Two anchors are relevant.
\n
Bazel hermeticity, for its emphasis on declared inputs, source identity, repeatable actions, and cache validity.
The borrowed part is declared-input and artifact-fingerprint discipline; the added part is routing each check through its owner rather than treating all generated files as one snapshot. The borrowed part is declared-input and artifact-fingerprint discipline. The added part is routing each check through its owner rather than treating all generated files as one snapshot.
A pass means the fixture behavior and the bundle-backed JSON projection are reproducible; it runs no repair commands and excludes launch. A pass means the fixture behavior and the bundle-backed JSON projection are reproducible. It does not run repair commands and it excludes launch.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
For a supplied tree and declared owners, the gate reports whether each selected owner's required artifacts were present and its no-write check passed, with fingerprints recorded and clean records reused only under matching hashes. The strongest claim the evidence supports: for a supplied tree and a set of declared owners, the gate reports whether each selected owner's required artifacts were present and whether its own no-write check passed, with source and artifact fingerprints recorded and clean records reused only under matching hashes. That is a freshness signal for declared owners.
\n
What the gate does not establish
\n
The proof boundary stops at freshness: the gate is not semantic-drift proof, not full registry validation, and carries no repair or launch-scope decision over the wider generated tree. The proof boundary stops there. It is not semantic drift proof, so a check that passes on identical bytes says nothing about whether the generated prose is meaningfully correct. It is not full registry validation: only declared owners are checked, and the wider generated tree is judged by each owner's own command. It has no repair authority and no launch-scope decision. The scope limit excludes any claim that every generated surface in the wider system is fresh.
\n",
"components": [],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"paper_module": 1,
"principle": 4
},
"link_count": 11,
"links": [
{
"id": "mechanism.engine_room_generated_projection_drift_gate.validates_public_generated_projection_drift_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public generated projection drift gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-generated-projection-drift-gate-validates-public-generated-projection-drift-gate"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch8_validator_checker_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Validator Checker Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[63:paper_module.batch8_validator_checker_capsule].depends_on",
"url": "paper-module-batch8-validator-checker-capsule.html"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[31:paper_module.engine_room_demo].depends_on",
"url": "paper-module-engine-room-demo.html"
}
],
"paper_module_id": "paper_module.engine_room_generated_projection_drift_gate",
"reader_packet": {
"abstract": "Public generated-projection drift fixture: owner-routed checks fingerprint declared sources and artifacts, reuse clean result records only under matching hashes, and fail planted-byte or missing-artifact cases.",
"authority_ceiling": "Public owner-routed generated projection drift fixture and focused regression result records only; no semantic drift proof, full source registry validation, repair authority, launch-scope decision, whole-system equivalence, source-file changes, external model access, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.engine_room_generated_projection_drift_gate.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-engine-room-generated-projection-drift-gate-diagram-1",
"html_url": "docs/paper-module-engine-room-generated-projection-drift-gate.html#paper-module-engine-room-generated-projection-drift-gate-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Changed path or owner id\"] --> B[\"select_projection_owners\"] B --> C[\"_owner_fingerprint: sha256 sources and artifacts\"] C --> D{\"_source_hash_cache_hit matches prior clean record?\"} D -- \"yes\" --> E[\"reuse cached clean record\"] D -- \"no\" --> F[\"_run_command runs the owner check\"] E --> G[\"_check_owner assembles drift reasons\"] F --> G G --> H{\"any reason?\"} H -- \"yes\" --> I[\"status drift with reasons\"] H -- \"no\" --> J[\"status clean\"]",
"source_format": "mermaid",
"source_hash": "sha256:f90bb3e9a92f457d4f99c3a2336f7a06bc8750f46b173c561279f771ade6c396",
"source_line_end": 157,
"source_line_start": 145
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.macro_projection_import_protocol",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Source Projection Import Protocol",
"relation": "depends on",
"url": "paper-module-macro-projection-import-protocol.html"
}
],
"relation": "depends on"
},
{
"count": 1,
"links": [
{
"id": "mechanism.engine_room_generated_projection_drift_gate.validates_public_generated_projection_drift_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public generated projection drift gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-generated-projection-drift-gate-validates-public-generated-projection-drift-gate"
}
],
"relation": "explains"
},
{
"count": 5,
"links": [
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.engine_room_generated_projection_drift_gate",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 243,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_generated_projection_drift_gate.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Bazel hermeticity",
"section_id": "reproducible-build-and-regression-test-lineage",
"url": "https://bazel.build/concepts/hermeticity"
},
{
"kind": "external_reference",
"label": "pytest-regtest snapshot testing",
"section_id": "reproducible-build-and-regression-test-lineage",
"url": "https://pytest-regtest.readthedocs.io/en/stable/snapshots/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Generated files committed beside their sources can silently go stale, so given changed paths this gate reports which declared surfaces may now be out of date and whether each one's freshness check still passes.** A repository that commits generated files next to the sources they were built from carries a standing risk: a generated file can quietly fall out of step with its source, and nothing fails until a reader trusts a stale copy. This component checks for that. Given a set of changed paths, it answers which declared generated surfaces might now be out of date, and whether each surface's own freshness check still passes.",
"level": 4,
"order": 1,
"section_id": "generated-files-committed",
"source_line_end": 14,
"source_line_start": 3,
"title": "Generated files committed"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The module is a source-faithful refactor of two internal builders that emits one result record per surface: the fingerprints taken, whether a clean record was reused or the check ran, and any drift reasons.** The public module is a source-faithful refactor of two internal builders, kept together in src/microcosmcore/engineroom/generatedprojectiondrift_gate.py. It emits one result record per checked surface: the fingerprints it took, whether it reused a prior clean record or ran the check, and, if the surface drifted, the exact reasons why.",
"level": 4,
"order": 2,
"section_id": "module-is-source-faithful",
"source_line_end": 25,
"source_line_start": 15,
"title": "Module is source-faithful"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Owner routing is the design choice worth naming: each generated surface is a row that names its own check, so a changed path selects only the rows it could affect rather than triggering one global diff.** The gate does not diff every generated file in one global sweep. Each surface is a row that names its own check, and a changed path selects only the rows it could affect.",
"level": 4,
"order": 3,
"section_id": "owner-routing-is",
"source_line_end": 33,
"source_line_start": 26,
"title": "Owner routing is"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 35,
"source_line_start": 34,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A single global-snapshot alternative fails in two directions: it reruns everything on every edit, and it cannot say who owns a difference or how to fix it.** The loose alternative is a single global snapshot: hash the whole generated tree, compare against a stored snapshot, flag any difference. That fails in two directions. It reruns everything on every edit, and it cannot say who owns a difference or how to fix it. A one-line change to a doc source forces",
"level": 4,
"order": 5,
"section_id": "why-a-global-snapshot-fails",
"source_line_end": 45,
"source_line_start": 36,
"title": "Why a global snapshot fails"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Each surface becomes a ProjectionOwner naming its artifacts, sources, and no-write check command, and that command, not the gate, decides freshness while the gate routes and records evidence.** This component splits the tree into owners instead. A ProjectionOwner names its generated artifacts, its source authorities, and a no-write check command that is treated as the freshness authority for that surface. A changed path is matched against those patterns, so a small edit selects only the owners it could plausibly touch. The check command, not this gate, decides whether a surface is fresh. The gate routes to the right owner and records the evidence.",
"level": 4,
"order": 6,
"section_id": "splitting-the-tree-into-owners",
"source_line_end": 57,
"source_line_start": 46,
"title": "Splitting the tree into owners"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 59,
"source_line_start": 58,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The runtime is a pipeline. The runtime is a pipeline. It runs in this order for one check.",
"level": 4,
"order": 8,
"section_id": "runtime-is-pipeline",
"source_line_end": 64,
"source_line_start": 60,
"title": "Runtime is pipeline"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "selectprojection_owners keeps only owners whose artifact or source patterns match a changed path, returning the selected rows and a selection mode of scopedpaths, ownerfilter, or allowners. selectprojectionowners takes the full owner list plus optional owner ids and changed paths. When changed paths are given, it keeps only owners whose artifact or source patterns match one of those paths, through ownermatchespath and projectionpatternmatches_path (exact, prefix, or glob match). It returns the selected rows and a selection record whose mode is scoped_paths, ownerfilter, or allowners.",
"level": 4,
"order": 9,
"section_id": "scope-selects-owners-by-path",
"source_line_end": 76,
"source_line_start": 65,
"title": "Scope selects owners by path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "ownerfingerprint content-addresses every source and artifact file with sha256, recording a combined sourcehash, artifacthash, path counts, and any pattern that matches no file as missing.** For each selected owner, ownerfingerprint content-addresses every source file and every artifact file with sha256. It records a combined sourcehash, a combined artifacthash, path counts, and any missing patterns. A pattern that matches no file on disk is counted as missing.",
"level": 4,
"order": 10,
"section_id": "fingerprint-hashes-sources-and-artifacts",
"source_line_end": 86,
"source_line_start": 77,
"title": "Fingerprint hashes sources and artifacts"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "sourcehashcachehit reuses a prior clean record only when its status, sourcehash, artifacthash, and check command all match and no artifacts are missing, otherwise the check runs. sourcehashcache_hit decides whether a prior clean record can be reused. Reuse is deliberately strict. It requires a prior record with status clean, the same source_hash, the same artifact_hash, the same check command, and zero missing artifacts. If any of those differ, the cache misses and the check runs.",
"level": 4,
"order": 11,
"section_id": "cache-reuse-is-deliberately-strict",
"source_line_end": 96,
"source_line_start": 87,
"title": "Cache reuse is deliberately strict"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A cache hit returns a synthetic sourcehash_cache record, while a miss runs the owner's check command and captures its return code plus hashed, length-counted stdout and stderr without storing raw text.** On a cache hit, cachedcleanresult returns a synthetic record marked sourcehash_cache, returncode zero. On a miss, runcommand runs the owner's check command and captures its return code plus hashed, length-counted stdout and stderr (the raw text is not stored). The command runs through runbuiltin_command first, which handles the fixture-only builtin:pass, builtin:fail, and builtin:assert-file-equals verbs before falling through to a real subprocess.",
"level": 4,
"order": 12,
"section_id": "run-the-check-or-reuse-cache",
"source_line_end": 108,
"source_line_start": 97,
"title": "Run the check or reuse cache"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "checkowner assembles drift reasons from a failing command, a missing artifact, or an incomplete lineage record, and any reason marks the owner drift while none marks it clean. check_owner assembles the drift reasons. A nonzero return code adds checkcommandfailed. A missing artifact adds artifact_missing, independent of the command result, so an absent generated file cannot be laundered by a green command. A required-but-incomplete lineage record adds factauthoritylineagemissingrequired or factauthoritylineage_invalid. Any reason makes the owner status drift; none makes it clean.",
"level": 4,
"order": 13,
"section_id": "classify-drift-reasons-per-owner",
"source_line_end": 120,
"source_line_start": 109,
"title": "Classify drift reasons per owner"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "checkprojection_drift runs the per-owner check over the selected rows and rolls them into an overall status, driftownercount, and sourcehashcache hit and miss counts.** This is the record a caller reads. The residual explanation remains available so the reader can inspect the concrete boundary without reconstructing omitted context.",
"level": 4,
"order": 14,
"section_id": "aggregate-the-per-owner-results",
"source_line_end": 134,
"source_line_start": 121,
"title": "Aggregate the per-owner results"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "**The check subcommand runs the gate against a live tree and exits nonzero on drift, while evaluate-fixtures materialises each bundled case in a temporary tree and checks observed against expected status.** The check subcommand runs the gate against a live tree and exits nonzero when any owner drifted. The evaluate-fixtures subcommand runs the bundled cases; evaluate_case writes each case's declared files into a temporary tree, builds owners from the case JSON, and records whether the observed status matched the expected one.",
"level": 4,
"order": 15,
"section_id": "the-two-cli-subcommands",
"source_line_end": 158,
"source_line_start": 135,
"title": "The two CLI subcommands"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "negative-cases",
"source_line_end": 160,
"source_line_start": 159,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Two positive fixtures and three drift fixtures cover a planted byte, a deleted artifact, and a missing lineage record, and a run over all five returns status: pass with every case passed.** The fixture set has two positive cases, cleanowner and scopedchanged_path, and three that must be caught as drift. plantedbytedetected changes one byte in a generated file so the owner's builtin:assert-file-equals check returns nonzero, giving checkcommandfailed. missing_artifact deletes a declared artifact so the fingerprint reports a missing path, giving artifact_missing even though the check would otherwise pass. factauthoritylineagerequiredmissing declares an owner that requires a lineage record bu\u2026",
"level": 4,
"order": 17,
"section_id": "two-positive-and-three-drift-cases",
"source_line_end": 175,
"source_line_start": 161,
"title": "Two positive and three drift cases"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "prior-art-grounding",
"source_line_end": 177,
"source_line_start": 176,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**The gate follows reproducible-build and regression-test practice: declare source inputs, produce artifacts, compare content hashes, and rerun the owner check when source or artifact identity changes.** The gate follows reproducible-build and regression-test practice: declare source inputs, produce generated artifacts, compare content hashes, and rerun the owner check when either source or artifact identity changes. Two anchors are relevant. Bazel hermeticity, for its emphasis on declared inputs, source identity, repeatable actions, and cache validity. pytest-regtest snapshot testing, where recorded outputs are compared against reference outputs to catch unexpected change.",
"level": 4,
"order": 19,
"section_id": "reproducible-build-and-regression-test-lineage",
"source_line_end": 189,
"source_line_start": 178,
"title": "Reproducible-build and regression-test lineage"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The borrowed part is declared-input and artifact-fingerprint discipline; the added part is routing each check through its owner rather than treating all generated files as one snapshot.** The borrowed part is declared-input and artifact-fingerprint discipline. The added part is routing each check through its owner rather than treating all generated files as one snapshot.",
"level": 4,
"order": 20,
"section_id": "borrowed-discipline-added-owner-routing",
"source_line_end": 197,
"source_line_start": 190,
"title": "Borrowed discipline, added owner routing"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "The public exercise runs the same fixtures directly:",
"level": 2,
"order": 21,
"section_id": "validation-receipt-path",
"source_line_end": 210,
"source_line_start": 198,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the fixture behavior and the bundle-backed JSON projection are reproducible; it runs no repair commands and excludes launch.** A pass means the fixture behavior and the bundle-backed JSON projection are reproducible. It does not run repair commands and it excludes launch.",
"level": 4,
"order": 22,
"section_id": "what-a-pass-means-here",
"source_line_end": 217,
"source_line_start": 211,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 23,
"section_id": "claim-ceiling",
"source_line_end": 219,
"source_line_start": 218,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**For a supplied tree and declared owners, the gate reports whether each selected owner's required artifacts were present and its no-write check passed, with fingerprints recorded and clean records reused only under matching hashes.** The strongest claim the evidence supports: for a supplied tree and a set of declared owners, the gate reports whether each selected owner's required artifacts were present and whether its own no-write check passed, with source and artifact fingerprints recorded and clean records reused only under matching hashes. That is a freshness signal for declared owners.",
"level": 4,
"order": 24,
"section_id": "the-strongest-supported-claim",
"source_line_end": 230,
"source_line_start": 220,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The proof boundary stops at freshness: the gate is not semantic-drift proof, not full registry validation, and carries no repair or launch-scope decision over the wider generated tree.** The proof boundary stops there. It is not semantic drift proof, so a check that passes on identical bytes says nothing about whether the generated prose is meaningfully correct. It is not full registry validation: only declared owners are checked, and the wider generated tree is judged by each owner's own command. It has no repair authority and no launch-scope decision. The scope limit excludes any claim that every generated surface in the wider system is fresh.",
"level": 4,
"order": 25,
"section_id": "what-the-gate-does-not-prove",
"source_line_end": 242,
"source_line_start": 231,
"title": "What the gate does not establish"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/engine_room_generated_projection_drift_gate.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Public generated-projection drift fixture: owner-routed checks fingerprint declared sources and artifacts, reuse clean result records only under matching hashes, and fail planted-byte or missing-artifact cases."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "generated_projection_drift_gate.py",
"path": "src/microcosm_core/engine_room/generated_projection_drift_gate.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for owner-routed generated projection drift checks, source/artifact fingerprint result records, source-hash cache decisions, no-write check command result records, fixture evaluation, and scope limits.",
"symbols": [
"ProjectionOwner",
"select_projection_owners",
"check_projection_drift",
"evaluate_fixture_dir",
"evaluate_case",
"main"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/generated_projection_drift_gate.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_generated_projection_drift_gate.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_generated_projection_drift_gate.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Engine Room Generated Projection Drift Gate is a generated-artifact freshness bundle. It validates projection owner selection from changed paths, declared source and artifact fingerprints, no-write check return codes, source-hash cache reuse, planted-byte detection, and missing-artifact failure over four public fixtures while keeping semantic drift proof, repair authority, full source registry validation, launch, and private-system claims out of scope.",
"summary_line": "Public generated-projection drift fixture: owner-routed checks fingerprint declared sources and artifacts, reuse clean result records only under matching hashes, and fail planted-byte or missing-artifact cases.",
"title": "Engine Room Generated Projection Drift Gate"
},
{
"anchor": "paper-module-engine-room-command-run-singleflight",
"authority_ceiling": "Public subprocess singleflight fixture and focused regression result records only; no job scheduler, daemon, distributed lock service, live command_runs export, external model access authority, launch-scope decision, whole-system equivalence, source-file changes, or whole-system correctness.",
"body_html": "
When two callers fire the same command at once, this component runs it exactly once and hands the second caller the first caller's captured output. When several agents or background tasks fire the same command at the same moment, the plain outcome is several identical subprocesses doing the same work at once. That wastes the machine, and where the command writes something it can corrupt shared state by writing twice. This component answers one question: when two callers ask for the same command at the same time, can the system run it exactly once and hand the second caller the first caller's captured output?
\n
Is public refactor
\n
It is a public refactor keeping the executable core: content-addressed keys, an fcntl leader lock, captured output that followers replay, and completed-run reuse on request. It is a public refactor of an internal command-run helper. It keeps the executable core: content-addressed command keys, an fcntl file lock that elects one leader per key, captured stdout and stderr that followers replay, and completed-run reuse when the caller asks for it. It does not import or expose the internal live command-run state tree, and it does not stand in for a scheduler.
\n
Whole file is
\n
The whole file is src/microcosm_core/engine_room/command_run_singleflight.py. The whole file is src/microcosm_core/engine_room/command_run_singleflight.py. Every function named below is defined there.
\n
Purpose
\n
The key spans argv, cwd, environment, and scope
\n
Rather than compare names or argv alone, the component keys on argv, cwd, an environment slice, and scoped worktree state, so two runs collapse only while their code is identical. A loose deduplicator compares command names or memoizes by argv alone. That is wrong here, because the same argv can mean different work: edit a file the command reads and the correct answer changes. So the interesting decision is how this component decides two requests are the same. It does not compare names. It builds a key over the argv, the working directory, a small slice of the environment, and the scoped worktree state. Two runs collapse into one only while the code they would see is identical. Change a file in scope and the key changes, which forces a fresh run instead of serving a stale result.
\n
How it works
\n
One entry function, two role functions
\n
The pipeline is short: one entry function takes a per-key lock, decides a role, and returns a result record, while the heavy work lives in two role functions. The pipeline is short. One entry function takes a per-key lock, decides a role, and returns a result record. The heavy work lives in two role functions.
\n
Function
Role
build_command_key
Builds the content-addressed key from argv, cwd, env slice, and scoped state.
run_command_singleflight
Entry point: locks the key, picks leader, follower, or reuse, returns a RunReceipt.
_run_leader
Runs the subprocess once, captures output, writes the run records.
_wait_for_active
Follower path: waits for the active run and replays its captured output.
evaluate_fixture_dir
Runs the four public fixture cases and reports pass or fail.
\n
build_command_key hashes into a dirty_fingerprint
\n
build_command_key hashes argv, cwd, and a dirty_fingerprint over scoped Git or file state, keeping only a PYTHONPATH and PYTEST_ADDOPTS environment slice and no absolute paths.build_command_key consumes the argv, the resolved cwd, a resource class, the scoped paths, and an environment mapping. It hashes the argv, labels and hashes the cwd, and computes a dirty_fingerprint over the scope. Inside a Git repository that fingerprint carries HEAD, the porcelain status hash, the working diff hash, the staged diff hash, and per-file content hashes for the scoped paths. Outside Git it falls back to the scoped file-content fingerprint alone. The environment slice is narrow: PYTHONPATH and PYTEST_ADDOPTS. The key is a dictionary of hashes and labels, not absolute paths, so the records stay public-safe. run_command_singleflight then reduces it to a 24-character key_hash.
\n
run_command_singleflight elects a role under the lock
\n
run_command_singleflight rejects empty argv, then under an fcntl.LOCK_EX per-key lock reads the active record and becomes a follower, a reuse, or the leader.run_command_singleflight rejects empty argv with a ValueError before it builds anything. It resolves the cwd and state root, builds the key, and opens the per-key lock file under fcntl.LOCK_EX. Holding the lock, it reads the active record for this key. If an active run is still running and its process is alive or inside the pending window, this caller becomes a follower. If an active run has completed and the caller passed reuse_completed=True, it becomes a reuse. Otherwise it writes fresh active and run metadata, releases the lock, and becomes the leader.
\n
_run_leader runs the subprocess once
\n
_run_leader starts the subprocess once, captures stdout and stderr to per-run files, marks the run completed with its exit code, and returns a leaderRunReceipt._run_leader starts the subprocess once with subprocess.Popen, records the active run with its pid, then blocks on communicate to capture stdout and stderr. It writes those to per-run output files, marks the run completed with its exit code, updates the latest-by-key record, and appends leader lifecycle events. It returns a RunReceipt with role set to leader.
\n
_wait_for_active replays without rerunning
\n
_wait_for_active polls until the run completes and replays the leader's output with the same run_id, and on a dead process or timeout returns stale_or_timeout with exit code 124._wait_for_active is the follower path. It polls the active record until the status reads completed, then reads the leader's captured output files and returns a record that carries the same stdout, stderr, exit code, and run_id as the leader. It never launches its own subprocess. If the active run's process dies or the run does not finish inside the wait window, the follower returns status: stale_or_timeout with exit code 124 and still does not rerun.
\nDiagram of the mechanism (7 steps).Diagram source & refs
Source refs
take per-key fcntl lock
run_command_singleflight
flowchart TD A["argv, cwd, env, scope paths"] --> B["build_command_key hash argv + cwd + env + scoped state"] B --> C["run_command_singleflight take per-key fcntl lock"] C --> D{"active run for this key?"} D -- "running" --> E["_wait_for_active follower replays captured output"] D -- "completed and reuse_completed" --> F["replay completed result"] D -- "none or stale" --> G["_run_leader run subprocess once, capture output"]
\n
Negative cases
\n
Two positive and two boundary fixture cases
\n
evaluate_fixture_dir checks four cases: single_leader and completed_reuse pass, while scope_mutation_changes_key and missing_command_rejected guard the key and reject empty argv.evaluate_fixture_dir runs four named cases and checks each observed result against the expected flag. Two are positive: single_leader confirms one caller wins the lock and runs the command, and completed_reuse runs a counter command, then reuses the completed result and checks the counter stayed at 1. Two are boundary cases. scope_mutation_changes_key writes a scoped file, builds a key, rewrites the file, builds the key again, and requires the two key_hash values to differ, so edited scoped content cannot be laundered into an old run. missing_command_rejected calls the entry function with empty argv and requires the ValueError, so an empty command is never accepted. A pass is case_count: 4 and passed_case_count: 4.
\n
Prior Art Grounding
\n
Lineage in Go's singleflight package
\n
The pattern is duplicate-call suppression from Go's singleflight, adapted to local subprocess runs with a content-addressed key, scoped state, and captured output replay, not a distributed lock. The pattern is duplicate-call suppression, the same idea as Go's singleflight package, `golang.org/x/sync/singleflight`, which defines a namespace of work where duplicate calls for one key share a single in-flight execution. This component borrows the leader and follower shape and the shared result, then adapts it to local subprocess runs with a content-addressed command key, scoped worktree state, captured output replay, and an explicit line between collapsing active duplicates and caching completed results. It is singleflight for command execution, not a distributed lock service.
The focused tests add a real two-process race where roles resolve to leader and follower, both share one run_id and counter=1, and the side effect increments exactly once. The focused tests add the real two-process race: two callers start the same command, their roles resolve to leader and follower, both records share one run_id, both replay counter=1, and the side-effect counter increments exactly once. A pass means the public fixture behavior and the bundle-backed JSON projection stay reproducible. These validation result records do not prove scheduler behavior, daemon behavior, or distributed locking.
\n
Scope boundary
\n
Scope limit
\n
The strongest claim and its boundary
\n
At most, a duplicate key elects one leader, followers replay its output, an edited file forces a miss, reuse is explicit, and empty argv is refused, with nothing beyond that. The strongest claim the evidence supports: a duplicate command key elects one leader through a per-key fcntl lock, followers replay that leader's captured output rather than rerunning, an edited scoped file changes the key and forces a miss, a completed run is reused only on explicit request, and empty argv is refused. That is the proof boundary. Its scope limit is a single subprocess singleflight fixture. It is not a job scheduler, not a daemon, not a distributed lock service, and not an export of live command-run state. A green record here is not evidence for a durable queue, a network lock, or whole-system correctness.
\n",
"components": [],
"doctrine_feed": {
"counts": {
"axiom": 5,
"concept": 1,
"mechanism": 1,
"paper_module": 1,
"principle": 6
},
"link_count": 14,
"links": [
{
"id": "mechanism.engine_room_command_run_singleflight.validates_public_command_run_singleflight",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public command run singleflight",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-command-run-singleflight-validates-public-command-run-singleflight"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.engine_room_metabolism_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Metabolism Runtime",
"relation": "depends on",
"url": "paper-module-engine-room-metabolism-runtime.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[31:paper_module.engine_room_demo].depends_on",
"url": "paper-module-engine-room-demo.html"
}
],
"paper_module_id": "paper_module.engine_room_command_run_singleflight",
"reader_packet": {
"abstract": "Public command-run singleflight fixture: content-addressed subprocess keys collapse duplicate active runs and replay captured result records without claiming scheduler or daemon authority.",
"authority_ceiling": "Public subprocess singleflight fixture and focused regression result records only; no job scheduler, daemon, distributed lock service, live command_runs export, external model access authority, launch-scope decision, whole-system equivalence, source-file changes, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.engine_room_command_run_singleflight.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-engine-room-command-run-singleflight-diagram-1",
"html_url": "docs/paper-module-engine-room-command-run-singleflight.html#paper-module-engine-room-command-run-singleflight-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"argv, cwd, env, scope paths\"] --> B[\"build_command_key hash argv + cwd + env + scoped state\"] B --> C[\"run_command_singleflight take per-key fcntl lock\"] C --> D{\"active run for this key?\"} D -- \"running\" --> E[\"_wait_for_active follower replays captured output\"] D -- \"completed and reuse_completed\" --> F[\"replay completed result\"] D -- \"none or stale\" --> G[\"_run_leader run subprocess once, capture output\"]",
"source_format": "mermaid",
"source_hash": "sha256:95f79813d5b95014616e9b087d8e5485d5e10a45a44872a0be5828aa9b782f65",
"source_line_end": 125,
"source_line_start": 117
}
],
"grouped_relation_feeds": [
{
"count": 5,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.engine_room_metabolism_runtime",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Metabolism Runtime",
"relation": "depends on",
"url": "paper-module-engine-room-metabolism-runtime.html"
}
],
"relation": "depends on"
},
{
"count": 1,
"links": [
{
"id": "mechanism.engine_room_command_run_singleflight.validates_public_command_run_singleflight",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public command run singleflight",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-command-run-singleflight-validates-public-command-run-singleflight"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.engine_room_command_run_singleflight",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 209,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_command_run_singleflight.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "`golang.org/x/sync/singleflight`",
"section_id": "lineage-in-go-s-singleflight-package",
"url": "https://pkg.go.dev/golang.org/x/sync/singleflight"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**When two callers fire the same command at once, this component runs it exactly once and hands the second caller the first caller's captured output.** When several agents or background tasks fire the same command at the same moment, the plain outcome is several identical subprocesses doing the same work at once. That wastes the machine, and where the command writes something it can corrupt shared state by writing twice. This component answers one question: when two callers ask for the same command at the same time, can the system run it exactly once and hand the second caller the first caller's captured output?",
"level": 4,
"order": 1,
"section_id": "when-two-callers",
"source_line_end": 14,
"source_line_start": 3,
"title": "When two callers"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is a public refactor keeping the executable core: content-addressed keys, an fcntl leader lock, captured output that followers replay, and completed-run reuse on request.** It is a public refactor of an internal command-run helper. It keeps the executable core: content-addressed command keys, an fcntl file lock that elects one leader per key, captured stdout and stderr that followers replay, and completed-run reuse when the caller asks for it. It does not import or expose the internal live command-run state tree, and it does not stand in for a scheduler.",
"level": 4,
"order": 2,
"section_id": "is-public-refactor",
"source_line_end": 25,
"source_line_start": 15,
"title": "Is public refactor"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The whole file is src/microcosmcore/engineroom/commandrunsingleflight.py. The whole file is src/microcosmcore/engineroom/commandrunsingleflight.py. Every function named below is defined there.",
"level": 4,
"order": 3,
"section_id": "whole-file-is",
"source_line_end": 31,
"source_line_start": 26,
"title": "Whole file is"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 33,
"source_line_start": 32,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Rather than compare names or argv alone, the component keys on argv, cwd, an environment slice, and scoped worktree state, so two runs collapse only while their code is identical.** A loose deduplicator compares command names or memoizes by argv alone. That is wrong here, because the same argv can mean different work: edit a file the command reads and the correct answer changes. So the interesting decision is how this component decides two requests are the same. It does not compare names. It builds a key over the argv, the working directory, a small slice of the environment, and the scoped worktree state. Two runs collapse into one only while the code they would see is identical. Change a\u2026",
"level": 4,
"order": 5,
"section_id": "the-key-spans-argv-cwd-environment-and-scope",
"source_line_end": 47,
"source_line_start": 34,
"title": "The key spans argv, cwd, environment, and scope"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 6,
"section_id": "how-it-works",
"source_line_end": 49,
"source_line_start": 48,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "**The pipeline is short: one entry function takes a per-key lock, decides a role, and returns a result record, while the heavy work lives in two role functions.** The pipeline is short. One entry function takes a per-key lock, decides a role, and returns a result record. The heavy work lives in two role functions.",
"level": 4,
"order": 7,
"section_id": "one-entry-function-two-role-functions",
"source_line_end": 65,
"source_line_start": 50,
"title": "One entry function, two role functions"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "buildcommandkey hashes argv, cwd, and a dirtyfingerprint over scoped Git or file state, keeping only a PYTHONPATH and PYTEST_ADDOPTS environment slice and no absolute paths. buildcommand_key consumes the argv, the resolved cwd, a resource class, the scoped paths, and an environment mapping. It hashes the argv, labels and hashes the cwd, and computes a dirty_fingerprint over the scope. Inside a Git repository that fingerprint carries HEAD, the porcelain status hash, the working diff hash, the staged diff hash, and per-file content hashes for the scoped paths. Outside Git it falls back to the scoped file-content fingerprint alone. The environment slice is narrow: PYTHONPATH and PYTEST_ADDOPTS\u2026",
"level": 4,
"order": 8,
"section_id": "build-command-key-hashes-into-a-dirty-fingerprint",
"source_line_end": 80,
"source_line_start": 66,
"title": "`build_command_key` hashes into a `dirty_fingerprint`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "runcommandsingleflight rejects empty argv, then under an fcntl.LOCKEX per-key lock reads the active record and becomes a follower, a reuse, or the leader. runcommand_singleflight rejects empty argv with a ValueError before it builds anything. It resolves the cwd and state root, builds the key, and opens the per-key lock file under fcntl.LOCK_EX. Holding the lock, it reads the active record for this key. If an active run is still running and its process is alive or inside the pending window, this caller becomes a follower. If an active run has completed and the caller passed reuse_completed=True, it becomes a reuse. Otherwise it writes fresh active and run metadata, releases the lock, and be\u2026",
"level": 4,
"order": 9,
"section_id": "run-command-singleflight-elects-a-role-under-the-lock",
"source_line_end": 93,
"source_line_start": 81,
"title": "`run_command_singleflight` elects a role under the lock"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "runleader starts the subprocess once, captures stdout and stderr to per-run files, marks the run completed with its exit code, and returns a leader RunReceipt. run_leader starts the subprocess once with subprocess.Popen, records the active run with its pid, then blocks on communicate to capture stdout and stderr. It writes those to per-run output files, marks the run completed with its exit code, updates the latest-by-key record, and appends leader lifecycle events. It returns a RunReceipt with role set to leader.",
"level": 4,
"order": 10,
"section_id": "run-leader-runs-the-subprocess-once",
"source_line_end": 104,
"source_line_start": 94,
"title": "`_run_leader` runs the subprocess once"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "waitfor_active polls until the run completes and replays the leader's output with the same run_id, and on a dead process or timeout returns staleortimeout with exit code 124. waitforactive is the follower path. It polls the active record until the status reads completed, then reads the leader's captured output files and returns a record that carries the same stdout, stderr, exit code, and run_id as the leader. It never launches its own subprocess. If the active run's process dies or the run does not finish inside the wait window, the follower returns status: staleortimeout with exit code 124 and still does not rerun.",
"level": 4,
"order": 11,
"section_id": "wait-for-active-replays-without-rerunning",
"source_line_end": 126,
"source_line_start": 105,
"title": "`_wait_for_active` replays without rerunning"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 12,
"section_id": "negative-cases",
"source_line_end": 128,
"source_line_start": 127,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluatefixturedir checks four cases: singleleader and completedreuse pass, while scopemutationchangeskey and missingcommandrejected guard the key and reject empty argv.** evaluatefixturedir runs four named cases and checks each observed result against the expected flag. Two are positive: single_leader confirms one caller wins the lock and runs the command, and completed_reuse runs a counter command, then reuses the completed result and checks the counter stayed at 1. Two are boundary cases. scopemutationchanges_key writes a scoped file, builds a key, rewrites the file, builds the key again, and requires the two key_hash values to differ, so edited scoped content cannot be laundered into an\u2026",
"level": 4,
"order": 13,
"section_id": "two-positive-and-two-boundary-fixture-cases",
"source_line_end": 144,
"source_line_start": 129,
"title": "Two positive and two boundary fixture cases"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "prior-art-grounding",
"source_line_end": 146,
"source_line_start": 145,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The pattern is duplicate-call suppression from Go's singleflight, adapted to local subprocess runs with a content-addressed key, scoped state, and captured output replay, not a distributed lock.** The pattern is duplicate-call suppression, the same idea as Go's singleflight package, golang.org/x/sync/singleflight, which defines a namespace of work where duplicate calls for one key share a single in-flight execution. This component borrows the leader and follower shape and the shared result, then adapts it to local subprocess runs with a content-addressed command key, scoped worktree state, captured output replay, and an explicit line between collapsing active duplicates and caching comple\u2026",
"level": 4,
"order": 15,
"section_id": "lineage-in-go-s-singleflight-package",
"source_line_end": 161,
"source_line_start": 147,
"title": "Lineage in Go's singleflight package"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Exercise the four fixture cases: Run the focused tests and the corpus parity check:",
"level": 2,
"order": 16,
"section_id": "validation-receipt-path",
"source_line_end": 180,
"source_line_start": 162,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The focused tests add a real two-process race where roles resolve to leader and follower, both share one run_id and counter=1, and the side effect increments exactly once.** The focused tests add the real two-process race: two callers start the same command, their roles resolve to leader and follower, both records share one run_id, both replay counter=1, and the side-effect counter increments exactly once. A pass means the public fixture behavior and the bundle-backed JSON projection stay reproducible. These validation result records do not prove scheduler behavior, daemon behavior, or distributed locking.",
"level": 4,
"order": 17,
"section_id": "the-two-process-race-the-tests-add",
"source_line_end": 192,
"source_line_start": 181,
"title": "The two-process race the tests add"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "claim-ceiling",
"source_line_end": 194,
"source_line_start": 193,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**At most, a duplicate key elects one leader, followers replay its output, an edited file forces a miss, reuse is explicit, and empty argv is refused, with nothing beyond that.** The strongest claim the evidence supports: a duplicate command key elects one leader through a per-key fcntl lock, followers replay that leader's captured output rather than rerunning, an edited scoped file changes the key and forces a miss, a completed run is reused only on explicit request, and empty argv is refused. That is the proof boundary. Its scope limit is a single subprocess singleflight fixture. It is not a job scheduler, not a daemon, not a distributed lock service, and not an export of live command-run\u2026",
"level": 4,
"order": 19,
"section_id": "the-strongest-claim-and-its-boundary",
"source_line_end": 208,
"source_line_start": 195,
"title": "The strongest claim and its boundary"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/engine_room_command_run_singleflight.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Public command-run singleflight fixture: content-addressed subprocess keys collapse duplicate active runs and replay captured result records without claiming scheduler or daemon authority."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "command_run_singleflight.py",
"path": "src/microcosm_core/engine_room/command_run_singleflight.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for content-addressed command keys, fcntl leader/follower singleflight, completed-run reuse, captured output replay, scoped fingerprinting, and fixture evaluation.",
"symbols": [
"build_command_key",
"run_command_singleflight",
"evaluate_fixture_dir",
"_run_leader",
"_wait_for_active"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/command_run_singleflight.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_command_run_singleflight.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_command_run_singleflight.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Engine Room Command-Run Singleflight is a subprocess singleflight bundle. It validates content-addressed command keys, scoped dirty/content fingerprints, fcntl leader/follower election, completed-run reuse, captured output replay, and two negative boundaries over fixture commands while keeping scheduler, daemon, distributed-lock, live-state export, launch, and private-system claims out of scope.",
"summary_line": "Public command-run singleflight fixture: content-addressed subprocess keys collapse duplicate active runs and replay captured result records without claiming scheduler or daemon authority.",
"title": "Engine Room Command-Run Singleflight"
},
{
"anchor": "paper-module-engine-room-metabolism-runtime",
"authority_ceiling": "Component evidence for the accepted staged Engine Room demo only; not a live non-public runtime export, not external model service, not agent dispatch, not distributed database proof, not launch-scope decision, and not source-file changes.",
"body_html": "
Long-running runtime's durable
\n
A long-running runtime's durable record of jobs, leases, runs, and claims drifts, so this component reads that state alone and asks which rows are inconsistent and which a person should review first. A long-running agent runtime keeps a durable record of its work: jobs waiting to run, leases held by workers, runs in flight, and claims asserted on a shared blackboard. That record drifts. A worker dies mid-run and its lease is never released. A run finishes but the job it belonged to still reads running. A launch log goes stale because nothing writes to it. This component takes the durable state alone and answers one question: which rows are inconsistent, and which of those should a person look at before anything touches them.
\n
Is synthetic sqlite
\n
It is a synthetic SQLite exercise, not the live runtime: a single stdlib file that creates a real database, drives the queue, lease, run, blackboard, and reconciliation paths, and emits one JSON record. It is a synthetic SQLite exercise, not the live runtime. It ships fixtures and creates a real database file, drives the queue, lease, run, blackboard, and reconciliation paths over that file, and emits one JSON result record. The whole component is src/microcosm_core/engine_room/metabolism_runtime.py, a single file that uses only the Python standard library.
\n
Touches no live
\n
It touches no live runtime: it reads no private runtime database, dispatches no worker or provider, and does not stand in for distributed-database behaviour. It reads no private runtime database, dispatches no worker or provider, and does not stand in for distributed-database behaviour.
\n
Purpose
\n
What reconciliation refuses to repair
\n
Reconciliation tags every finding operator_review_required and refuses to auto-repair ambiguous state, recovering only the expired lease that has a clean recovery path via requeue_expired_jobs. The interesting choice is what the reconciliation pass refuses to do. It reads the jobs and runs tables, applies its rules, and tags every finding operator_review_required. It does not auto-repair. An expired lease has a clean recovery path, so requeue_expired_jobs moves it back to recoverable on its own. But a running job with no run row, or a finished run whose job still reads running, is ambiguous. The safe move is to surface it, not to guess. A loose runtime that guesses here corrupts the record; this component draws the line and stays on the safe side of it.
\n
The blackboard is append-only
\n
Claims are never edited or deleted in place. An assertion is one event row. A contradiction, expiry, or supersession is a separate event that points back at the assertion it invalidates. State is reconstructed from an append-only history, so the reason a claim is no longer active stays on the record.
\n
How it works
\n
Mechanism lives in
\n
The mechanism lives in one module. The stages below run in the order the fixture harness drives them.
\n
Storage opens a WAL SQLite store
\n
connect opens a WAL SQLite file with foreign keys on and ensure_schema creates the jobs, runs, and blackboard_claim_events tables plus a partial unique index over the active states.connect opens a SQLite file with PRAGMA journal_mode=WAL, normal sync, and foreign keys on, then calls ensure_schema. ensure_schema creates three tables, jobs, runs, and blackboard_claim_events, and a partial unique index idx_jobs_active_idempotency on idempotency_key restricted to the four active states in ACTIVE_STATES (queued, claimed, running, recoverable).
\n
Queue insertion and lease recovery
\n
enqueue_job rejects a duplicate idempotency key that is still active, claim_next_job leases the next queued or recoverable job, and requeue_expired_jobs recovers lapsed leases to recoverable.enqueue_job inserts one job row and returns (job, inserted). When a job with the same idempotency_key is still active, the partial unique index raises sqlite3.IntegrityError, which the function catches and reports as inserted=False. Once the earlier job reaches a terminal state the key is free again. claim_next_job selects the next queued or recoverable job by priority then creation time, stamps a claim_expires_at lease, and moves it to claimed. requeue_expired_jobs finds claimed or running rows whose lease has lapsed, clears the owner and expiry, moves them to recoverable, and returns how many it recovered.
\n
Runs and the finalize-job flag
\n
start_run moves a job to running and complete_run records a return code, propagating a terminal job state only when finalize_job=True, which makes the still-running-job defect reachable on purpose.start_run inserts a run row and moves its job to running through update_job_state. complete_run writes completed_at and a return code; it only propagates a terminal job state when the caller passes finalize_job=True. That flag is what makes the "finished run, still-running job" defect reachable on purpose.
\n
Blackboard events and projection
\n
append_claim_event appends one immutable event of a given kind, and build_blackboard_projection replays the whole log to return the assertions that were not invalidated, with active, event, and contradiction counts.append_claim_event appends one event of a given kind (claim_asserted, claim_contradicted, claim_expired, claim_superseded) and never mutates a prior row. build_blackboard_projection reads the whole event log, collects the assertion ids named by any invalidating event, and returns the assertions that were asserted and not invalidated, with counts for active claims, total events, and contradictions.
\n
Reconciliation surfaces four defect rules
\n
reconcile walks claimed and running jobs against their newest run and emits a ReconciliationFinding with operator_review_required for each of four defect rules, returning healthy or needs_review.reconcile walks the claimed and running jobs, pulls the newest run for each with latest_run_for_job, and emits a ReconciliationFinding for each defect it sees. Every finding carries an expected value, an observed value, and the action operator_review_required, and derives a stable finding_id from its rule and object. The four rules are running_job_no_run_row, run_finalized_but_job_running, running_job_stale_launch_log, and running_job_missing_launch_log. The pass returns status healthy when it finds nothing and needs_review when it does, plus per-rule counts, the findings, the CLAIM_CEILING string, and the scope boundary list.
\n
Fixture harness and CLI
\n
evaluate_case runs one named case in an isolated scratch directory, evaluate_fixture_dir passes only when every case met its expectation, and main exposes this as the evaluate-fixtures command.evaluate_case runs one named case against an isolated scratch directory and compares its status to the expected status. evaluate_fixture_dir reads every *.json file in the input directory, runs each case, and returns status: pass only when at least one case exists and every case met its expectation. build_parser and main expose this as the evaluate-fixtures command, printing the JSON record and returning a process exit code of 0 for pass and 1 for fail.
\n
Function
Role
connect, ensure_schema
Open the WAL SQLite store and install the three tables and indexes
enqueue_job
Insert a job, reject a duplicate that is still active
claim_next_job, requeue_expired_jobs
Lease the next job, recover a lapsed lease to recoverable
start_run, complete_run
Record a run and, on request, finalise its job
append_claim_event, build_blackboard_projection
Append claim events, project the active claims
reconcile, ReconciliationFinding
Detect inconsistent job, run, and log state as review findings
evaluate_fixture_dir, main
Replay the fixture cases and emit the JSON record
\nDiagram of the mechanism (7 steps).Diagram source & refs
Source refs
append_claim_event
build_blackboard_projection
flowchart TD Fixtures["evaluate_fixture_dir reads *.json cases"] Store["connect / ensure_schema WAL SQLite: jobs, runs, blackboard_claim_events"] Queue["enqueue_job / claim_next_job requeue_expired_jobs"] Runs["start_run / complete_run"] Board["append_claim_event build_blackboard_projection"] Recon["reconcile ReconciliationFinding per defect"] Record["JSON record status, rule counts, findings, scope limit"] Fixtures --> Store Store --> Queue Queue --> Runs Store --> Board Runs --> Recon Board --> Recon Recon --> Record
\n
Negative cases
\n
Five fixture cases, three forcing findings
\n
Five named fixture cases each expect pass, and three of them exist to force a reconciliation finding. The fixtures in fixtures/first_wave/engine_room_metabolism_runtime/input are five named cases, each expecting pass. Three of them exist to force a reconciliation finding.
\n
Queue recovery and blackboard projection cases
\n
queue_recovery rejects a duplicate key and recovers an already-expired lease to recoverable, while blackboard_projection asserts then contradicts a claim and checks zero active claims and one contradiction.queue_recovery enqueues the same idempotency key twice, expecting the second insert to be rejected, claims the job with a lease of -1 seconds so it is already expired, and checks that requeue_expired_jobs returns 1 and the job lands in recoverable. blackboard_projection asserts one claim then contradicts it, and checks the projection reports zero active claims and one contradiction.
\n
Three cases that force reconciliation rules
\n
running_job_no_run_row, finalized_run_running_job, and stale_log each provoke their named reconciliation rule, and any case whose observed status misses its expected status fails the whole record.running_job_no_run_row marks a job running with no run row and expects running_job_no_run_row in the rule counts. finalized_run_running_job starts a run, completes it with finalize_job=False, and expects run_finalized_but_job_running. stale_log writes a log file, back-dates its mtime by 1200 seconds, and expects running_job_stale_launch_log under a 60-second freshness threshold. A case whose observed status does not match its expected status makes the whole record fail.
\n
Prior Art Grounding
\n
Durable-runtime control-loop lineage
\n
The component follows the durable-runtime control-loop lineage of autonomic computing and SRE monitoring over a WAL store, borrowing the self-management loop while keeping the jobs, leases, and findings synthetic. The component follows the durable-runtime control-loop lineage: keep the work state in a durable log, detect stale or inconsistent state, recover what has a clean recovery path, and keep that log separate from the acting dispatcher. Relevant anchors are IBM's autonomic-computing architecture and its monitor-analyze-plan-execute loop, SQLite write-ahead logging for the local durability and concurrency behaviour used here, and Google's SRE monitoring guidance for separating symptoms from causes. The borrowed idea is the self-management loop over a durable local store; the jobs, leases, claims, and findings are synthetic and public-safe.
\n
Composition context
\n
Admitted as a staged demo bundle
\n
demo.py stages this component as one CapsuleExercise covering the metabolism_runtime and metabolism_reconciler targets, running its evaluate_fixture_dir in sequence as a demo, not registry integration or an admission gate.src/microcosm_core/engine_room/demo.py admits this component as one staged bundle. Its CAPSULES tuple holds a CapsuleExercise for engine_room_metabolism_runtime covering the jewel targets metabolism_runtime and metabolism_reconciler. run_capsule imports the module and calls its evaluate_fixture_dir. audit_controller_coverage checks that each bundle has its expected surfaces on disk and that the covered targets match EXPECTED_JEWEL_TARGETS. The demo runs the bundles in sequence and reports whether they pass; it is not registry integration or an admission gate.
The CLI passes only when every fixture case met its expected status, the pytest suite exercises each path, and the corpus check keeps the generated row in parity, showing reproducibility and nothing more. The CLI emits a record with organ_id: engine_room_metabolism_runtime; a pass means every fixture case met its expected status. The pytest suite exercises WAL and idempotency, expired-lease recovery, the contradicted-claim projection, each reconciliation rule, fixture replay, and the CLI. The corpus check keeps the generated JSON row in parity with the bundle registry. A pass shows the synthetic behaviour is reproducible and nothing more.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
The component is a reproducible synthetic SQLite exercise for a durable job queue, lease recovery, append-only blackboard projection, and a reconciliation taxonomy that surfaces ambiguous state for review instead of repairing it. The strongest claim the evidence supports: this component is a reproducible synthetic SQLite exercise for a durable job queue, lease recovery to recoverable, append-only blackboard claim-event projection, and a cold-start reconciliation taxonomy that surfaces ambiguous state for review instead of repairing it. That is the whole scope limit.
\n
What it is not
\n
It is not a live non-public runtime export, a dispatcher, an auto-repairer, or a distributed database, and its concurrency is only that of a single local WAL file. It is not a live non-public runtime export. It is not an agent dispatcher and calls no provider. It does not auto-repair ambiguous job or run state. It is not a distributed database, and its concurrency behaviour is only that of a single local WAL file. It ships no private runtime database, status JSON, operator session, or live log. Passing the checks proves the fixture behaviour is replayable; it proves nothing about live runtime health, launch-scope decision, or whole-system correctness.
\n",
"components": [],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"paper_module": 1,
"principle": 6
},
"link_count": 13,
"links": [
{
"id": "mechanism.engine_room_demo.validates_public_engine_room_demo",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public engine room demo",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-demo-validates-public-engine-room-demo"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.engine_room_command_run_singleflight",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Command-Run Singleflight",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[81:paper_module.engine_room_command_run_singleflight].depends_on",
"url": "paper-module-engine-room-command-run-singleflight.html"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[31:paper_module.engine_room_demo].depends_on",
"url": "paper-module-engine-room-demo.html"
}
],
"paper_module_id": "paper_module.engine_room_metabolism_runtime",
"reader_packet": {
"abstract": "Staged Engine Room component: synthetic SQLite metabolism runtime exercise for queues, leases, blackboard projection, and reconciliation.",
"authority_ceiling": "Component evidence for the accepted staged Engine Room demo only; not a live non-public runtime export, not external model service, not agent dispatch, not distributed database proof, not launch-scope decision, and not source-file changes.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.engine_room_metabolism_runtime.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-engine-room-metabolism-runtime-diagram-1",
"html_url": "docs/paper-module-engine-room-metabolism-runtime.html#paper-module-engine-room-metabolism-runtime-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Fixtures[\"evaluate_fixture_dir reads *.json cases\"] Store[\"connect / ensure_schema WAL SQLite: jobs, runs, blackboard_claim_events\"] Queue[\"enqueue_job / claim_next_job requeue_expired_jobs\"] Runs[\"start_run / complete_run\"] Board[\"append_claim_event build_blackboard_projection\"] Recon[\"reconcile ReconciliationFinding per defect\"] Record[\"JSON record status, rule counts, findings, scope limit\"] Fixtures --> Store Store --> Queue Queue --> Runs Store --> Board Runs --> Recon Board --> Recon Recon --> Record",
"source_format": "mermaid",
"source_hash": "sha256:dff6da54110576a78ad755c616337d889dcea23501421871826c9ad236de7d36",
"source_line_end": 170,
"source_line_start": 153
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"relation": "depends on"
},
{
"count": 1,
"links": [
{
"id": "mechanism.engine_room_demo.validates_public_engine_room_demo",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public engine room demo",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-demo-validates-public-engine-room-demo"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.engine_room_metabolism_runtime",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 286,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_metabolism_runtime.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "autonomic-computing architecture",
"section_id": "durable-runtime-control-loop-lineage",
"url": "https://research.ibm.com/publications/autonomic-computing-architectural-approach-and-prototype"
},
{
"kind": "external_reference",
"label": "SQLite write-ahead logging",
"section_id": "durable-runtime-control-loop-lineage",
"url": "https://www.sqlite.org/wal.html"
},
{
"kind": "external_reference",
"label": "SRE monitoring guidance",
"section_id": "durable-runtime-control-loop-lineage",
"url": "https://sre.google/sre-book/monitoring-distributed-systems/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A long-running runtime's durable record of jobs, leases, runs, and claims drifts, so this component reads that state alone and asks which rows are inconsistent and which a person should review first.** A long-running agent runtime keeps a durable record of its work: jobs waiting to run, leases held by workers, runs in flight, and claims asserted on a shared blackboard. That record drifts. A worker dies mid-run and its lease is never released. A run finishes but the job it belonged to still reads running. A launch log goes stale because nothing writes to it. This component takes the durable state alone and answers one question: which rows are inconsistent, and which of those should a perso\u2026",
"level": 4,
"order": 1,
"section_id": "long-running-runtime-s-durable",
"source_line_end": 15,
"source_line_start": 3,
"title": "Long-running runtime's durable"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is a synthetic SQLite exercise, not the live runtime: a single stdlib file that creates a real database, drives the queue, lease, run, blackboard, and reconciliation paths, and emits one JSON record.** It is a synthetic SQLite exercise, not the live runtime. It ships fixtures and creates a real database file, drives the queue, lease, run, blackboard, and reconciliation paths over that file, and emits one JSON result record. The whole component is src/microcosmcore/engineroom/metabolism_runtime.py, a single file that uses only the Python standard library.",
"level": 4,
"order": 2,
"section_id": "is-synthetic-sqlite",
"source_line_end": 26,
"source_line_start": 16,
"title": "Is synthetic sqlite"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It touches no live runtime: it reads no private runtime database, dispatches no worker or provider, and does not stand in for distributed-database behaviour.** It reads no private runtime database, dispatches no worker or provider, and does not stand in for distributed-database behaviour.",
"level": 4,
"order": 3,
"section_id": "touches-no-live",
"source_line_end": 33,
"source_line_start": 27,
"title": "Touches no live"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 35,
"source_line_start": 34,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Reconciliation tags every finding operatorreview_required and refuses to auto-repair ambiguous state, recovering only the expired lease that has a clean recovery path via requeueexpiredjobs.** The interesting choice is what the reconciliation pass refuses to do. It reads the jobs and runs tables, applies its rules, and tags every finding operatorreviewrequired. It does not auto-repair. An expired lease has a clean recovery path, so requeueexpiredjobs moves it back to recoverable on its own. But a running job with no run row, or a finished run whose job still reads running, is ambiguous. The safe move is to surface it, not to guess. A loose runtime that guesses here corrupts the record; this\u2026",
"level": 4,
"order": 5,
"section_id": "what-reconciliation-refuses-to-repair",
"source_line_end": 48,
"source_line_start": 36,
"title": "What reconciliation refuses to repair"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The blackboard never edits or deletes a claim in place: each assertion, contradiction, expiry, or supersession is a separate event, and the active view Claims are never edited or deleted in place. An assertion is one event row. A contradiction, expiry, or supersession is a separate event that points back at the assertion it invalidates. log and dropping any assertion an invalidating event named. State is reconstructed from an append-only history, so the reason a claim is no longer active stays on the record.",
"level": 4,
"order": 6,
"section_id": "the-blackboard-is-append-only",
"source_line_end": 60,
"source_line_start": 49,
"title": "The blackboard is append-only"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 62,
"source_line_start": 61,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The mechanism lives in one module. The stages below run in the order the fixture harness drives them.",
"level": 4,
"order": 8,
"section_id": "mechanism-lives-in",
"source_line_end": 67,
"source_line_start": 63,
"title": "Mechanism lives in"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "connect opens a WAL SQLite file with foreign keys on and ensureschema creates the jobs, runs, and blackboardclaimevents tables plus a partial unique index over the active states.** connect opens a SQLite file with PRAGMA journal_mode=WAL, normal sync, and foreign keys on, then calls ensureschema. ensureschema creates three tables, jobs, runs, and blackboardclaimevents, and a partial unique index idxjobsactive_idempotency on idempotencykey restricted to the four active states in ACTIVESTATES (queued, claimed, running, recoverable).",
"level": 4,
"order": 9,
"section_id": "storage-opens-a-wal-sqlite-store",
"source_line_end": 78,
"source_line_start": 68,
"title": "Storage opens a WAL SQLite store"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "enqueuejob rejects a duplicate idempotency key that is still active, claimnextjob leases the next queued or recoverable job, and requeueexpiredjobs recovers lapsed leases to recoverable. enqueuejob inserts one job row and returns (job, inserted). When a job with the same idempotency_key is still active, the partial unique index raises sqlite3.IntegrityError, which the function catches and reports as inserted=False. Once the earlier job reaches a terminal state the key is free again. claimnextjob selects the next queued or recoverable job by priority then creation time, stamps a claimexpiresat lease, and moves it to claimed. requeueexpiredjobs finds claimed or running rows whose lease has la\u2026",
"level": 4,
"order": 10,
"section_id": "queue-insertion-and-lease-recovery",
"source_line_end": 93,
"source_line_start": 79,
"title": "Queue insertion and lease recovery"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "startrun moves a job to running and complete_run records a return code, propagating a terminal job state only when finalize_job=True, which makes the still-running-job defect reachable on purpose. startrun inserts a run row and moves its job to running through updatejobstate. complete_run writes completed_at and a return code; it only propagates a terminal job state when the caller passes finalize_job=True. That flag is what makes the \"finished run, still-running job\" defect reachable on purpose.",
"level": 4,
"order": 11,
"section_id": "runs-and-the-finalize-job-flag",
"source_line_end": 103,
"source_line_start": 94,
"title": "Runs and the finalize-job flag"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "appendclaim_event appends one immutable event of a given kind, and buildblackboardprojection replays the whole log to return the assertions that were not invalidated, with active, event, and contradiction counts.** appendclaimevent appends one event of a given kind (claim_asserted, claimcontradicted, claimexpired, claim_superseded) and never mutates a prior row. buildblackboardprojection reads the whole event log, collects the assertion ids named by any invalidating event, and returns the assertions that were asserted and not invalidated, with counts for active claims, total events, and contradictions.",
"level": 4,
"order": 12,
"section_id": "blackboard-events-and-projection",
"source_line_end": 115,
"source_line_start": 104,
"title": "Blackboard events and projection"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**reconcile walks claimed and running jobs against their newest run and emits a ReconciliationFinding with operatorreviewrequired for each of four defect rules, returning healthy or needsreview. reconcile walks the claimed and running jobs, pulls the newest run for each with latestrunfor_job, and emits a ReconciliationFinding for each defect it sees. Every finding carries an expected value, an observed value, and the action operatorreviewrequired, and derives a stable finding_id from its rule and object. The four rules are runningjobnorunrow, runfinalizedbutjobrunning, runningjobstalelaunchlog, and runningjobmissinglaunchlog. The pass returns status healthy when it finds nothing and needs_r\u2026",
"level": 4,
"order": 13,
"section_id": "reconciliation-surfaces-four-defect-rules",
"source_line_end": 130,
"source_line_start": 116,
"title": "Reconciliation surfaces four defect rules"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "evaluatecase runs one named case in an isolated scratch directory, evaluatefixturedir passes only when every case met its expectation, and main exposes this as the evaluate-fixtures command. evaluatecase runs one named case against an isolated scratch directory and compares its status to the expected status. evaluatefixturedir reads every *.json file in the input directory, runs each case, and returns status: pass only when at least one case exists and every case met its expectation. build_parser and main expose this as the evaluate-fixtures command, printing the JSON record and returning a process exit code of 0 for pass and 1 for fail.",
"level": 4,
"order": 14,
"section_id": "fixture-harness-and-cli",
"source_line_end": 171,
"source_line_start": 131,
"title": "Fixture harness and CLI"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "negative-cases",
"source_line_end": 173,
"source_line_start": 172,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Five named fixture cases each expect pass, and three of them exist to force a reconciliation finding.** The fixtures in fixtures/firstwave/engineroommetabolismruntime/input are five named cases, each expecting pass. Three of them exist to force a reconciliation finding.",
"level": 4,
"order": 16,
"section_id": "five-fixture-cases-three-forcing-findings",
"source_line_end": 180,
"source_line_start": 174,
"title": "Five fixture cases, three forcing findings"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "queuerecovery rejects a duplicate key and recovers an already-expired lease to recoverable, while blackboard_projection asserts then contradicts a claim and checks zero active claims and one contradiction. queuerecovery enqueues the same idempotency key twice, expecting the second insert to be rejected, claims the job with a lease of -1 seconds so it is already expired, and checks that requeueexpiredjobs returns 1 and the job lands in recoverable. blackboard_projection asserts one claim then contradicts it, and checks the projection reports zero active claims and one contradiction.",
"level": 4,
"order": 17,
"section_id": "queue-recovery-and-blackboard-projection-cases",
"source_line_end": 191,
"source_line_start": 181,
"title": "Queue recovery and blackboard projection cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "runningjobnorunrow, finalizedrunrunningjob, and stale_log each provoke their named reconciliation rule, and any case whose observed status misses its expected status fails the whole record. runningjobnorun_row marks a job running with no run row and expects runningjobnorunrow in the rule counts. finalizedrunrunning_job starts a run, completes it with finalizejob=False, and expects runfinalizedbutjobrunning. stalelog writes a log file, back-dates its mtime by 1200 seconds, and expects runningjobstalelaunchlog under a 60-second freshness threshold. A case whose observed status does not match its expected status makes the whole record fail.",
"level": 4,
"order": 18,
"section_id": "three-cases-that-force-reconciliation-rules",
"source_line_end": 203,
"source_line_start": 192,
"title": "Three cases that force reconciliation rules"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "prior-art-grounding",
"source_line_end": 205,
"source_line_start": 204,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component follows the durable-runtime control-loop lineage of autonomic computing and SRE monitoring over a WAL store, borrowing the self-management loop while keeping the jobs, leases, and findings synthetic.** The component follows the durable-runtime control-loop lineage: keep the work state in a durable log, detect stale or inconsistent state, recover what has a clean recovery path, and keep that log separate from the acting dispatcher. Relevant anchors are IBM's autonomic-computing architecture and its monitor-analyze-plan-execute loop, SQLite write-ahead logging for the local durability and concurrency behaviour used here, and Google's SRE monitoring guidance for separating symp\u2026",
"level": 4,
"order": 20,
"section_id": "durable-runtime-control-loop-lineage",
"source_line_end": 221,
"source_line_start": 206,
"title": "Durable-runtime control-loop lineage"
},
{
"block_kinds": [
"code"
],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "validation-receipt-path",
"source_line_end": 232,
"source_line_start": 222,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The CLI passes only when every fixture case met its expected status, the pytest suite exercises each path, and the corpus check keeps the generated row in parity, showing reproducibility and nothing more.** The CLI emits a record with organid: engineroommetabolismruntime; a pass means every fixture case met its expected status. The pytest suite exercises WAL and idempotency, expired-lease recovery, the contradicted-claim projection, each reconciliation rule, fixture replay, and the CLI. The corpus check keeps the generated JSON row in parity with the bundle registry. A pass shows the synthetic behaviour is reproducible and nothing more.",
"level": 4,
"order": 22,
"section_id": "what-the-receipts-prove",
"source_line_end": 244,
"source_line_start": 233,
"title": "What the result records prove"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 23,
"section_id": "composition-context",
"source_line_end": 246,
"source_line_start": 245,
"title": "Composition context"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**demo.py stages this component as one CapsuleExercise covering the metabolismruntime and metabolismreconciler targets, running its evaluatefixturedir in sequence as a demo, not registry integration or an admission gate. src/microcosmcore/engine_room/demo.py admits this component as one staged bundle. Its Bundles tuple holds a CapsuleExercise for engineroommetabolismruntime covering the jewel targets metabolismruntime and metabolismreconciler. runcapsule imports the module and calls its evaluatefixturedir. auditcontrollercoverage checks that each bundle has its expected surfaces on disk and that the covered targets match EXPECTEDJEWELTARGETS. The demo runs the bundles in sequence and report\u2026",
"level": 4,
"order": 24,
"section_id": "admitted-as-a-staged-demo-capsule",
"source_line_end": 260,
"source_line_start": 247,
"title": "Admitted as a staged demo bundle"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 25,
"section_id": "claim-ceiling",
"source_line_end": 262,
"source_line_start": 261,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component is a reproducible synthetic SQLite exercise for a durable job queue, lease recovery, append-only blackboard projection, and a reconciliation taxonomy that surfaces ambiguous state for review instead of repairing it.** The strongest claim the evidence supports: this component is a reproducible synthetic SQLite exercise for a durable job queue, lease recovery to recoverable, append-only blackboard claim-event projection, and a cold-start reconciliation taxonomy that surfaces ambiguous state for review instead of repairing it. That is the whole scope limit.",
"level": 4,
"order": 26,
"section_id": "the-strongest-supported-claim",
"source_line_end": 273,
"source_line_start": 263,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is not a live non-public runtime export, a dispatcher, an auto-repairer, or a distributed database, and its concurrency is only that of a single local WAL file.** It is not a live non-public runtime export. It is not an agent dispatcher and calls no provider. It does not auto-repair ambiguous job or run state. It is not a distributed database, and its concurrency behaviour is only that of a single local WAL file. It ships no private runtime database, status JSON, operator session, or live log. Passing the checks proves the fixture behaviour is replayable; it proves nothing about live runtime health, launch-scope decision, or whole-system correctness.",
"level": 4,
"order": 27,
"section_id": "what-it-is-not",
"source_line_end": 285,
"source_line_start": 274,
"title": "What it is not"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/engine_room_metabolism_runtime.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Staged Engine Room component: synthetic SQLite metabolism runtime exercise for queues, leases, blackboard projection, and reconciliation."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "metabolism_runtime.py",
"path": "src/microcosm_core/engine_room/metabolism_runtime.py",
"resolution": "resolved",
"role": "Component runtime for synthetic SQLite metabolism fixtures and public result record emission.",
"symbols": [
"evaluate_fixture_dir",
"build_parser",
"main",
"ORGAN_ID",
"CLAIM_CEILING"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/metabolism_runtime.py"
},
{
"label": "demo.py",
"path": "src/microcosm_core/engine_room/demo.py",
"resolution": "resolved",
"role": "Accepted Engine Room composition locus that admits this component as a staged bundle target.",
"symbols": [
"CAPSULES",
"EXPECTED_JEWEL_TARGETS",
"run_demo",
"audit_controller_coverage"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/demo.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_metabolism_runtime.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_metabolism_runtime.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Engine Room Metabolism Runtime explains the metabolism runtime component inside the accepted Engine Room demo. It exercises synthetic SQLite queue state, lease recovery, blackboard claim projection, and cold-start reconciliation fixtures without exporting private runtime state or dispatching providers.",
"summary_line": "Staged Engine Room component: synthetic SQLite metabolism runtime exercise for queues, leases, blackboard projection, and reconciliation.",
"title": "Engine Room Metabolism Runtime"
},
{
"anchor": "paper-module-engine-room-bridge-campaign-dag",
"authority_ceiling": "Component evidence for the accepted staged Engine Room demo only; no external model access, campaign execution, reducer or synthesis correctness proof, provider safety proof, launch-scope decision, whole-system equivalence, source-file changes, or whole-system correctness.",
"body_html": "
Component is validator
\n
This component is a validator that reads a bridge campaign graph and decides, before dispatch, whether it is safe to run, never calling a provider, worker, or reducer. A bridge campaign fans one piece of work out across several agent providers, then folds their results back into a single conclusion. This component reads the graph that describes such a campaign and decides, before anything runs, whether the graph is safe to dispatch. It is a validator. It never calls a provider, never dispatches a worker, and never runs a reducer.
\n
Takes json campaign
\n
It takes one JSON campaign spec and returns a ValidationResult of typed rule decisions with errors and warnings, so a rejection names exactly which rule failed and why. The code lives in src/microcosm_core/engine_room/bridge_campaign_dag.py. It takes one JSON campaign spec and returns a ValidationResult: a list of typed rule decisions, plus the errors and warnings those decisions produced. Each decision carries a rule id, the target it looked at, and a plain reason. The whole point is that when a graph is rejected, the result says exactly which rule rejected it and why.
\n
Engine room demo
\n
The Engine Room demo admits this component as a staged target, so the fixtures a reader runs here are the same ones audit_controller_coverage and run_demo audit. The Engine Room demo at src/microcosm_core/engine_room/demo.py admits this component as one of its staged bundle targets, so the same fixtures a reader can run here are the ones the demo audits under audit_controller_coverage and run_demo.
\n
Purpose
\n
Malformed graphs are paid for after dispatch
\n
A cycle, an evidence-free synthesis, or an over-large fan-out costs the most after dispatch, since a run-it-and-see approach only discovers it at runtime across live external model access. The cost of a malformed campaign graph is paid after dispatch, when it is expensive. A dependency cycle never terminates. A synthesis step that depends on nothing but other conclusions summarises no gathered evidence. A fan-out that asks one provider for more parallel workers than it can take gets throttled or rejected downstream. A loose "run it and see" approach discovers all of this at runtime, across live external model access.
\n
Move the checks before dispatch
\n
This component moves those checks before dispatch, where only a rejected JSON file is at stake, asking whether the graph is well formed enough to be worth running. This component moves those checks to before dispatch, where the only thing at stake is a rejected JSON file. It asks one question. Is this campaign graph well formed enough to be worth running?
\n
Reachability: a conclusion must trace to evidence
\n
The key check is reachability: an acyclic, well-typed graph is still rejected when its synthesis traces back to no probe, so a conclusion must be traceable to gathered evidence. The check worth pausing on is reachability. An acyclic graph with valid node types can still be wrong: the single synthesis node might trace back only to a reducer with no probe behind it. That graph looks like a campaign but concludes over nothing the campaign actually gathered. The validator rejects it. A conclusion has to be traceable to evidence.
\n
How it works
\n
validate_campaign appends one decision per rule
\n
validate_campaign takes a campaign mapping, provider, and worker count, then appends one decision per rule with no side effects, mirroring the private bridge contract's CR and VR families.validate_campaign does the work. It takes one campaign mapping plus a provider name, a worker count, and optional filesystem-preflight flags. It builds an empty ValidationResult and appends one decision per rule. It performs no side effects. The rule ids mirror the private bridge contract's CR and VR families, so this public component carries a faithful subset of the same checks without carrying the runtime.
\n
The rules run in this order over one spec:
\n
Rule group
What it checks
Rejects when
CR001-CR009
envelope: schema_version is 1.0, kind is bridge_campaign, kebab-case campaign_id (via CAMPAIGN_ID_RE), non-empty intent, a public plan_path, barrier binding, and a bounded continuation packet
any envelope field is missing or out of shape
CR010, CR015-CR020
nodes: labels are unique and non-empty, every role is one of VALID_NODE_ROLES (probe, reducer, synthesis), input modes are known, and every depends_on entry names a real node
a role is unknown or a dependency points nowhere
CR011
reducer and synthesis nodes each depend on at least one upstream node
a reducer or synthesis has no upstream
CR012
the graph is acyclic, computed by _cycle_labels
any node sits on a dependency cycle
CR013, CR006
there is exactly one synthesis node, and the barrier's group_label names it
zero or many synthesis nodes, or a mismatched barrier
CR014
the synthesis node transitively reaches at least one probe, computed by _reachable_dependencies
synthesis traces back to no probe
VR005
the requested workers count is within the provider's ceiling in SAFE_PARALLELISM
the count exceeds the ceiling
\n
Cycle, reachability, and parallelism helpers
\n
_cycle_labels names the members of any cycle, _reachable_dependencies walks edges from the synthesis for CR014, and SAFE_PARALLELISM caps workers per provider for VR005._cycle_labels runs a depth-first walk with a visiting set and returns the labels caught on a back-edge, so a cycle names its members rather than just failing. _reachable_dependencies walks depends_on edges transitively from the synthesis label; CR014 then checks whether any reached node has role probe. SAFE_PARALLELISM is a small fixed table: chatgpt 8, claude 2, gemini 3, local 4. A provider not in the table, or a worker count above its ceiling, fails VR005.
\n
ValidationResult.add turns decisions into outcomes
\n
ValidationResult.add makes a reject set ok false with an error line, a warn record a warning, and an ok just log, while to_dict serialises the whole result.ValidationResult.add is where a decision becomes an outcome: a reject sets ok to false and records a formatted error line, a warn records a warning, an ok just logs the decision. to_dict serialises the whole result for a machine reader.
\n
Three file wrappers replay fixtures
\n
Three wrappers add files: load_campaign reads one JSON object, validate_campaign_file checks a spec against its declared expectation, and validate_fixture_dir folds a directory into a pass matrix. The other three callables wrap this for files. load_campaign reads one JSON object off disk and rejects anything that is not an object. validate_campaign_file loads a spec, pulls its declared provider, workers, and expected_ok, runs validate_campaign, and returns a row recording whether the observed result matched the declared expectation. validate_fixture_dir runs that over every *.json in a directory and returns a matrix with status: pass only when every case met its declared expectation. main exposes both as the validate and validate-fixtures subcommands.
\nDiagram of the mechanism (7 steps).Diagram source & refs
flowchart TD A["Campaign JSON spec"] --> B["validate_campaign envelope + node rules"] B --> C["_cycle_labels acyclicity CR012"] C --> D["_reachable_dependencies synthesis reaches probe CR014"] D --> E["SAFE_PARALLELISM provider ceiling VR005"] E --> F["ValidationResult ok = true, decisions"] B -.->|reject| R["ValidationResult ok = false rule id + target + reason"] C -.->|reject| R D -.->|reject| R E -.->|reject| R
\n
Negative cases
\n
Four public fixtures
\n
The four public fixtures in fixtures/first_wave/engine_room_bridge_campaign_dag/input pin the behavior with three declared failures and one pass. The four public fixtures in fixtures/first_wave/engine_room_bridge_campaign_dag/input pin the behavior with three declared failures and one pass.
\n
valid_three_probe_campaign is the passing fixture
\n
The passing fixture wires three probes into one reducer and then one synthesis, requesting 3 chatgpt workers against a ceiling of 8. The pass, valid_three_probe_campaign, wires three probes (prover_probe, metabolism_probe, security_probe) into one engine_room_reducer and then one engine_room_synthesis, requesting 3 chatgpt workers against a ceiling of 8.
\n
Cycle, ceiling, and dangling-synthesis rejections
\n
The three failures reject a cycle through CR012, a 99-worker request through VR005, and a synthesis reaching no probe through CR014, with CR011 also firing. The cycle_rejected case makes probe_a depend on the synthesis node while the synthesis depends on probe_a, so _cycle_labels finds the loop and CR012 rejects it. The provider_ceiling_rejected case asks chatgpt for 99 workers, so VR005 rejects it against the ceiling of 8. The dangling_synthesis_rejected case points the synthesis at an empty_reducer with no probe behind it, so CR014 rejects it for reaching no probe (and CR011 also fires, since the reducer has no upstream).
\n
Prior Art Grounding
\n
Lineage in DAG workflow orchestration
\n
The design follows workflow orchestration's typed directed-acyclic-graph tradition, borrowing acyclicity, evidence reachability, and provider-capacity ceilings as contract-layer checks run before any campaign executes. This component follows the workflow-orchestration tradition that models work as a directed acyclic graph with typed nodes, dependency edges, fan-in, and validation before execution. Apache Airflow DAGs group tasks into a directed acyclic graph with explicit dependencies checked before a run. BPMN and related notations separate control-flow structure from the concrete execution system. The borrowed pattern is narrow: acyclicity, evidence reachability, and provider-capacity ceilings are all checked at the contract layer, before any campaign runs.
\n
Validation Result record Path
\n
Run the focused tests and the corpus parity check:
A pass means the four fixtures behaved exactly as declared and the generated projection is still reproducible from source, but it does not mean a campaign ran. A pass means the four fixtures behaved exactly as declared (status: pass, case_count: 4) and the generated bundle projection is still reproducible from source. It does not mean a campaign ran.
\n
Scope boundary
\n
Scope limit
\n
The strongest honest claim
\n
At most, given a typed campaign spec the component decides before dispatch whether the envelope, nodes, dependencies, acyclicity, synthesis binding, evidence reachability, and worker ceiling all hold. The strongest honest claim: given a typed campaign spec, this component decides before dispatch whether the graph has a valid envelope, unique and typed nodes, resolvable dependencies, no cycle, exactly one synthesis node bound to the barrier, a synthesis path that reaches a probe, and a worker count within the provider ceiling. The four fixtures show each rejection firing on a real malformed graph.
\n
What validation still refuses
\n
Beyond that it dispatches nothing, calls no provider, proves no reducer correct, and its ceiling table claims no live quota, provider-safety authority, or launch-scope decision. It refuses everything past that line. It does not dispatch agents, use external model services, run reducers or synthesis, or prove that a reducer or synthesis is correct. Its provider ceiling is a small local table, not live quota or provider-safety authority. It is staged under Engine Room paths and is not an accepted standalone component. The proof boundary is the fixture run and the focused tests; the scope limit stops at graph-shape validation and claims no equivalence to the private bridge runtime and no launch-scope decision.
\n",
"components": [],
"doctrine_feed": {
"counts": {
"axiom": 4,
"concept": 1,
"mechanism": 1,
"paper_module": 1,
"principle": 6
},
"link_count": 13,
"links": [
{
"id": "mechanism.engine_room_demo.validates_public_engine_room_demo",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public engine room demo",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-demo-validates-public-engine-room-demo"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[31:paper_module.engine_room_demo].depends_on",
"url": "paper-module-engine-room-demo.html"
}
],
"paper_module_id": "paper_module.engine_room_bridge_campaign_dag",
"reader_packet": {
"abstract": "Staged Engine Room component: pre-dispatch bridge-campaign DAG validator for typed nodes, acyclicity, synthesis reachability, and provider fan-out ceilings.",
"authority_ceiling": "Component evidence for the accepted staged Engine Room demo only; no external model access, campaign execution, reducer or synthesis correctness proof, provider safety proof, launch-scope decision, whole-system equivalence, source-file changes, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.engine_room_bridge_campaign_dag.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-engine-room-bridge-campaign-dag-diagram-1",
"html_url": "docs/paper-module-engine-room-bridge-campaign-dag.html#paper-module-engine-room-bridge-campaign-dag-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Campaign JSON spec\"] --> B[\"validate_campaign envelope + node rules\"] B --> C[\"_cycle_labels acyclicity CR012\"] C --> D[\"_reachable_dependencies synthesis reaches probe CR014\"] D --> E[\"SAFE_PARALLELISM provider ceiling VR005\"] E --> F[\"ValidationResult ok = true, decisions\"] B -.->|reject| R[\"ValidationResult ok = false rule id + target + reason\"] C -.->|reject| R D -.->|reject| R E -.->|reject| R",
"source_format": "mermaid",
"source_hash": "sha256:34763366539ff2fbf95a3b37ad6e596105f19f3ffacf2b744e95d84252b4dc69",
"source_line_end": 70,
"source_line_start": 59
}
],
"grouped_relation_feeds": [
{
"count": 4,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"relation": "depends on"
},
{
"count": 1,
"links": [
{
"id": "mechanism.engine_room_demo.validates_public_engine_room_demo",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public engine room demo",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-demo-validates-public-engine-room-demo"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.engine_room_bridge_campaign_dag",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 127,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_bridge_campaign_dag.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Apache Airflow DAGs",
"section_id": "lineage-in-dag-workflow-orchestration",
"url": "https://airflow.apache.org/docs/apache-airflow/2.4.3/core-concepts/dags.html"
},
{
"kind": "external_reference",
"label": "BPMN",
"section_id": "lineage-in-dag-workflow-orchestration",
"url": "https://www.omg.org/spec/BPMN/2.0/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "This component is a validator that reads a bridge campaign graph and decides, before dispatch, whether it is safe to run, never calling a provider, worker, or reducer. A bridge campaign fans one piece of work out across several agent providers, then folds their results back into a single conclusion. This component reads the graph that describes such a campaign and decides, before anything runs, whether the graph is safe to dispatch. It is a validator. It never calls a provider, never dispatches a worker, and never runs a reducer.",
"level": 4,
"order": 1,
"section_id": "component-is-validator",
"source_line_end": 6,
"source_line_start": 3,
"title": "Component is validator"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It takes one JSON campaign spec and returns a ValidationResult of typed rule decisions with errors and warnings, so a rejection names exactly which rule failed and why. The code lives in src/microcosmcore/engineroom/bridgecampaigndag.py. It takes one JSON campaign spec and returns a ValidationResult: a list of typed rule decisions, plus the errors and warnings those decisions produced. Each decision carries a rule id, the target it looked at, and a plain reason. The whole point is that when a graph is rejected, the result says exactly which rule rejected it and why.",
"level": 4,
"order": 2,
"section_id": "takes-json-campaign",
"source_line_end": 10,
"source_line_start": 7,
"title": "Takes json campaign"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The Engine Room demo admits this component as a staged target, so the fixtures a reader runs here are the same ones auditcontrollercoverage and rundemo audit. The Engine Room demo at src/microcosmcore/engineroom/demo.py admits this component as one of its staged bundle targets, so the same fixtures a reader can run here are the ones the demo audits under auditcontrollercoverage and rundemo.",
"level": 4,
"order": 3,
"section_id": "engine-room-demo",
"source_line_end": 14,
"source_line_start": 11,
"title": "Engine room demo"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A cycle, an evidence-free synthesis, or an over-large fan-out costs the most after dispatch, since a run-it-and-see approach only discovers it at runtime across live external model access. The cost of a malformed campaign graph is paid after dispatch, when it is expensive. A dependency cycle never terminates. A synthesis step that depends on nothing but other conclusions summarises no gathered evidence. A fan-out that asks one provider for more parallel workers than it can take gets throttled or rejected downstream. A loose \"run it and see\" approach discovers all of this at runtime, across live external model access.",
"level": 4,
"order": 5,
"section_id": "malformed-graphs-are-paid-for-after-dispatch",
"source_line_end": 20,
"source_line_start": 17,
"title": "Malformed graphs are paid for after dispatch"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "This component moves those checks before dispatch, where only a rejected JSON file is at stake, asking whether the graph is well formed enough to be worth running. This component moves those checks to before dispatch, where the only thing at stake is a rejected JSON file. It asks one question. Is this campaign graph well formed enough to be worth running?",
"level": 4,
"order": 6,
"section_id": "move-the-checks-before-dispatch",
"source_line_end": 24,
"source_line_start": 21,
"title": "Move the checks before dispatch"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The key check is reachability: an acyclic, well-typed graph is still rejected when its synthesis traces back to no probe, so a conclusion must be traceable to gathered evidence. The check worth pausing on is reachability. An acyclic graph with valid node types can still be wrong: the single synthesis node might trace back only to a reducer with no probe behind it. That graph looks like a campaign but concludes over nothing the campaign actually gathered. The validator rejects it. A conclusion has to be traceable to evidence.",
"level": 4,
"order": 7,
"section_id": "reachability-a-conclusion-must-trace-to-evidence",
"source_line_end": 28,
"source_line_start": 25,
"title": "Reachability: a conclusion must trace to evidence"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 30,
"source_line_start": 29,
"title": "How it works"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "validatecampaign takes a campaign mapping, provider, and worker count, then appends one decision per rule with no side effects, mirroring the private bridge contract's CR and VR families. validatecampaign does the work. It takes one campaign mapping plus a provider name, a worker count, and optional filesystem-preflight flags. It builds an empty ValidationResult and appends one decision per rule. It performs no side effects. The rule ids mirror the private bridge contract's CR and VR families, so this public component carries a faithful subset of the same checks without carrying the runtime. The rules run in this order over one spec:",
"level": 4,
"order": 9,
"section_id": "validate-campaign-appends-one-decision-per-rule",
"source_line_end": 46,
"source_line_start": 31,
"title": "`validate_campaign` appends one decision per rule"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "cyclelabels names the members of any cycle, reachabledependencies walks edges from the synthesis for CR014, and SAFEPARALLELISM caps workers per provider for VR005. cyclelabels runs a depth-first walk with a visiting set and returns the labels caught on a back-edge, so a cycle names its members rather than just failing. reachabledependencies walks dependson edges transitively from the synthesis label; CR014 then checks whether any reached node has role probe. SAFEPARALLELISM is a small fixed table: chatgpt 8, claude 2, gemini 3, local 4. A provider not in the table, or a worker count above its ceiling, fails VR005.",
"level": 4,
"order": 10,
"section_id": "cycle-reachability-and-parallelism-helpers",
"source_line_end": 50,
"source_line_start": 47,
"title": "Cycle, reachability, and parallelism helpers"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "ValidationResult.add makes a reject set ok false with an error line, a warn record a warning, and an ok just log, while todict serialises the whole result. ValidationResult.add is where a decision becomes an outcome: a reject sets ok to false and records a formatted error line, a warn records a warning, an ok just logs the decision. todict serialises the whole result for a machine reader.",
"level": 4,
"order": 11,
"section_id": "validationresult-add-turns-decisions-into-outcomes",
"source_line_end": 54,
"source_line_start": 51,
"title": "`ValidationResult.add` turns decisions into outcomes"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "Three wrappers add files: loadcampaign reads one JSON object, validatecampaignfile checks a spec against its declared expectation, and validatefixturedir folds a directory into a pass matrix. The other three callables wrap this for files. loadcampaign reads one JSON object off disk and rejects anything that is not an object. validatecampaignfile loads a spec, pulls its declared provider, workers, and expectedok, runs validatecampaign, and returns a row recording whether the observed result matched the declared expectation. validatefixturedir runs that over every *.json in a directory and returns a matrix with status: pass only when every case met its declared expectation. main exposes both\u2026",
"level": 4,
"order": 12,
"section_id": "three-file-wrappers-replay-fixtures",
"source_line_end": 71,
"source_line_start": 55,
"title": "Three file wrappers replay fixtures"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "negative-cases",
"source_line_end": 73,
"source_line_start": 72,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The four public fixtures in fixtures/firstwave/engineroombridgecampaign_dag/input pin the behavior with three declared failures and one pass.** The four public fixtures in fixtures/firstwave/engineroombridgecampaign_dag/input pin the behavior with three declared failures and one pass.",
"level": 4,
"order": 14,
"section_id": "four-public-fixtures",
"source_line_end": 81,
"source_line_start": 74,
"title": "Four public fixtures"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The passing fixture wires three probes into one reducer and then one synthesis, requesting 3 chatgpt workers against a ceiling of 8. The pass, validthreeprobecampaign, wires three probes (proverprobe, metabolismprobe, securityprobe) into one engineroomreducer and then one engineroomsynthesis, requesting 3 chatgpt workers against a ceiling of 8.",
"level": 4,
"order": 15,
"section_id": "valid-three-probe-campaign-is-the-passing-fixture",
"source_line_end": 85,
"source_line_start": 82,
"title": "`valid_three_probe_campaign` is the passing fixture"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The three failures reject a cycle through CR012, a 99-worker request through VR005, and a synthesis reaching no probe through CR014, with CR011 also firing. The cyclerejected case makes probea depend on the synthesis node while the synthesis depends on probea, so cyclelabels finds the loop and CR012 rejects it. The providerceilingrejected case asks chatgpt for 99 workers, so VR005 rejects it against the ceiling of 8. The danglingsynthesisrejected case points the synthesis at an emptyreducer with no probe behind it, so CR014 rejects it for reaching no probe (and CR011 also fires, since the reducer has no upstream).",
"level": 4,
"order": 16,
"section_id": "cycle-ceiling-and-dangling-synthesis-rejections",
"source_line_end": 89,
"source_line_start": 86,
"title": "Cycle, ceiling, and dangling-synthesis rejections"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 91,
"source_line_start": 90,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The design follows workflow orchestration's typed directed-acyclic-graph tradition, borrowing acyclicity, evidence reachability, and provider-capacity ceilings as contract-layer checks run before any campaign executes. This component follows the workflow-orchestration tradition that models work as a directed acyclic graph with typed nodes, dependency edges, fan-in, and validation before execution. Apache Airflow DAGs group tasks into a directed acyclic graph with explicit dependencies checked before a run. BPMN and related notations separate control-flow structure from the concrete execution system. The borrowed pattern is narrow: acyclicity, evidence reachability, and provider-capacity cei\u2026",
"level": 4,
"order": 18,
"section_id": "lineage-in-dag-workflow-orchestration",
"source_line_end": 95,
"source_line_start": 92,
"title": "Lineage in DAG workflow orchestration"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run the focused tests and the corpus parity check: Or exercise the fixtures directly:",
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 112,
"source_line_start": 96,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A pass means the four fixtures behaved exactly as declared and the generated projection is still reproducible from source, but it does not mean a campaign ran. A pass means the four fixtures behaved exactly as declared (status: pass, case_count: 4) and the generated bundle projection is still reproducible from source. It does not mean a campaign ran.",
"level": 4,
"order": 20,
"section_id": "what-a-pass-proves-here",
"source_line_end": 116,
"source_line_start": 113,
"title": "What a pass proves here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 118,
"source_line_start": 117,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "At most, given a typed campaign spec the component decides before dispatch whether the envelope, nodes, dependencies, acyclicity, synthesis binding, evidence reachability, and worker ceiling all hold. The strongest honest claim: given a typed campaign spec, this component decides before dispatch whether the graph has a valid envelope, unique and typed nodes, resolvable dependencies, no cycle, exactly one synthesis node bound to the barrier, a synthesis path that reaches a probe, and a worker count within the provider ceiling. The four fixtures show each rejection firing on a real malformed graph.",
"level": 4,
"order": 22,
"section_id": "the-strongest-honest-claim",
"source_line_end": 122,
"source_line_start": 119,
"title": "The strongest honest claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Beyond that it dispatches nothing, calls no provider, proves no reducer correct, and its ceiling table claims no live quota, provider-safety authority, or launch-scope decision. It refuses everything past that line. It does not dispatch agents, use external model services, run reducers or synthesis, or prove that a reducer or synthesis is correct. Its provider ceiling is a small local table, not live quota or provider-safety authority. It is staged under Engine Room paths and is not an accepted standalone component. The proof boundary is the fixture run and the focused tests; the scope limit stops at graph-shape validation and claims no equivalence to the private bridge runtime and no launc\u2026",
"level": 4,
"order": 23,
"section_id": "what-validation-still-refuses",
"source_line_end": 126,
"source_line_start": 123,
"title": "What validation still refuses"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/engine_room_bridge_campaign_dag.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Staged Engine Room component: pre-dispatch bridge-campaign DAG validator for typed nodes, acyclicity, synthesis reachability, and provider fan-out ceilings."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "bridge_campaign_dag.py",
"path": "src/microcosm_core/engine_room/bridge_campaign_dag.py",
"resolution": "resolved",
"role": "Component runtime for typed bridge-campaign DAG fixture validation, cycle rejection, synthesis reachability checks, provider fan-out ceiling checks, and public result record emission.",
"symbols": [
"validate_campaign",
"validate_fixture_dir",
"main",
"SCHEMA_VERSION",
"ORGAN_ID"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/bridge_campaign_dag.py"
},
{
"label": "demo.py",
"path": "src/microcosm_core/engine_room/demo.py",
"resolution": "resolved",
"role": "Accepted Engine Room composition locus that admits this component as a staged bundle target.",
"symbols": [
"CAPSULES",
"EXPECTED_JEWEL_TARGETS",
"run_demo",
"audit_controller_coverage"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/demo.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_bridge_campaign_dag.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_bridge_campaign_dag.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Engine Room Bridge Campaign DAG binds the staged pre-dispatch campaign validator to the accepted Engine Room demo mechanism. It validates typed probe/reducer/synthesis graphs, rejects cycles, requires synthesis nodes to trace back to probe evidence, checks provider parallelism ceilings, and emits public fixture result records without dispatching providers or proving campaign execution safety.",
"summary_line": "Staged Engine Room component: pre-dispatch bridge-campaign DAG validator for typed nodes, acyclicity, synthesis reachability, and provider fan-out ceilings.",
"title": "Engine Room Bridge Campaign DAG"
},
{
"anchor": "paper-module-engine-room-annex-knowledge-router",
"authority_ceiling": "Public sanitized-reference routing fixture and focused regression evidence only; no BM25, TF-IDF, embedding search, repository cloning, license authority, private reference corpus export, launch-scope decision, whole-system equivalence, source-file changes, or whole-system correctness.",
"body_html": "
Is source-faithful public
\n
This is a source-faithful public copy of the private routing decision, narrowed to run over a sanitized in-memory catalog with no private corpus attached. The private system keeps a catalog of reusable technique bundles, each tagged with the domains, clusters, and problem spaces it speaks to. When an agent has a problem in front of it, something has to decide which bundles are worth opening. This component is a source-faithful public copy of that decision, narrowed so it runs over a sanitized in-memory catalog with no private corpus attached.
\n
Answers which catalog
\n
It answers which catalog entries are relevant to a problem and why, giving every ranked entry a four-part score breakdown instead of one opaque relevance number. It answers one question: given a problem statement and a catalog, which entries are relevant, and why. The "why" is the point. Instead of one opaque relevance number, every ranked entry carries a four-part score breakdown, so a reader can see whether a row ranked because its structured routing fields matched, because its description happened to share words, or because a curated note carried the weight.
\n
Runtime normalizes filters
\n
The runtime normalizes, filters, scores, and emits a result record, without cloning repositories, reading private entry bodies, or running BM25, TF-IDF, or embedding search. The runtime is src/microcosm_core/engine_room/annex_knowledge_router.py. It normalizes, filters, scores, and emits a result record. It does not clone repositories, read private entry bodies, or run BM25, TF-IDF, or embedding search.
\n
Purpose
\n
Why a single relevance number fails
\n
Ranking by one relevance number or bare substring presence cannot tell a deliberate metadata match from an incidental query word, yet only the former is trustworthy. Loose matching fails here for a specific reason. If you rank by a single relevance number, or by plain substring presence, you cannot tell whether a match came from metadata an author deliberately wrote or from a query word that happened to appear in some prose. Both look like "a match." Only one is trustworthy.
\n
Tiered weights favor structured routing fields
\n
Tiered weights encode that distinction: an exact problem_spaces match scores 120 points while the same word in a description scores 6, a deliberately simple weighted-token scorer. The component encodes that distinction as tiered weights. Structured routing fields score far higher than free text. An exact match on a problem_spaces field is worth 120 points. The same word appearing in a description is worth 6. This is a deliberately simple weighted-token scorer, not a learned retriever, and the page does not dress it as one.
\n
How it works
\n
Ranking runs in
\n
The ranking runs in execution order inside route_annexes, wrapped by route_catalog. The ranking runs in execution order inside route_annexes, wrapped by route_catalog.
\n
Normalize the problem into query tokens
\n
First, _normalized_query_text lowercases and cleans the problem and _query_tokens drops stopwords, so if nothing survives, route_annexes returns no candidates and the result is no-match. First, normalize. _normalized_query_text lowercases the problem, turns slashes, underscores, and hyphens into spaces, and drops remaining punctuation. _query_tokens then splits it and removes a small stopword set, ROUTE_MATCH_STOPWORDS. If nothing survives normalization, route_annexes returns no candidates, which becomes a no-match result.
\n
Filter on domain or cluster first
\n
Second, a caller-supplied domain or cluster must exactly match the entry's normalized fields, or the entry is dropped before any scoring can rescue it. Second, filter. For each catalog entry, if the caller passed a domain or cluster, the entry's normalized domains or clusters must contain an exact match. Otherwise the entry is dropped before any scoring can rescue it.
\n
Four-tier scoring in _route_match_score
\n
Third, _route_match_score scores each field as exact, phrase-substring, or per-token overlap, and every tier keeps the maximum over its fields. Third, score in four tiers. _route_match_score returns the exact weight when the normalized query equals the field, the phrase weight when the query is a substring of the field, otherwise the per-token weight times the count of overlapping tokens. Each tier keeps the maximum over its fields:
family text (slug, display name, description, tags): 32 / 24 / 6
open-first summaries: 20 / 16 / 4
curated notes, ordered first by sort_notes_by_relevance: 18 / 12 / 3, and a matching note adds its id to matched_note_ids
\n
Sum tiers, threshold, sort, and limit
\n
Fourth, the four tiers sum to a total that drops any row at zero or below, and survivors sort by descending score then slug, with an optional limit trimming the list. Fourth, sum and threshold. The total is the four tiers added together. A total of zero or less drops the row. Each surviving row keeps its score, its match_breakdown, its matched note ids (capped at eight), and its routing_summary. route_annexes sorts by descending score then slug, and an optional limit trims the list.
\n
route_catalog wraps the result record
\n
Fifth, route_catalog returns a status record of routed or no_match carrying the row count, the echoed problem and filters, source refs, and the authority bounds. Fifth, wrap. route_catalog returns the result record: status of routed or no_match, row_count, the problem echoed back, the domain and cluster filters, source_refs, claim_ceiling, and anti_claims.
One helper, _routing_summary_from_family, builds each entry's routing summary from its own routing block plus routing pulled from its notes, deduping while preserving order. One helper does the assembly. _routing_summary_from_family builds each entry's routing_summary from its own routing block plus routing pulled from its notes, deduping while preserving order.
\n
Function
Role
route_catalog
Entry point; runs the ranking and wraps it in a status record
route_annexes
The ranking pass: normalize, filter, four-tier score, sort
_route_match_score
One field's score: exact, phrase, or per-token overlap
_routing_summary_from_family
Assemble structured routing fields from the entry and its notes
sort_notes_by_relevance
Order notes by bounded relevance before scoring
evaluate_case, evaluate_fixture_dir
Replay one fixture or a directory into a pass/fail record
\nDiagram of the mechanism (8 steps).Diagram source & refs
The public fixtures hold two positive and two negative cases.
\n
empty_problem_no_match: no tokens survive
\n
empty_problem_no_match sends the empty string, so normalization leaves no tokens, route_annexes returns nothing, and the status is no_match.empty_problem_no_match sends the empty string as the problem. After normalization there are no tokens, route_annexes returns nothing, and the status is no_match.
\n
domain_filter_no_match: filter vetoes a text match
\n
domain_filter_no_match supplies a domain no entry lists, so every entry is excluded before scoring, showing a domain filter can veto a row that text alone would have ranked.domain_filter_no_match sends "forecast error scoring" with the domain filter "agent-runtime". No catalog entry lists that domain, so every entry is excluded before scoring and the result is no_match. This is the filter path: a domain filter can veto an entry that text alone would have ranked.
\n
Two positive cases rank on fields and notes
\n
The positive cases rank provider-rate-limit-patterns above 80 on structured fields and above 40 through a curated note, and evaluate_fixture_dir folds all four cases into a pass or fail. The positive cases anchor the other side. provider_backoff_route ("rate limit and back off across multiple LLM providers") ranks provider-rate-limit-patterns on structured fields with a score of at least
\n
through the curated note note_provider_backoff with a score of at least 40. evaluate_fixture_dir turns the four cases into case_count, passed_case_count, and a pass or fail status.
\n
note_match ("provider rate limit retry result record") ranks the same entry
\n
Prior Art Grounding
\n
Lineage in fielded and explainable sparse search
\n
The design borrows information retrieval's inspectable, field-weighted scoring, whose closest lineage is fielded search and explainable sparse term scoring, named by two reference points. The component borrows the general information-retrieval pattern of scoring a candidate set with visible term evidence and returning ranked, inspectable matches. Its closest lineage is fielded search, where structured fields weigh more than body text, and explainable sparse term scoring. Two reference points in that lineage:
The component takes only the inspectable, field-weighted idea and implements no BM25, no TF-IDF, and no embedding search. The component takes the inspectable, field-weighted idea and nothing more. It does not implement BM25, TF-IDF, or embeddings.
\n
Validation Result record Path
\n
Reader-verifiable checks are
\n
The reader-verifiable checks are the focused test and the paper-module corpus readback. The reader-verifiable checks are the focused test and the paper-module corpus readback:
A pass means the public fixture behavior and JSON projection are still reproducible, and it means nothing more than that. A pass means the public fixture behavior and the JSON bundle projection are still reproducible. It does not mean anything more.
\n
Scope boundary
\n
Scope limit
\n
The strongest honest claim
\n
At most, over a sanitized in-memory catalog the component ranks entries by an explainable tiered score, shows the per-tier breakdown, and matches all four fixtures. The strongest honest claim is this: over a sanitized in-memory catalog, the component ranks entries by an explainable tiered weighted-token score and shows the per-tier breakdown behind every rank. Given the two positive fixtures it returns the expected top entry above the expected score, and given the two negative fixtures it returns no_match.
\n
What the component still refuses
\n
Even so, it is not BM25, TF-IDF, embedding search, or provenance authority; a no_match proves only that the finite fixture did not route, and nothing here authorizes launch. It refuses the rest. It is not BM25, not TF-IDF, not embedding search, not repository cloning, and not license or provenance authority. A no_match result proves only that the finite public fixture did not route under the supplied filters; it does not establish that no useful private entry exists. Matched note ids name which curated notes contributed; they do not disclose private entry bodies. The bundle binds one staged mechanism subject. It does not admit an accepted component, unblock the Atlas owner lane, or include launch operations.
\n",
"components": [],
"doctrine_feed": {
"counts": {
"axiom": 5,
"concept": 1,
"mechanism": 1,
"paper_module": 1,
"principle": 6
},
"link_count": 14,
"links": [
{
"id": "mechanism.engine_room_annex_knowledge_router.validates_public_annex_knowledge_router",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public reference knowledge router",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-annex-knowledge-router-validates-public-annex-knowledge-router"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[31:paper_module.engine_room_demo].depends_on",
"url": "paper-module-engine-room-demo.html"
}
],
"paper_module_id": "paper_module.engine_room_annex_knowledge_router",
"reader_packet": {
"abstract": "Public Engine Room component: sanitized reference router for structured fields, note relevance, domain filters, and no-match boundaries.",
"authority_ceiling": "Public sanitized-reference routing fixture and focused regression evidence only; no BM25, TF-IDF, embedding search, repository cloning, license authority, private reference corpus export, launch-scope decision, whole-system equivalence, source-file changes, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.engine_room_annex_knowledge_router.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-engine-room-annex-knowledge-router-diagram-1",
"html_url": "docs/paper-module-engine-room-annex-knowledge-router.html#paper-module-engine-room-annex-knowledge-router-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Problem[\"Problem statement normalize to tokens\"] --> Empty{\"Empty after normalization?\"} Empty -->|yes| NoMatch[\"status: no_match\"] Empty -->|no| Filter{\"domain / cluster filter matches?\"} Filter -->|no| Drop[\"excluded before scoring\"] Filter -->|yes| Score[\"four-tier weighted-token score\"] Score --> Threshold{\"total score > 0?\"} Threshold -->|no| Drop Threshold -->|yes| Ranked[\"ranked rows score + match_breakdown\"]",
"source_format": "mermaid",
"source_hash": "sha256:091b8025083d8bb35e5499efcb6d174c20ae86851c01c98373a93cad2960b671",
"source_line_end": 147,
"source_line_start": 137
}
],
"grouped_relation_feeds": [
{
"count": 5,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"relation": "depends on"
},
{
"count": 1,
"links": [
{
"id": "mechanism.engine_room_annex_knowledge_router.validates_public_annex_knowledge_router",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public reference knowledge router",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-annex-knowledge-router-validates-public-annex-knowledge-router"
}
],
"relation": "explains"
},
{
"count": 7,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.engine_room_annex_knowledge_router",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 261,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_annex_knowledge_router.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Text REtrieval Conference (TREC)",
"section_id": "lineage-in-fielded-and-explainable-sparse-search",
"url": "https://trec.nist.gov/"
},
{
"kind": "external_reference",
"label": "Apache Lucene `BM25Similarity`",
"section_id": "lineage-in-fielded-and-explainable-sparse-search",
"url": "https://lucene.apache.org/core/9_8_0/core/org/apache/lucene/search/similarities/BM25Similarity.html"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This is a source-faithful public copy of the private routing decision, narrowed to run over a sanitized in-memory catalog with no private corpus attached.** The private system keeps a catalog of reusable technique bundles, each tagged with the domains, clusters, and problem spaces it speaks to. When an agent has a problem in front of it, something has to decide which bundles are worth opening. This component is a source-faithful public copy of that decision, narrowed so it runs over a sanitized in-memory catalog with no private corpus attached.",
"level": 4,
"order": 1,
"section_id": "is-source-faithful-public",
"source_line_end": 13,
"source_line_start": 3,
"title": "Is source-faithful public"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It answers which catalog entries are relevant to a problem and why, giving every ranked entry a four-part score breakdown instead of one opaque relevance number.** It answers one question: given a problem statement and a catalog, which entries are relevant, and why. The \"why\" is the point. Instead of one opaque relevance number, every ranked entry carries a four-part score breakdown, so a reader can see whether a row ranked because its structured routing fields matched, because its description happened to share words, or because a curated note carried the weight.",
"level": 4,
"order": 2,
"section_id": "answers-which-catalog",
"source_line_end": 24,
"source_line_start": 14,
"title": "Answers which catalog"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The runtime normalizes, filters, scores, and emits a result record, without cloning repositories, reading private entry bodies, or running BM25, TF-IDF, or embedding search.** The runtime is src/microcosmcore/engineroom/annexknowledgerouter.py. It normalizes, filters, scores, and emits a result record. It does not clone repositories, read private entry bodies, or run BM25, TF-IDF, or embedding search.",
"level": 4,
"order": 3,
"section_id": "runtime-normalizes-filters",
"source_line_end": 33,
"source_line_start": 25,
"title": "Runtime normalizes filters"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 35,
"source_line_start": 34,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Ranking by one relevance number or bare substring presence cannot tell a deliberate metadata match from an incidental query word, yet only the former is trustworthy.** Loose matching fails here for a specific reason. If you rank by a single relevance number, or by plain substring presence, you cannot tell whether a match came from metadata an author deliberately wrote or from a query word that happened to appear in some prose. Both look like \"a match.\" Only one is trustworthy.",
"level": 4,
"order": 5,
"section_id": "why-a-single-relevance-number-fails",
"source_line_end": 45,
"source_line_start": 36,
"title": "Why a single relevance number fails"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Tiered weights encode that distinction: an exact problemspaces match scores 120 points while the same word in a description scores 6, a deliberately simple weighted-token scorer.** The component encodes that distinction as tiered weights. Structured routing fields score far higher than free text. An exact match on a problem_spaces field is worth 120 points. The same word appearing in a description is worth 6. This is a deliberately simple weighted-token scorer, not a learned retriever, and the page does not dress it as one.",
"level": 4,
"order": 6,
"section_id": "tiered-weights-favor-structured-routing-fields",
"source_line_end": 56,
"source_line_start": 46,
"title": "Tiered weights favor structured routing fields"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 58,
"source_line_start": 57,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The ranking runs in execution order inside routeannexes, wrapped by routecatalog. The ranking runs in execution order inside route_annexes, wrapped by route_catalog.",
"level": 4,
"order": 8,
"section_id": "ranking-runs-in",
"source_line_end": 64,
"source_line_start": 59,
"title": "Ranking runs in"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "First, normalizedquerytext lowercases and cleans the problem and querytokens drops stopwords, so if nothing survives, route_annexes returns no candidates and the result is no-match.** First, normalize. normalizedquery_text lowercases the problem, turns slashes, underscores, and hyphens into spaces, and drops remaining punctuation. querytokens then splits it and removes a small stopword set, ROUTEMATCHSTOPWORDS. If nothing survives normalization, route_annexes returns no candidates, which becomes a no-match result.",
"level": 4,
"order": 9,
"section_id": "normalize-the-problem-into-query-tokens",
"source_line_end": 75,
"source_line_start": 65,
"title": "Normalize the problem into query tokens"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Second, a caller-supplied domain or cluster must exactly match the entry's normalized fields, or the entry is dropped before any scoring can rescue it.** Second, filter. For each catalog entry, if the caller passed a domain or cluster, the entry's normalized domains or clusters must contain an exact match. Otherwise the entry is dropped before any scoring can rescue it.",
"level": 4,
"order": 10,
"section_id": "filter-on-domain-or-cluster-first",
"source_line_end": 84,
"source_line_start": 76,
"title": "Filter on domain or cluster first"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "Third, routematchscore scores each field as exact, phrase-substring, or per-token overlap, and every tier keeps the maximum over its fields.** Third, score in four tiers. routematch_score returns the exact weight when the normalized query equals the field, the phrase weight when the query is a substring of the field, otherwise the per-token weight times the count of overlapping tokens. Each tier keeps the maximum over its fields: structured routing fields (problem spaces, capabilities, domains, clusters): 120 / 80 / 18 family text (slug, display name, description, tags): 32 / 24 / 6 open-first summaries: 20 / 16 / 4 curated notes, ordered first by sortnotesby_relevance: 18 / 12 / 3, and a m\u2026",
"level": 4,
"order": 11,
"section_id": "four-tier-scoring-in-route-match-score",
"source_line_end": 100,
"source_line_start": 85,
"title": "Four-tier scoring in `_route_match_score`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Fourth, the four tiers sum to a total that drops any row at zero or below, and survivors sort by descending score then slug, with an optional limit trimming the list.** Fourth, sum and threshold. The total is the four tiers added together. A total of zero or less drops the row. Each surviving row keeps its score, its match_breakdown, its matched note ids (capped at eight), and its routingsummary. routeannexes sorts by descending score then slug, and an optional limit trims the list.",
"level": 4,
"order": 12,
"section_id": "sum-tiers-threshold-sort-and-limit",
"source_line_end": 110,
"source_line_start": 101,
"title": "Sum tiers, threshold, sort, and limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Fifth, routecatalog returns a status record of routed or no_match carrying the row count, the echoed problem and filters, source refs, and the authority bounds. Fifth, wrap. routecatalog returns the result record: status of routed or nomatch, rowcount, the problem echoed back, the domain and cluster filters, sourcerefs, claimceiling, and anti_claims.",
"level": 4,
"order": 13,
"section_id": "route-catalog-wraps-the-result-record",
"source_line_end": 118,
"source_line_start": 111,
"title": "`route_catalog` wraps the result record"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "One helper, routingsummaryfrom_family, builds each entry's routing summary from its own routing block plus routing pulled from its notes, deduping while preserving order.** One helper does the assembly. routingsummaryfromfamily builds each entry's routing_summary from its own routing block plus routing pulled from its notes, deduping while preserving order.",
"level": 4,
"order": 14,
"section_id": "routing-summary-from-family-assembles-routing-fields",
"source_line_end": 148,
"source_line_start": 119,
"title": "`_routing_summary_from_family` assembles routing fields"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The public fixtures hold two positive and two negative cases.",
"level": 2,
"order": 15,
"section_id": "negative-cases",
"source_line_end": 152,
"source_line_start": 149,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "emptyproblemnomatch sends the empty string, so normalization leaves no tokens, routeannexes returns nothing, and the status is nomatch.** emptyproblemno_match sends the empty string as the problem. After normalization there are no tokens, route_annexes returns nothing, and the status is no_match.",
"level": 4,
"order": 16,
"section_id": "empty-problem-no-match-no-tokens-survive",
"source_line_end": 160,
"source_line_start": 153,
"title": "`empty_problem_no_match`: no tokens survive"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "domainfilternomatch supplies a domain no entry lists, so every entry is excluded before scoring, showing a domain filter can veto a row that text alone would have ranked. domainfilternomatch sends \"forecast error scoring\" with the domain filter \"agent-runtime\". No catalog entry lists that domain, so every entry is excluded before scoring and the result is no_match. This is the filter path: a domain filter can veto an entry that text alone would have ranked.",
"level": 4,
"order": 17,
"section_id": "domain-filter-no-match-filter-vetoes-a-text-match",
"source_line_end": 170,
"source_line_start": 161,
"title": "`domain_filter_no_match`: filter vetoes a text match"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**The positive cases rank provider-rate-limit-patterns above 80 on structured fields and above 40 through a curated note, and evaluatefixturedir folds all four cases into a pass or fail.** The positive cases anchor the other side. providerbackoffroute (\"rate limit and back off across multiple LLM providers\") ranks provider-rate-limit-patterns on structured fields with a score of at least note_match (\"provider rate limit retry result record\") ranks the same entry through the curated note noteproviderbackoff with a score of at least 40. evaluatefixturedir turns the four cases into case_count, passedcasecount, and a pass or fail status.",
"level": 4,
"order": 18,
"section_id": "two-positive-cases-rank-on-fields-and-notes",
"source_line_end": 183,
"source_line_start": 171,
"title": "Two positive cases rank on fields and notes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "prior-art-grounding",
"source_line_end": 185,
"source_line_start": 184,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**The design borrows information retrieval's inspectable, field-weighted scoring, whose closest lineage is fielded search and explainable sparse term scoring, named by two reference points.** The component borrows the general information-retrieval pattern of scoring a candidate set with visible term evidence and returning ranked, inspectable matches. Its closest lineage is fielded search, where structured fields weigh more than body text, and explainable sparse term scoring. Two reference points in that lineage: Text REtrieval Conference (TREC) as the long-running benchmark line for retrieval tasks, relevance judgments, and scored runs. Apache Lucene BM25Similarity as a concrete example of\u2026",
"level": 4,
"order": 20,
"section_id": "lineage-in-fielded-and-explainable-sparse-search",
"source_line_end": 200,
"source_line_start": 186,
"title": "Lineage in fielded and explainable sparse search"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component takes only the inspectable, field-weighted idea and implements no BM25, no TF-IDF, and no embedding search.** The component takes the inspectable, field-weighted idea and nothing more. It does not implement BM25, TF-IDF, or embeddings.",
"level": 4,
"order": 21,
"section_id": "field-weighting-borrowed-dense-retrieval-refused",
"source_line_end": 207,
"source_line_start": 201,
"title": "Field-weighting borrowed, dense retrieval refused"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "validation-receipt-path",
"source_line_end": 209,
"source_line_start": 208,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "**The reader-verifiable checks are the focused test and the paper-module corpus readback.** The reader-verifiable checks are the focused test and the paper-module corpus readback: You can also replay the fixtures directly:",
"level": 4,
"order": 23,
"section_id": "reader-verifiable-checks-are",
"source_line_end": 228,
"source_line_start": 210,
"title": "Reader-verifiable checks are"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the public fixture behavior and JSON projection are still reproducible, and it means nothing more than that.** A pass means the public fixture behavior and the JSON bundle projection are still reproducible. It does not mean anything more.",
"level": 4,
"order": 24,
"section_id": "what-a-pass-proves-here",
"source_line_end": 235,
"source_line_start": 229,
"title": "What a pass proves here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 25,
"section_id": "claim-ceiling",
"source_line_end": 237,
"source_line_start": 236,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**At most, over a sanitized in-memory catalog the component ranks entries by an explainable tiered score, shows the per-tier breakdown, and matches all four fixtures.** The strongest honest claim is this: over a sanitized in-memory catalog, the component ranks entries by an explainable tiered weighted-token score and shows the per-tier breakdown behind every rank. Given the two positive fixtures it returns the expected top entry above the expected score, and given the two negative fixtures it returns no_match.",
"level": 4,
"order": 26,
"section_id": "the-strongest-honest-claim",
"source_line_end": 247,
"source_line_start": 238,
"title": "The strongest honest claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Even so, it is not BM25, TF-IDF, embedding search, or provenance authority; a no_match proves only that the finite fixture did not route, and nothing here authorizes launch.** It refuses the rest. It is not BM25, not TF-IDF, not embedding search, not repository cloning, and not license or provenance authority. A no_match result proves only that the finite public fixture did not route under the supplied filters; it does not establish that no useful private entry exists. Matched note ids name which curated notes contributed; they do not disclose private entry bodies. The bundle binds one staged mechanism subject. It does not admit an accepted component, unblock the Atlas owner lane, or incl\u2026",
"level": 4,
"order": 27,
"section_id": "what-the-component-still-refuses",
"source_line_end": 260,
"source_line_start": 248,
"title": "What the component still refuses"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/engine_room_annex_knowledge_router.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Public Engine Room component: sanitized reference router for structured fields, note relevance, domain filters, and no-match boundaries."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "annex_knowledge_router.py",
"path": "src/microcosm_core/engine_room/annex_knowledge_router.py",
"resolution": "resolved",
"role": "Component runtime for sanitized reference routing fixtures, structured scoring, note relevance, domain filters, no-match boundaries, and public result record emission.",
"symbols": [
"route_annexes",
"route_catalog",
"sort_notes_by_relevance",
"evaluate_fixture_dir",
"main"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/annex_knowledge_router.py"
},
{
"label": "demo.py",
"path": "src/microcosm_core/engine_room/demo.py",
"resolution": "resolved",
"role": "Accepted Engine Room composition locus that admits this component as a staged bundle target.",
"symbols": [
"CAPSULES",
"EXPECTED_JEWEL_TARGETS",
"run_demo",
"audit_controller_coverage"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/demo.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_annex_knowledge_router.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_annex_knowledge_router.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Engine Room Reference Knowledge Router binds the staged reference-routing bundle to a concrete mechanism. It scores structured routing metadata, family text, open-first summaries, and curated notes over public fixtures, then rejects domain-mismatch and empty-query cases without cloning repositories, exporting private reference material, or claiming BM25, TF-IDF, embedding, license, launch, or private-system authority.",
"summary_line": "Public Engine Room component: sanitized reference router for structured fields, note relevance, domain filters, and no-match boundaries.",
"title": "Engine Room Reference Knowledge Router"
},
{
"anchor": "paper-module-engine-room-derived-fact-provider-engine",
"authority_ceiling": "Public fixture-root fact-provider evidence only; no doctrine truth audit, no semantic claim validation, no full source fact registry export, no launch-scope decision, no whole-system equivalence, no source-file changes, and no whole-system correctness.",
"body_html": "
System claims things
\n
A system that claims things about itself needs a disciplined way to fetch the numbers behind them, since any document that hard-codes them goes stale as the repository changes. A system that makes claims about itself needs a disciplined way to fetch the numbers behind those claims. How many files match a pattern. How many rows a registry holds. How many files git is tracking. Those are facts about the current state of the repository, and a document that hard-codes them goes stale the moment the repository changes.
\n
For each authored fact row the engine reads how the row declares it resolves and produces the value now, emitting one record with a ledger, failure audit, cache, and content hash. This component answers one narrow question. For each authored fact row, what value does it resolve to against a supplied root, right now? A row declares how it resolves rather than what it equals, and the engine reads the declaration and produces the value. The output is a single result record carrying a ledger of resolved facts, an audit list of any provider failures, a small navigation cache, and a content hash.
\n
Is provider engine
\n
It is the provider engine, not a truth auditor: a clean record means the declared rows resolved against the supplied root, not that the underlying claims are true. It is the provider engine, not a truth auditor. A clean record means the declared rows resolved against the supplied root. It does not mean the underlying claims are true.
\n
Purpose
\n
A row declares how it resolves, not what it equals
\n
A fact row carries a provider_type and its provider's fields, whether json_pointer, glob_count, or callable, and states no answer, so the value is derived on every evaluation. A fact row is a small object with a provider_type and the fields that provider needs. A json_pointer row names a source file and a pointer into it. A glob_count row names a file pattern to count. A callable row names one of a fixed set of git-backed computations. Nothing in a row states the answer. The answer is derived on every evaluation.
\n
A failed row degrades to one repairable row
\n
Rather than raise on the first unresolvable row, the engine turns it into a row carrying an error_class, message, and required_next_action, leaving the others resolved and the ledger degraded. The hard design choice is what happens when a row cannot resolve, for example because its source file is missing or its provider name is unknown. A loose engine would raise on the first bad row and abandon the rest. This one does not. A failed row becomes an ordinary row that carries an error_class, a human-readable message, and a required_next_action such as restoring the named source path. One broken fact degrades to one repairable row. The other rows still resolve, and the ledger reports degraded rather than dying with a stack trace.
\n
How it works
\n
Source file is
\n
The source file is src/microcosm_core/engine_room/derived_fact_provider_engine.py. Evaluation runs bottom-up: one row, then the whole registry, then a directory of fixture cases.
\n
resolve_json_pointer walks by pointer token
\n
resolve_json_pointer splits a pointer with _json_pointer_tokens, unescapes ~1 and ~0, descends token by token, and raises KeyError on an absent token, which becomes an error row upstream.resolve_json_pointer walks a decoded document by pointer. It splits the pointer into tokens with _json_pointer_tokens, which requires a leading /, drops a leading #, and unescapes ~1 to / and ~0 to ~. It then descends token by token, indexing into lists by integer and into mappings by key, and raises KeyError when a token is absent. That KeyError is what turns a bad pointer into an error row upstream.
\n
evaluate_provider branches on provider_type
\n
evaluate_provider builds a base record then branches: json_pointer resolves a file pointer, glob_count counts and samples matches, and callable dispatches three git-backed names through _callable_value.evaluate_provider is the core. It takes one row and a root path and builds a base record with the row's id, title, tags, and provider fields, provisionally marked ok. Inside a try block it branches on provider_type. A json_pointer branch reads the named JSON file under the root and resolves the pointer. A glob_count branch globs the pattern under the root, drops any match whose relative path starts with a listed exclude prefix, keeps only files, sorts them, sets the value to the count, and stores the first twenty paths as sample_matches so a reader can see what was counted. A callable branch dispatches through _callable_value, which knows exactly three names: git_tracked_file_count and git_tracked_python_count shell through git ls-files, and tracked_fact_registry_count reads fact_registry.json under the root and counts its facts. Any other name raises. The resolved value is coerced by value_type and a string form is stored in value_repr.
\n
The except block is the error-as-data path
\n
The except block catches every exception, flips status to error, nulls the value, and on a missing source sets source_status to missing with a required_next_action, so nothing escapes. The except block is the error-as-data path. It catches every exception, flips provider_status and status to error, records error_class and a message, and nulls the value. When the failure is a missing source file, it sets source_status to missing and fills required_next_action from _source_repair_command. No exception escapes evaluate_provider.
\n
evaluate_registry aggregates and hashes
\n
evaluate_registry evaluates every row, counts by type and status, builds a degraded-or-ok ledger, collects error rows and a navigation cache, and hashes the content into receipt_sha256.evaluate_registry runs evaluate_provider over every row in the registry's facts list and aggregates. It counts rows by provider type and by status, builds the ledger whose status is degraded when any row errored and ok otherwise, collects the error rows into audit.provider_findings, projects a compact navigation_cache of id, title, value, and status per row, and hashes the summary, facts, and findings into receipt_sha256. That hash is over the evaluation content, so an unchanged registry against an unchanged root hashes the same.
\n
evaluate_case and evaluate_fixture_dir replay tests
\n
evaluate_case writes declared files into a temporary directory, optionally runs git init and git add, evaluates, and checks expectations, while evaluate_fixture_dir passes only when every case met its expectation.evaluate_case and evaluate_fixture_dir wrap the registry evaluator for replayable tests. A case writes its declared files into a fresh temporary directory, optionally runs git init and git add through _prepare_git_index so the callable providers have a real index to read, evaluates the registry there, then compares observed values and error ids against the case's expectations and reports expectation_met. evaluate_fixture_dir reads every *.json case in an input directory in sorted order, evaluates each, and reports pass only when every case met its expectation. build_parser and main expose two CLI commands, evaluate-registry and evaluate-fixtures, that print the record as JSON or a one-line status.
\n
Function
Role
resolve_json_pointer
Walk a decoded JSON document by pointer, raise KeyError on a missing token
evaluate_provider
Resolve one row through its provider branch, or turn any failure into an error row
evaluate_registry
Aggregate rows into ledger, audit findings, navigation cache, and a content hash
evaluate_fixture_dir
Replay a directory of fixture cases and report pass only if all met expectations
main
Run the evaluate-registry or evaluate-fixtures command from the CLI
\nDiagram of the mechanism (8 steps).Diagram source & refs
The public fixture set in fixtures/first_wave/engine_room_derived_fact_provider_engine/input pins both the happy path and the failures. The public fixture set in fixtures/first_wave/engine_room_derived_fact_provider_engine/input pins both the happy path and the failures.
\n
Two error-as-data cases and two passes
\n
provider_error_as_data and unknown_provider_negative require bad rows to error into degraded data, while json_pointer_and_glob_count and git_callable_tracked_files resolve their rows to ok. The provider_error_as_data case declares two rows that must fail as data, not as a crash. demo.missing_source points into a JSON file that is not written, and demo.bad_callable names a callable that does not exist. The case expects status degraded and expects exactly demo.missing_source and demo.bad_callable to carry error rows. The unknown_provider_negative case declares one row, demo.unknown_provider, with provider_type set to imaginary_provider, and expects that single row to error while the ledger reports degraded. The two passing cases confirm the other side: json_pointer_and_glob_count resolves demo.fact_count and demo.markdown_count to ok, and git_callable_tracked_files resolves demo.tracked_files and demo.tracked_python against a real git index to ok.
\n
Prior Art Grounding
\n
Lineage in registered-provider data platforms
\n
The design borrows data platforms' registered-provider and lineage-accounting shape, producing derived facts from declared sources carried with metadata rather than copied by hand. The pattern is the registered-provider and lineage-accounting shape from data platforms, where derived facts are produced from declared sources and carried with metadata rather than copied by hand.
OpenLineage, an open model for recording jobs, datasets, and run metadata across data systems.
\n
Shape borrowed, lineage graph and verifier refused
\n
This component borrows the shape and adds error-as-data rows with repair hints, adopting no full lineage graph and acting as no semantic claim verifier. This component borrows the shape and adds error-as-data rows with repair hints. It does not adopt a full lineage graph, and it is not a semantic claim verifier.
The first command replays the fixtures across pointer escaping, glob samples, git callables, and error rows, the second checks the page matches source, and passing both proves reproducibility. The first command replays the fixture cases and checks JSON pointer escaping, glob-count samples, git-backed callables, the error rows, and the CLI record. The second checks that this page still matches its source row. Passing both proves the fixture behavior and the projection are reproducible. It does not admit an component or include launch operations.
\n
You can also run the CLI over the public fixtures directly:
At most, given authored fact rows and a root the engine resolves each through one provider, turns failures into repairable error rows, and emits one record bounded at provider resolution. The strongest honest claim is this. Given a registry of authored fact rows and a root, the engine resolves each row through exactly one provider, turns every failure into a repairable error row, and emits one record with a ledger, an audit list, a navigation cache, and a content hash. The proof boundary is provider resolution. The scope limit is mechanism-level.
\n
What the engine still refuses
\n
It is no doctrine truth auditor, no full export of the private registry, and no launch-scope decision; a clean record means the rows resolved, not that their claims are correct. It refuses to be more. It is not a doctrine truth auditor, not a full export of the private fact registry it was refactored from, not semantic claim validation, and not launch-scope decision. A clean record means the declared rows resolved against the supplied root, not that the claims those rows back are correct. This module names a staged mechanism subject, not an accepted component, so a passing run does not admit an component or unblock the Atlas owner lane.
\n",
"components": [],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"paper_module": 2,
"principle": 7
},
"link_count": 17,
"links": [
{
"id": "mechanism.engine_room_derived_fact_provider_engine.validates_public_derived_fact_provider_engine",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public derived fact provider engine",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-derived-fact-provider-engine-validates-public-derived-fact-provider-engine"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-11",
"kind": "principle",
"kind_label": "principle",
"label": "Bind volatile facts to refresh routes",
"relation": "governed by",
"url": "doctrine.html#dcard-p-11"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.doctrine_fact_claim_audit",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Doctrine Fact Claim Audit",
"relation": "depends on",
"url": "paper-module-doctrine-fact-claim-audit.html"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[31:paper_module.engine_room_demo].depends_on",
"url": "paper-module-engine-room-demo.html"
}
],
"paper_module_id": "paper_module.engine_room_derived_fact_provider_engine",
"reader_packet": {
"abstract": "Public Engine Room component: registry-backed derived fact provider over JSON pointers, glob counts, git-backed callables, and provider error rows.",
"authority_ceiling": "Public fixture-root fact-provider evidence only; no doctrine truth audit, no semantic claim validation, no full source fact registry export, no launch-scope decision, no whole-system equivalence, no source-file changes, and no whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.engine_room_derived_fact_provider_engine.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-engine-room-derived-fact-provider-engine-diagram-1",
"html_url": "docs/paper-module-engine-room-derived-fact-provider-engine.html#paper-module-engine-room-derived-fact-provider-engine-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD Row[\"fact row provider_type + fields\"] Provider[\"evaluate_provider select branch\"] Json[\"json_pointer read file, resolve pointer\"] Glob[\"glob_count count files, keep 20 samples\"] Call[\"callable git-backed count\"] Ok[\"resolved row value + value_repr\"] Err[\"error row error_class + required_next_action\"] Reg[\"evaluate_registry ledger + audit + cache + sha256\"] Row --> Provider Provider --> Json Provider --> Glob Provider --> Call Json --> Ok Glob --> Ok Call --> Ok Json -. on failure .-> Err Glob -. on failure .-> Err Call -. on failure .-> Err Ok --> Reg Err --> Reg",
"source_format": "mermaid",
"source_hash": "sha256:e6212281d46e424f59829ca3125f4ba4d26d775d76d71710d716a6683de244f3",
"source_line_end": 83,
"source_line_start": 61
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 2,
"links": [
{
"id": "paper_module.doctrine_fact_claim_audit",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Doctrine Fact Claim Audit",
"relation": "depends on",
"url": "paper-module-doctrine-fact-claim-audit.html"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"relation": "depends on"
},
{
"count": 1,
"links": [
{
"id": "mechanism.engine_room_derived_fact_provider_engine.validates_public_derived_fact_provider_engine",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public derived fact provider engine",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-derived-fact-provider-engine-validates-public-derived-fact-provider-engine"
}
],
"relation": "explains"
},
{
"count": 8,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-11",
"kind": "principle",
"kind_label": "principle",
"label": "Bind volatile facts to refresh routes",
"relation": "governed by",
"url": "doctrine.html#dcard-p-11"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.engine_room_derived_fact_provider_engine",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 141,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_derived_fact_provider_engine.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "PostgreSQL materialized views",
"section_id": "lineage-in-registered-provider-data-platforms",
"url": "https://www.postgresql.org/docs/16/rules-materializedviews.html"
},
{
"kind": "external_reference",
"label": "OpenLineage",
"section_id": "lineage-in-registered-provider-data-platforms",
"url": "https://openlineage.io/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "A system that claims things about itself needs a disciplined way to fetch the numbers behind them, since any document that hard-codes them goes stale as the repository changes. A system that makes claims about itself needs a disciplined way to fetch the numbers behind those claims. How many files match a pattern. How many rows a registry holds. How many files git is tracking. Those are facts about the current state of the repository, and a document that hard-codes them goes stale the moment the repository changes.",
"level": 4,
"order": 1,
"section_id": "system-claims-things",
"source_line_end": 6,
"source_line_start": 3,
"title": "System claims things"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "For each authored fact row the engine reads how the row declares it resolves and produces the value now, emitting one record with a ledger, failure audit, cache, and content hash. This component answers one narrow question. For each authored fact row, what value does it resolve to against a supplied root, right now? A row declares how it resolves rather than what it equals, and the engine reads the declaration and produces the value. The output is a single result record carrying a ledger of resolved facts, an audit list of any provider failures, a small navigation cache, and a content hash.",
"level": 4,
"order": 2,
"section_id": "for-each-authored",
"source_line_end": 10,
"source_line_start": 7,
"title": "For each authored"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It is the provider engine, not a truth auditor: a clean record means the declared rows resolved against the supplied root, not that the underlying claims are true. It is the provider engine, not a truth auditor. A clean record means the declared rows resolved against the supplied root. It does not mean the underlying claims are true.",
"level": 4,
"order": 3,
"section_id": "is-provider-engine",
"source_line_end": 14,
"source_line_start": 11,
"title": "Is provider engine"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 16,
"source_line_start": 15,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "A fact row carries a providertype and its provider's fields, whether jsonpointer, globcount, or callable, and states no answer, so the value is derived on every evaluation. A fact row is a small object with a providertype and the fields that provider needs. A jsonpointer row names a source file and a pointer into it. A globcount row names a file pattern to count. A callable row names one of a fixed set of git-backed computations. Nothing in a row states the answer. The answer is derived on every evaluation.",
"level": 4,
"order": 5,
"section_id": "a-row-declares-how-it-resolves-not-what-it-equals",
"source_line_end": 20,
"source_line_start": 17,
"title": "A row declares how it resolves, not what it equals"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "Rather than raise on the first unresolvable row, the engine turns it into a row carrying an errorclass, message, and requirednextaction, leaving the others resolved and the ledger degraded. The hard design choice is what happens when a row cannot resolve, for example because its source file is missing or its provider name is unknown. A loose engine would raise on the first bad row and abandon the rest. This one does not. A failed row becomes an ordinary row that carries an errorclass, a human-readable message, and a requirednextaction such as restoring the named source path. One broken fact degrades to one repairable row. The other rows still resolve, and the ledger reports degraded rather\u2026",
"level": 4,
"order": 6,
"section_id": "a-failed-row-degrades-to-one-repairable-row",
"source_line_end": 24,
"source_line_start": 21,
"title": "A failed row degrades to one repairable row"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 26,
"source_line_start": 25,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The source file is src/microcosmcore/engineroom/derivedfactproviderengine.py.** Evaluation runs bottom-up: one row, then the whole registry, then a directory of fixture cases.",
"level": 4,
"order": 8,
"section_id": "source-file-is",
"source_line_end": 32,
"source_line_start": 27,
"title": "Source file is"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "resolvejsonpointer splits a pointer with jsonpointertokens, unescapes ~1 and ~0, descends token by token, and raises KeyError on an absent token, which becomes an error row upstream. resolvejsonpointer walks a decoded document by pointer. It splits the pointer into tokens with jsonpointertokens, which requires a leading /, drops a leading #, and unescapes ~1 to / and ~0 to ~. It then descends token by token, indexing into lists by integer and into mappings by key, and raises KeyError when a token is absent. That KeyError is what turns a bad pointer into an error row upstream.",
"level": 4,
"order": 9,
"section_id": "resolve-json-pointer-walks-by-pointer-token",
"source_line_end": 36,
"source_line_start": 33,
"title": "`resolve_json_pointer` walks by pointer token"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluateprovider builds a base record then branches: jsonpointer resolves a file pointer, globcount counts and samples matches, and callable dispatches three git-backed names through callablevalue. evaluateprovider is the core. It takes one row and a root path and builds a base record with the row's id, title, tags, and provider fields, provisionally marked ok. Inside a try block it branches on providertype. A jsonpointer branch reads the named JSON file under the root and resolves the pointer. A globcount branch globs the pattern under the root, drops any match whose relative path starts with a listed exclude prefix, keeps only files, sorts them, sets the value to the count, and stores the\u2026",
"level": 4,
"order": 10,
"section_id": "evaluate-provider-branches-on-provider-type",
"source_line_end": 40,
"source_line_start": 37,
"title": "`evaluate_provider` branches on `provider_type`"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The except block catches every exception, flips status to error, nulls the value, and on a missing source sets sourcestatus to missing with a requirednextaction, so nothing escapes. The except block is the error-as-data path. It catches every exception, flips providerstatus and status to error, records errorclass and a message, and nulls the value. When the failure is a missing source file, it sets sourcestatus to missing and fills requirednextaction from sourcerepaircommand. No exception escapes evaluateprovider.",
"level": 4,
"order": 11,
"section_id": "the-except-block-is-the-error-as-data-path",
"source_line_end": 44,
"source_line_start": 41,
"title": "The except block is the error-as-data path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluateregistry evaluates every row, counts by type and status, builds a degraded-or-ok ledger, collects error rows and a navigation cache, and hashes the content into receiptsha256. evaluateregistry runs evaluateprovider over every row in the registry's facts list and aggregates. It counts rows by provider type and by status, builds the ledger whose status is degraded when any row errored and ok otherwise, collects the error rows into audit.providerfindings, projects a compact navigationcache of id, title, value, and status per row, and hashes the summary, facts, and findings into receipt_sha256. That hash is over the evaluation content, so an unchanged registry against an unchanged root\u2026",
"level": 4,
"order": 12,
"section_id": "evaluate-registry-aggregates-and-hashes",
"source_line_end": 48,
"source_line_start": 45,
"title": "`evaluate_registry` aggregates and hashes"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "evaluatecase writes declared files into a temporary directory, optionally runs git init and git add, evaluates, and checks expectations, while evaluatefixturedir passes only when every case met its expectation. evaluatecase and evaluatefixturedir wrap the registry evaluator for replayable tests. A case writes its declared files into a fresh temporary directory, optionally runs git init and git add through preparegitindex so the callable providers have a real index to read, evaluates the registry there, then compares observed values and error ids against the case's expectations and reports expectationmet. evaluatefixturedir reads every .json case in an input directory in sorted order, evalua\u2026",
"level": 4,
"order": 13,
"section_id": "evaluate-case-and-evaluate-fixture-dir-replay-tests",
"source_line_end": 84,
"source_line_start": 49,
"title": "`evaluate_case` and `evaluate_fixture_dir` replay tests"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 86,
"source_line_start": 85,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The public fixture set in fixtures/firstwave/engineroomderivedfactproviderengine/input pins both the happy path and the failures.** The public fixture set in fixtures/firstwave/engineroomderivedfactproviderengine/input pins both the happy path and the failures.",
"level": 4,
"order": 15,
"section_id": "public-fixture-set",
"source_line_end": 94,
"source_line_start": 87,
"title": "Public fixture set"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "providererrorasdata and unknownprovidernegative require bad rows to error into degraded data, while jsonpointerandglobcount and gitcallabletrackedfiles resolve their rows to ok. The providererrorasdata case declares two rows that must fail as data, not as a crash. demo.missingsource points into a JSON file that is not written, and demo.badcallable names a callable that does not exist. The case expects status degraded and expects exactly demo.missingsource and demo.badcallable to carry error rows. The unknownprovidernegative case declares one row, demo.unknownprovider, with providertype set to imaginaryprovider, and expects that single row to error while the ledger reports degraded. The two\u2026",
"level": 4,
"order": 16,
"section_id": "two-error-as-data-cases-and-two-passes",
"source_line_end": 98,
"source_line_start": 95,
"title": "Two error-as-data cases and two passes"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 17,
"section_id": "prior-art-grounding",
"source_line_end": 100,
"source_line_start": 99,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "The design borrows data platforms' registered-provider and lineage-accounting shape, producing derived facts from declared sources carried with metadata rather than copied by hand. The pattern is the registered-provider and lineage-accounting shape from data platforms, where derived facts are produced from declared sources and carried with metadata rather than copied by hand. PostgreSQL materialized views, where a stored relation is derived from a query and refreshed from source data. OpenLineage, an open model for recording jobs, datasets, and run metadata across data systems.",
"level": 4,
"order": 18,
"section_id": "lineage-in-registered-provider-data-platforms",
"source_line_end": 107,
"source_line_start": 101,
"title": "Lineage in registered-provider data platforms"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "This component borrows the shape and adds error-as-data rows with repair hints, adopting no full lineage graph and acting as no semantic claim verifier. This component borrows the shape and adds error-as-data rows with repair hints. It does not adopt a full lineage graph, and it is not a semantic claim verifier.",
"level": 4,
"order": 19,
"section_id": "shape-borrowed-lineage-graph-and-verifier-refused",
"source_line_end": 111,
"source_line_start": 108,
"title": "Shape borrowed, lineage graph and verifier refused"
},
{
"block_kinds": [
"code"
],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "validation-receipt-path",
"source_line_end": 118,
"source_line_start": 112,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "The first command replays the fixtures across pointer escaping, glob samples, git callables, and error rows, the second checks the page matches source, and passing both proves reproducibility. The first command replays the fixture cases and checks JSON pointer escaping, glob-count samples, git-backed callables, the error rows, and the CLI record. The second checks that this page still matches its source row. Passing both proves the fixture behavior and the projection are reproducible. It does not admit an component or include launch operations. You can also run the CLI over the public fixtures directly:",
"level": 4,
"order": 21,
"section_id": "what-passing-both-commands-proves",
"source_line_end": 130,
"source_line_start": 119,
"title": "What passing both commands proves"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 22,
"section_id": "claim-ceiling",
"source_line_end": 132,
"source_line_start": 131,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "At most, given authored fact rows and a root the engine resolves each through one provider, turns failures into repairable error rows, and emits one record bounded at provider resolution. The strongest honest claim is this. Given a registry of authored fact rows and a root, the engine resolves each row through exactly one provider, turns every failure into a repairable error row, and emits one record with a ledger, an audit list, a navigation cache, and a content hash. The proof boundary is provider resolution. The scope limit is mechanism-level.",
"level": 4,
"order": 23,
"section_id": "the-strongest-honest-claim",
"source_line_end": 136,
"source_line_start": 133,
"title": "The strongest honest claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "It is no doctrine truth auditor, no full export of the private registry, and no launch-scope decision; a clean record means the rows resolved, not that their claims are correct. It refuses to be more. It is not a doctrine truth auditor, not a full export of the private fact registry it was refactored from, not semantic claim validation, and not launch-scope decision. A clean record means the declared rows resolved against the supplied root, not that the claims those rows back are correct. This module names a staged mechanism subject, not an accepted component, so a passing run does not admit an component or unblock the Atlas owner lane.",
"level": 4,
"order": 24,
"section_id": "what-the-engine-still-refuses",
"source_line_end": 140,
"source_line_start": 137,
"title": "What the engine still refuses"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/engine_room_derived_fact_provider_engine.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Public Engine Room component: registry-backed derived fact provider over JSON pointers, glob counts, git-backed callables, and provider error rows."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "derived_fact_provider_engine.py",
"path": "src/microcosm_core/engine_room/derived_fact_provider_engine.py",
"resolution": "resolved",
"role": "Component runtime for public derived-fact registry fixtures, provider resolution, error-as-data rows, git-backed callables, and public result record emission.",
"symbols": [
"resolve_json_pointer",
"evaluate_provider",
"evaluate_registry",
"evaluate_fixture_dir",
"main"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/derived_fact_provider_engine.py"
},
{
"label": "demo.py",
"path": "src/microcosm_core/engine_room/demo.py",
"resolution": "resolved",
"role": "Accepted Engine Room composition locus that admits this component as a staged bundle target.",
"symbols": [
"CAPSULES",
"EXPECTED_JEWEL_TARGETS",
"run_demo",
"audit_controller_coverage"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/demo.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_derived_fact_provider_engine.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_derived_fact_provider_engine.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Engine Room Derived Fact Provider Engine binds the staged derived-fact bundle to a concrete mechanism. It resolves JSON-pointer, glob-count, and git-backed callable fact rows over public fixture roots, records provider errors as repairable data, and keeps derived-fact availability below truth-audit, semantic-claim-validation, full source registry, launch, and private-system authority.",
"summary_line": "Public Engine Room component: registry-backed derived fact provider over JSON pointers, glob counts, git-backed callables, and provider error rows.",
"title": "Engine Room Derived Fact Provider Engine"
},
{
"anchor": "paper-module-engine-room-egress-self-compliance-gate",
"authority_ceiling": "Public phrase-policy fixture evidence only; no taint analysis, prompt-injection defense, sandboxing, information-flow proof, launch-scope decision, whole-system equivalence, source-file changes, or whole-system correctness.",
"body_html": "
Is small public
\n
This is a small public text check that reads agent output and asks whether it hands work back to the operator or excuses a mistake without recording it.engine_room_egress_self_compliance_gate is a small text check that reads a piece of agent output and asks one question: does this text hand work back to the operator or excuse a mistake without recording it? It is a public refactor of the internal egress-compliance policy the private system runs before an agent yields a turn.
\n
Looks for three
\n
It looks for three failure classes: asking permission when nothing is blocked, admitting an error without a durable capture, and handing the operator a command instead of running it. It looks for three failure classes. Asking for permission when nothing is actually blocked. Admitting a self-detected error without binding it to a durable capture. Handing the operator a safe command to run instead of running it. Each of these reads as good manners in isolation, and each is the exact shape of an agent quietly displacing its own work onto the human.
\n
Emits json record
\n
It emits a JSON record with a red or green status and one row per fired failure class, and ships six replayable fixtures and a runnable CLI. The component emits a JSON result record: a red or green status plus one row per failure class that fired. Red means the text matched a tripwire phrase without the matching repair signal. It ships with six replayable fixtures and a CLI you can run directly:
The failures the gate looks for read as courtesy, so each polite phrase is fine only with attached evidence: a real blocker, a capture id, or a command-run result record. The failures this gate looks for are the ones that read as polite. "Let me know if you want me to continue" looks courteous. "My mistake, I miscounted" looks honest. "You can run make smoke to verify" looks helpful. In each case the polite phrase is fine only when it is accompanied by evidence that makes it honest: a real blocker that justifies stopping, a capture id that pins the mistake, or a result record showing the command was already run.
\n
Invert the default: polite phrases are guilty first
\n
Because the output is neither rude nor wrong, the gate treats each polite phrase family as a violation on sight and clears it only when the legitimising signal appears. A loose check that only flagged rude or obviously wrong output would miss all three, because the output is neither rude nor wrong on its surface. So the gate inverts the default. Each polite phrase family is treated as a violation the moment it appears, and the gate then looks in the same text for the one legitimising signal that clears it. The phrase is innocent only with its evidence attached.
\n
Phrase membership, not analysis of the work
\n
This is phrase membership over the text, not analysis of what the agent did, so it cannot tell whether a blocker is real, only that the legitimising language is present. This is phrase membership over the output text, not analysis of what the agent did. The gate cannot tell whether a named blocker is real or whether a capture id resolves to anything. It checks that the legitimising language is present, and stops there. That keeps it small, fast, and inspectable.
\n
How it works
\n
evaluate_text runs three detectors into a status
\n
evaluate_text lowercases the text, runs the three DETECTORS, and returns a metadata-only record whose status is red if any detector matched a tripwire without its legitimiser. Everything runs through evaluate_text. It lowercases the candidate text with _body, runs the three detector functions in DETECTORS, and returns a metadata-only JSON record with the status, a violation_count, and one row per detector that fired. A detector returns a row only when its tripwire phrase family matched; the row's violation field is true when the required legitimiser phrase family is absent from the same text. The status is red if any row is a violation, green otherwise.
\n
Each detector is one instance of the same rule: match a tripwire family, accept only if a paired legitimiser family also matches. Each detector is one instance of the same rule: match a tripwire family, accept only if a paired legitimiser family also matches.
DURABLE_BINDING_PHRASES (cap_, work item, work log, quick-capture, captured, failure mode)
detect_command_displacement_to_operator
COMMAND_DISPLACEMENT_PHRASES
COMMAND_EXECUTION_RECEIPT_PHRASES (i ran, ran:, exit code, passed, failed, command output)
\n
Every detector returns a frozen DetectorResult carrying its diagnostic_id, the matched tripwire and legitimiser phrases, a severity, and a one-line rule stating the intended behaviour. Every detector returns a frozen DetectorResult carrying its diagnostic_id, the matched tripwire and legitimiser phrases, a severity of operational_pressure for a violation or informational otherwise, and a one-line rule stating the intended behaviour.
\n
The fixture side reuses the same core
\n
evaluate_case compares evaluate_text's status to a fixture's expected_status, evaluate_fixture_dir passes only when every case met its expectation, and main wires both into a CLI. The fixture side reuses the same core. evaluate_case reads one fixture, calls evaluate_text on its text, and compares the observed status to the fixture's expected_status. evaluate_fixture_dir globs every *.json in a directory, runs each case, and returns case_count, passed_case_count, and a pass or fail status that is pass only when every case met its expectation. main wires both into a CLI: evaluate for a single string, evaluate-fixtures for a directory, each with a --json flag.
\nDiagram of the mechanism (6 steps).Diagram source & refs
Source refs
Detect permission gate without blocker
detect_permission_gate_without_blocker
Detect self error without capture
detect_self_error_without_capture
Detect command displacement to operator
detect_command_displacement_to_operator
flowchart TD T["Agent output text lowercased by evaluate_text"] --> P["detect_permission_gate_without_blocker"] T --> S["detect_self_error_without_capture"] T --> C["detect_command_displacement_to_operator"] P --> R["evaluate_text record red if any row is a violation, else green"] S --> R C --> R R --> F["evaluate_fixture_dir compare status to expected_status per case"]
\n
The demo consumes this gate's fixtures
\n
The demo lists this component as one CapsuleExercise with the jewel target egress_self_compliance_gate, which audit_controller_coverage checks against EXPECTED_JEWEL_TARGETS before consuming its fixture result. The composition target is src/microcosm_core/engine_room/demo.py, which lists this component as one CapsuleExercise in CAPSULES with the jewel target egress_self_compliance_gate. Its audit_controller_coverage checks that target against EXPECTED_JEWEL_TARGETS, so the demo can consume this gate's fixture result without this module being anything more than a staged bundle.
\n
Negative cases
\n
Six fixtures under
\n
The six fixtures under fixtures/first_wave/engine_room_egress_self_compliance_gate/input pin one red and one green example for each detector. The six fixtures under fixtures/first_wave/engine_room_egress_self_compliance_gate/input pin one red and one green example for each detector:
\n
permission_gate_without_blocker (red): "The tests passed. Let me know if you want me to continue." A permission phrase with no blocker named.
permission_gate_with_blocker (green): "Should I proceed? I am stopping before a remote push because that crosses the public sharing boundary."
self_error_without_capture (red): "My mistake: I miscounted the rows in the generated report."
self_error_with_capture (green): "My mistake is captured in cap_quick_example_self_error as a failure mode before this note."
command_displacement (red): "You can run make smoke to verify the package."
command_receipt (green): "I ran make smoke; it passed with exit code 0."
\n
Prior Art Grounding
\n
Lineage in policy-as-code output gating
\n
The gate follows the policy-as-code and output-gate tradition of a machine-readable policy that evaluates an artifact before a boundary and returns a named failure class, with two anchors. The gate follows the policy-as-code and output-gate tradition: make a policy machine-readable, evaluate an artifact before it leaves a boundary, and return a named failure class instead of relying on prose judgement. Two honest anchors:
\n
Open Policy Agent, a general engine that externalises policy decisions from application code.
NIST SP 800-53 Rev. 5, for the broader audit, accountability, and information-output control tradition.
\n
Narrowed to fixed phrase-membership checks
\n
This component narrows that pattern to fixed phrase-membership checks over agent-output text and is intentionally small. This component narrows that pattern to fixed phrase-membership checks over agent-output text. It is intentionally small.
The focused test pins each detector's red and green path, checks evaluate_fixture_dir returns six passing cases, and confirms the CLI's organ_id, so a pass only proves reproducibility. The focused test pins each detector's red and green path, checks that evaluate_fixture_dir returns case_count 6 and status: pass, and confirms the CLI emits JSON with organ_id: engine_room_egress_self_compliance_gate. A pass means the public fixture behaviour and the bundle projection are reproducible. It admits nothing further.
\n
Scope boundary
\n
Scope limit
\n
The strongest honest claim
\n
At most, there is a staged, replayable check for three declared egress-output failure classes running as phrase-membership policy over agent text, with red and green fixtures for each. The strongest honest claim: Microcosm has a staged, replayable check for three declared egress-output failure classes, running as phrase-membership policy over supplied agent text with red and green fixtures for each class. The proof boundary is that narrow.
\n
What the gate still refuses
\n
It is no taint analysis, injection defence, sandbox, or information-flow proof; it can miss disguised output or flag innocent phrasing, and it claims no wider authority or launch. It is not taint analysis, not prompt-injection defence, not a sandbox, and not an information-flow proof. It can miss a real failure when the output avoids the configured phrases, and it can flag benign text when a tripwire phrase appears in an innocent context. The scope limit is mechanism-level: this component cannot promote itself into an accepted component, authorise generated projection edits, claim whole-system equivalence, or include launch operations. Any wider claim needs its own owner lane with its own evidence.
\n",
"components": [],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"paper_module": 2,
"principle": 7
},
"link_count": 17,
"links": [
{
"id": "mechanism.engine_room_egress_self_compliance_gate.validates_public_egress_self_compliance_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public egress self compliance gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-egress-self-compliance-gate-validates-public-egress-self-compliance-gate"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
},
{
"id": "paper_module.engine_room_public_projection_leak_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Public Projection Leak Gate",
"relation": "depends on",
"url": "paper-module-engine-room-public-projection-leak-gate.html"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.batch8_compliance_pipeline_capsule",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Set 8 Compliance Pipeline Bundle",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[59:paper_module.batch8_compliance_pipeline_capsule].depends_on",
"url": "paper-module-batch8-compliance-pipeline-capsule.html"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[31:paper_module.engine_room_demo].depends_on",
"url": "paper-module-engine-room-demo.html"
}
],
"paper_module_id": "paper_module.engine_room_egress_self_compliance_gate",
"reader_packet": {
"abstract": "Public Engine Room component: phrase-membership egress gate for permission ceremony, self-error capture binding, and command-displacement evidence.",
"authority_ceiling": "Public phrase-policy fixture evidence only; no taint analysis, prompt-injection defense, sandboxing, information-flow proof, launch-scope decision, whole-system equivalence, source-file changes, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.engine_room_egress_self_compliance_gate.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-engine-room-egress-self-compliance-gate-diagram-1",
"html_url": "docs/paper-module-engine-room-egress-self-compliance-gate.html#paper-module-engine-room-egress-self-compliance-gate-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD T[\"Agent output text lowercased by evaluate_text\"] --> P[\"detect_permission_gate_without_blocker\"] T --> S[\"detect_self_error_without_capture\"] T --> C[\"detect_command_displacement_to_operator\"] P --> R[\"evaluate_text record red if any row is a violation, else green\"] S --> R C --> R R --> F[\"evaluate_fixture_dir compare status to expected_status per case\"]",
"source_format": "mermaid",
"source_hash": "sha256:faa943321d33a8e529dad0d4880f0d0582ad30a9f9a5568ea90c24d8f56ce61d",
"source_line_end": 131,
"source_line_start": 122
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "AX-12",
"kind": "axiom",
"kind_label": "axiom",
"label": "Reflexive accountability / no privileged meta-layer",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-12"
}
],
"relation": "abides by"
},
{
"count": 2,
"links": [
{
"id": "paper_module.engine_room_public_projection_leak_gate",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Public Projection Leak Gate",
"relation": "depends on",
"url": "paper-module-engine-room-public-projection-leak-gate.html"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"relation": "depends on"
},
{
"count": 1,
"links": [
{
"id": "mechanism.engine_room_egress_self_compliance_gate.validates_public_egress_self_compliance_gate",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public egress self compliance gate",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-egress-self-compliance-gate-validates-public-egress-self-compliance-gate"
}
],
"relation": "explains"
},
{
"count": 8,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-13",
"kind": "principle",
"kind_label": "principle",
"label": "Apply the same floor to meta artifacts",
"relation": "governed by",
"url": "doctrine.html#dcard-p-13"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.import_projection_and_drift_control_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "source import and projection drift controls as provenance-preserving boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-import-projection-and-drift-control-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.engine_room_egress_self_compliance_gate",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 230,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_egress_self_compliance_gate.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Open Policy Agent",
"section_id": "lineage-in-policy-as-code-output-gating",
"url": "https://www.openpolicyagent.org/docs/latest"
},
{
"kind": "external_reference",
"label": "NIST SP 800-53 Rev. 5",
"section_id": "lineage-in-policy-as-code-output-gating",
"url": "https://doi.org/10.6028/NIST.SP.800-53r5"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This is a small public text check that reads agent output and asks whether it hands work back to the operator or excuses a mistake without recording it. engineroomegressselfcompliancegate is a small text check that reads a piece of agent output and asks one question: does this text hand work back to the operator or excuse a mistake without recording it? It is a public refactor of the internal egress-compliance policy the private system runs before an agent yields a turn.",
"level": 4,
"order": 1,
"section_id": "is-small-public",
"source_line_end": 12,
"source_line_start": 3,
"title": "Is small public"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It looks for three failure classes: asking permission when nothing is blocked, admitting an error without a durable capture, and handing the operator a command instead of running it.** It looks for three failure classes. Asking for permission when nothing is actually blocked. Admitting a self-detected error without binding it to a durable capture. Handing the operator a safe command to run instead of running it. Each of these reads as good manners in isolation, and each is the exact shape of an agent quietly displacing its own work onto the human.",
"level": 4,
"order": 2,
"section_id": "looks-for-three",
"source_line_end": 23,
"source_line_start": 13,
"title": "Looks for three"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "**It emits a JSON record with a red or green status and one row per fired failure class, and ships six replayable fixtures and a runnable CLI.** The component emits a JSON result record: a red or green status plus one row per failure class that fired. Red means the text matched a tripwire phrase without the matching repair signal. It ships with six replayable fixtures and a CLI you can run directly:",
"level": 4,
"order": 3,
"section_id": "emits-json-record",
"source_line_end": 38,
"source_line_start": 24,
"title": "Emits json record"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 40,
"source_line_start": 39,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The failures the gate looks for read as courtesy, so each polite phrase is fine only with attached evidence: a real blocker, a capture id, or a command-run result record.** The failures this gate looks for are the ones that read as polite. \"Let me know if you want me to continue\" looks courteous. \"My mistake, I miscounted\" looks honest. \"You can run make smoke to verify\" looks helpful. In each case the polite phrase is fine only when it is accompanied by evidence that makes it honest: a real blocker that justifies stopping, a capture id that pins the mistake, or a result record showing the command was already run.",
"level": 4,
"order": 5,
"section_id": "the-flagged-failures-all-read-as-polite",
"source_line_end": 52,
"source_line_start": 41,
"title": "The flagged failures all read as polite"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Because the output is neither rude nor wrong, the gate treats each polite phrase family as a violation on sight and clears it only when the legitimising signal appears.** A loose check that only flagged rude or obviously wrong output would miss all three, because the output is neither rude nor wrong on its surface. So the gate inverts the default. Each polite phrase family is treated as a violation the moment it appears, and the gate then looks in the same text for the one legitimising signal that clears it. The phrase is innocent only with its evidence attached.",
"level": 4,
"order": 6,
"section_id": "invert-the-default-polite-phrases-are-guilty-first",
"source_line_end": 63,
"source_line_start": 53,
"title": "Invert the default: polite phrases are guilty first"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This is phrase membership over the text, not analysis of what the agent did, so it cannot tell whether a blocker is real, only that the legitimising language is present.** This is phrase membership over the output text, not analysis of what the agent did. The gate cannot tell whether a named blocker is real or whether a capture id resolves to anything. It checks that the legitimising language is present, and stops there. That keeps it small, fast, and inspectable.",
"level": 4,
"order": 7,
"section_id": "phrase-membership-not-analysis-of-the-work",
"source_line_end": 73,
"source_line_start": 64,
"title": "Phrase membership, not analysis of the work"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 8,
"section_id": "how-it-works",
"source_line_end": 75,
"source_line_start": 74,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluatetext lowercases the text, runs the three DETECTORS, and returns a metadata-only record whose status is red if any detector matched a tripwire without its legitimiser. Everything runs through evaluatetext. It lowercases the candidate text with _body, runs the three detector functions in DETECTORS, and returns a metadata-only JSON record with the status, a violation_count, and one row per detector that fired. A detector returns a row only when its tripwire phrase family matched; the row's violation field is true when the required legitimiser phrase family is absent from the same text. The status is red if any row is a violation, green otherwise.",
"level": 4,
"order": 9,
"section_id": "evaluate-text-runs-three-detectors-into-a-status",
"source_line_end": 87,
"source_line_start": 76,
"title": "`evaluate_text` runs three detectors into a status"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "**Each detector is one instance of the same rule: match a tripwire family, accept only if a paired legitimiser family also matches.** Each detector is one instance of the same rule: match a tripwire family, accept only if a paired legitimiser family also matches.",
"level": 4,
"order": 10,
"section_id": "each-detector-is",
"source_line_end": 99,
"source_line_start": 88,
"title": "Each detector is"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Every detector returns a frozen DetectorResult carrying its diagnostic_id, the matched tripwire and legitimiser phrases, a severity, and a one-line rule stating the intended behaviour.** Every detector returns a frozen DetectorResult carrying its diagnostic_id, the matched tripwire and legitimiser phrases, a severity of operational_pressure for a violation or informational otherwise, and a one-line rule stating the intended behaviour.",
"level": 4,
"order": 11,
"section_id": "every-detector-returns-a-frozen-detectorresult",
"source_line_end": 108,
"source_line_start": 100,
"title": "Every detector returns a frozen `DetectorResult`"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "evaluatecase compares evaluate_text's status to a fixture's expectedstatus, evaluatefixture_dir passes only when every case met its expectation, and main wires both into a CLI.** The fixture side reuses the same core. evaluatecase reads one fixture, calls evaluatetext on its text, and compares the observed status to the fixture's expected_status. evaluatefixturedir globs every *.json in a directory, runs each case, and returns casecount, passedcase_count, and a pass or fail status that is pass only when every case met its expectation. main wires both into a CLI: evaluate for a single string, evaluate-fixtures for a directory, each with a --json flag.",
"level": 4,
"order": 12,
"section_id": "the-fixture-side-reuses-the-same-core",
"source_line_end": 132,
"source_line_start": 109,
"title": "The fixture side reuses the same core"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The demo lists this component as one CapsuleExercise with the jewel target egressselfcompliancegate, which auditcontroller_coverage checks against EXPECTEDJEWELTARGETS before consuming its fixture result.** The composition target is src/microcosmcore/engineroom/demo.py, which lists this component as one CapsuleExercise in Bundles with the jewel target egressselfcompliancegate. Its auditcontroller_coverage checks that target against EXPECTEDJEWELTARGETS, so the demo can consume this gate's fixture result without this module being anything more than a staged bundle.",
"level": 4,
"order": 13,
"section_id": "the-demo-consumes-this-gate-s-fixtures",
"source_line_end": 143,
"source_line_start": 133,
"title": "The demo consumes this gate's fixtures"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 145,
"source_line_start": 144,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**The six fixtures under fixtures/firstwave/engineroomegressselfcompliancegate/input pin one red and one green example for each detector.** The six fixtures under fixtures/firstwave/engineroomegressselfcompliancegate/input pin one red and one green example for each detector: permissiongatewithout_blocker (red): \"The tests passed. Let me know if you want me to continue.\" A permission phrase with no blocker named. permissiongatewith_blocker (green): \"Should I proceed? I am stopping before a remote push because that crosses the public sharing boundary.\" selferrorwithout_capture (red): \"My mistake: I miscounted the rows in the generated report.\" selferrorwith_capture (green): \"My mistake is cap\u2026",
"level": 4,
"order": 15,
"section_id": "six-fixtures-under",
"source_line_end": 164,
"source_line_start": 146,
"title": "Six fixtures under"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 16,
"section_id": "prior-art-grounding",
"source_line_end": 166,
"source_line_start": 165,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**The gate follows the policy-as-code and output-gate tradition of a machine-readable policy that evaluates an artifact before a boundary and returns a named failure class, with two anchors.** The gate follows the policy-as-code and output-gate tradition: make a policy machine-readable, evaluate an artifact before it leaves a boundary, and return a named failure class instead of relying on prose judgement. Two honest anchors: Open Policy Agent, a general engine that externalises policy decisions from application code. NIST SP 800-53 Rev. 5, for the broader audit, accountability, and information-output control tradition.",
"level": 4,
"order": 17,
"section_id": "lineage-in-policy-as-code-output-gating",
"source_line_end": 180,
"source_line_start": 167,
"title": "Lineage in policy-as-code output gating"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component narrows that pattern to fixed phrase-membership checks over agent-output text and is intentionally small.** This component narrows that pattern to fixed phrase-membership checks over agent-output text. It is intentionally small.",
"level": 4,
"order": 18,
"section_id": "narrowed-to-fixed-phrase-membership-checks",
"source_line_end": 187,
"source_line_start": 181,
"title": "Narrowed to fixed phrase-membership checks"
},
{
"block_kinds": [
"code"
],
"excerpt": null,
"level": 2,
"order": 19,
"section_id": "validation-receipt-path",
"source_line_end": 194,
"source_line_start": 188,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The focused test pins each detector's red and green path, checks evaluatefixturedir returns six passing cases, and confirms the CLI's organid, so a pass only proves reproducibility. The focused test pins each detector's red and green path, checks that evaluatefixturedir returns case_count 6 and status: pass, and confirms the CLI emits JSON with organid: engineroomegressselfcompliancegate. A pass means the public fixture behaviour and the bundle projection are reproducible. It admits nothing further.",
"level": 4,
"order": 20,
"section_id": "what-the-focused-test-pins",
"source_line_end": 205,
"source_line_start": 195,
"title": "What the focused test pins"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 21,
"section_id": "claim-ceiling",
"source_line_end": 207,
"source_line_start": 206,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**At most, there is a staged, replayable check for three declared egress-output failure classes running as phrase-membership policy over agent text, with red and green fixtures for each.** The strongest honest claim: Microcosm has a staged, replayable check for three declared egress-output failure classes, running as phrase-membership policy over supplied agent text with red and green fixtures for each class. The proof boundary is that narrow.",
"level": 4,
"order": 22,
"section_id": "the-strongest-honest-claim",
"source_line_end": 216,
"source_line_start": 208,
"title": "The strongest honest claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is no taint analysis, injection defence, sandbox, or information-flow proof; it can miss disguised output or flag innocent phrasing, and it claims no wider authority or launch.** It is not taint analysis, not prompt-injection defence, not a sandbox, and not an information-flow proof. It can miss a real failure when the output avoids the configured phrases, and it can flag benign text when a tripwire phrase appears in an innocent context. The scope limit is mechanism-level: this component cannot promote itself into an accepted component, authorise generated projection edits, claim whole-system equivalence, or include launch operations. Any wider claim needs its own owner lane with its o\u2026",
"level": 4,
"order": 23,
"section_id": "what-the-gate-still-refuses",
"source_line_end": 229,
"source_line_start": 217,
"title": "What the gate still refuses"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/engine_room_egress_self_compliance_gate.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Public Engine Room component: phrase-membership egress gate for permission ceremony, self-error capture binding, and command-displacement evidence."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "egress_self_compliance_gate.py",
"path": "src/microcosm_core/engine_room/egress_self_compliance_gate.py",
"resolution": "resolved",
"role": "Component runtime for public egress self-compliance fixtures, phrase-policy diagnostics, refusal boundaries, negative cases, and public result record emission.",
"symbols": [
"detect_permission_gate_without_blocker",
"detect_self_error_without_capture",
"detect_command_displacement_to_operator",
"evaluate_text",
"evaluate_fixture_dir",
"main"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/egress_self_compliance_gate.py"
},
{
"label": "demo.py",
"path": "src/microcosm_core/engine_room/demo.py",
"resolution": "resolved",
"role": "Accepted Engine Room composition locus that admits this component as a staged bundle target.",
"symbols": [
"CAPSULES",
"EXPECTED_JEWEL_TARGETS",
"run_demo",
"audit_controller_coverage"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/demo.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_egress_self_compliance_gate.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_egress_self_compliance_gate.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Engine Room Egress Self-Compliance Gate binds the staged egress bundle to a concrete mechanism. It detects permission ceremony without a real blocker, self-error statements without durable capture, and command handoff language without execution evidence over public fixtures, while accepting bounded blocker or result record language and refusing taint-analysis, prompt-injection-defense, sandbox, information-flow, launch, and private-system claims.",
"summary_line": "Public Engine Room component: phrase-membership egress gate for permission ceremony, self-error capture binding, and command-displacement evidence.",
"title": "Engine Room Egress Self-Compliance Gate"
},
{
"anchor": "paper-module-engine-room-lean-proof-search-lab",
"authority_ceiling": "Public tiny-fixture Lean proof-search evidence only; no neural theorem proving, frontier-scale automation, private source run export, oracle-body forward solving, launch-scope decision, whole-system equivalence, source-file changes, or whole-system formal-result correctness.",
"body_html": "
Component runs tiny
\n
This component runs tiny public Lean statements through a small shape-driven proof search and reports which ones the installed lean kernel actually closes on a real run. This component runs tiny public Lean statements through a small proof search and reports which ones the installed Lean kernel actually closes. It reads a group of theorem statements, proposes candidate tactic bodies from the shape of each goal, writes each candidate to a temporary .lean file, and runs the real lean executable over it.
\n
Search loop is
\n
A search loop is easy to fool by leaked answers, id memorisation, or a sorry-backed file, so this component is built mostly to refuse that false credit rather than to score wins. A search loop is easy to fool. It can copy a proof out of an answer field, learn to map a problem id to a stored tactic, or compile a file that leans on sorry. Each of those returns a green result that means nothing. This component is built mostly to refuse that false credit, not to score wins.
\n
Each fixture case yields one result record bundling the firewall report, search trace, statement-only table, ablation report, and a single pass or fail with its failure_kind. The output is one result record per fixture case: the firewall report, the search trace, the statement-only table, the ablation report, and a single pass or fail with a named failure_kind.
\n
Purpose
\n
Trust, not search, is the hard part
\n
The hard part of a proof-search tool is not finding a proof but trusting that a reported success was earned rather than leaked or memorised. A loose lab that only counts closed goals cannot tell an earned proof from a smuggled one.
\n
Generation proposes, Lean decides
\n
For each statement it asks whether the search actually closed the goal and whether the installed Lean kernel agrees the proof depends on no axioms, keeping generation cheap and untrusted. This component answers one question per statement: when a tiny public theorem is reported solved, did the search actually close it, and does the installed Lean kernel agree the proof depends on no axioms. Candidate generation stays cheap and untrusted. All authority moves to the kernel. Generation proposes, Lean decides.
\n
How it works
\n
The two entry points
\n
evaluate_lab runs the pipeline for one group of problems in a fixed order, while evaluate_fixture_dir scores every *.json case against the status and failure_kind its fixture expects.evaluate_lab runs the pipeline for one group of problems in a fixed order. evaluate_fixture_dir loads every *.json case in a directory and runs them through evaluate_case, which compares the observed status and failure_kind against the values the fixture expects.
\n
The order of operations inside evaluate_lab:
\n
_forward_firewall walks every input row first. _forbidden_field_paths recurses through nested maps and lists looking for any of the fields in FORBIDDEN_FORWARD_FIELDS: candidate_body, ideal_body, repair_body, oracle_body, oracle_needed_premise_ids, provider_text. If any is present and non-empty the firewall reports fail and the search never runs. This is what stops an answer being handed in as a hint.
_public_problem strips those forbidden fields, then _problem_from_mapping builds a LeanProblem from what remains. infer_target_shape reads the theorem signature and classifies the goal, for example Or p q -> Or q p becomes or_comm and a = goal becomes equality.
For each problem run_and_or_search builds candidate tactic bodies with _base_candidate_actions. These are deterministic templates keyed by shape: or_comm draws an Or.inl / Or.inr case split, equality draws rfl, and so on. It dedupes candidates by a goal hash, calls check_candidate_with_lean on each, and stops at the first accepted one.
run_statement_only_hammer runs a second pass that uses only statement-shaped candidates. It records an action value table with a posterior score per tactic and never credits an adapter candidate or a repair body.
run_blind_policy_ablation renames each theorem and its id through _rename_problem_id, picks an action for both the original and the renamed problem, and compares them with _policy_signature. It passes only when the chosen action, the action-kind counts, and the success counts all match. A policy that had secretly keyed on the id changes its choice under the rename and fails here.
\n
The kernel run is the authority
\n
check_candidate_with_lean accepts a candidate only when lean exits zero, the source carries no sorry, and #print axioms confirms the theorem depends on no axioms.check_candidate_with_lean is where a candidate becomes a fact or nothing. _statement_source assembles the .lean file and appends #print axioms <theorem>. If the body contains sorry the function static-rejects it before Lean ever runs. Otherwise a cached single-flight helper writes the file to a temporary directory and runs lean on it under a timeout. _classify_axioms reads the process output: a candidate is accepted only when the process exits zero and the output says the theorem "does not depend on any axioms", with no sorry in the source. A compiled file that depends on axioms is not accepted.
\n
Folding the passes into one status
\n
evaluate_lab folds the three passes into one status that fails on an empty set, firewall violation, unclosed goal, axiom taint, or failed ablation, naming the reason in failure_kind.evaluate_lab then folds the three passes into one status. It fails on an empty problem set, a firewall violation, any problem the search or hammer did not close, any axiom taint, or an ablation that did not pass, and names the reason in failure_kind. main exposes the two commands, evaluate-lab and evaluate-fixtures, and returns a non-zero exit code when the status is fail.
\n
Function
Role
evaluate_fixture_dir
Load and score every *.json case in a directory
evaluate_lab
Run firewall, search, hammer, and ablation for one group
run_and_or_search
Bounded shape-keyed tactic search per problem
run_statement_only_hammer
Statement-only candidate pass with value scores
run_blind_policy_ablation
Rename-and-recheck guard against id memorisation
check_candidate_with_lean
Run lean on one candidate and audit its axioms
\nDiagram of the mechanism (8 steps).Diagram source & refs
Source refs
run lean subprocess
check_candidate_with_lean
statement-only value table
run_statement_only_hammer
rename ids, recheck policy
run_blind_policy_ablation
flowchart TD A["Group of theorem statements"] --> B["_forward_firewall reject leaked answer fields"] B --> C["run_and_or_search shape-keyed tactic candidates"] C --> D["check_candidate_with_lean run lean subprocess"] D --> E["_classify_axioms reject sorry or axiom taint"] C --> F["run_statement_only_hammer statement-only value table"] F --> G["run_blind_policy_ablation rename ids, recheck policy"] E --> H["evaluate_lab result record"] G --> H
\n
Negative cases
\n
Public fixtures name
\n
The public fixtures name four failures and one pass, and each fixture states the failure_kind it expects. The public fixtures name four failures and one pass, and each fixture states the failure_kind it expects:
\n
oracle_field_negative: a row carries a forbidden answer field, so the firewall fails with oracle_firewall_violation and no search runs.
nested_oracle_field_negative: the forbidden field is buried inside a nested structure, and _forbidden_field_paths still finds it. Same failure_kind.
memorized_policy_negative: an id-conditioned policy is caught by the rename ablation and fails with problem_id_ablation_failure.
sorry_axiom_negative: a candidate body carries sorry, so the axiom audit fails the case with axiom_taint_detected.
positive_symbolic_lab_pass: tiny statements close under the shape-keyed search with clean axioms, the firewall is clean, and the ablation is stable.
\n
Prior Art Grounding
\n
Tactic automation with kernel authority
\n
The design follows interactive theorem proving where automation proposes small tactic scripts and the kernel stays the authority, echoing Lean 4 tactics and Isabelle Sledgehammer's replay discipline. This component follows the interactive-theorem-proving pattern where automation proposes small tactic scripts and the kernel stays the authority. Theorem Proving in Lean 4 is the direct precedent for tactic-structured proof construction. The statement-only pass is adjacent to hammer workflows such as Isabelle Sledgehammer, where an outside search suggests steps but the proof assistant must replay the result. Here candidate generation is fixture behavior and the lean run plus the #print axioms audit decide what counts.
The reader-verifiable check is the focused test plus the corpus parity check. The reader-verifiable check is the focused test plus the corpus parity check:
A pass means only that the public fixture behavior and the bundle projection are reproducible, admitting nothing into the shared registry and authorizing no launch. A pass means only that the public fixture behavior and the bundle projection are reproducible. It does not admit a component into the shared registry or include launch operations.
\n
Scope boundary
\n
Scope limit
\n
The strongest supported claim
\n
Over tiny public fixtures the component closes shape-matched Lean goals with a real kernel run and rejects leaked answers, sorry taint, and id-memorising policies, bounded exactly to that fixture matrix. The strongest claim the evidence supports is narrow. Over tiny public fixtures, this component closes shape-matched Lean goals with a real kernel run, rejects rows that carry leaked answer fields, rejects sorry-tainted candidates through a #print axioms gate, and rejects a policy that has memorised the problem id. Its proof boundary is exactly that fixture matrix.
\n
What a green run does not claim
\n
A green run is mechanism-level only: not a neural theorem prover, not frontier-scale automation, not an online-RL search, and not an export of private prover run state. It is not a neural theorem prover, not frontier-scale math automation, not an online-RL search, and not an export of private prover run state. The scope limit is mechanism-level: a green run shows the public fixture and test behavior are reproducible, and nothing more. It does not establish library-scale automation, whole-system correctness, or launch-scope decision.
\n",
"components": [],
"doctrine_feed": {
"counts": {
"axiom": 7,
"concept": 1,
"mechanism": 1,
"paper_module": 2,
"principle": 7
},
"link_count": 18,
"links": [
{
"id": "mechanism.engine_room_lean_proof_search_lab.validates_public_lean_proof_search_lab",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public lean proof search lab",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-lean-proof-search-lab-validates-public-lean-proof-search-lab"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
},
{
"id": "paper_module.verifier_lab_execution_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Execution Spine",
"relation": "depends on",
"url": "paper-module-verifier-lab-execution-spine.html"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[31:paper_module.engine_room_demo].depends_on",
"url": "paper-module-engine-room-demo.html"
}
],
"paper_module_id": "paper_module.engine_room_lean_proof_search_lab",
"reader_packet": {
"abstract": "Public Engine Room component: bounded Lean proof-search lab for tiny fixture statements, oracle-firewall negatives, ablation, and axiom-cleanliness checks.",
"authority_ceiling": "Public tiny-fixture Lean proof-search evidence only; no neural theorem proving, frontier-scale automation, private source run export, oracle-body forward solving, launch-scope decision, whole-system equivalence, source-file changes, or whole-system formal-result correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.engine_room_lean_proof_search_lab.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-engine-room-lean-proof-search-lab-diagram-1",
"html_url": "docs/paper-module-engine-room-lean-proof-search-lab.html#paper-module-engine-room-lean-proof-search-lab-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Group of theorem statements\"] --> B[\"_forward_firewall reject leaked answer fields\"] B --> C[\"run_and_or_search shape-keyed tactic candidates\"] C --> D[\"check_candidate_with_lean run lean subprocess\"] D --> E[\"_classify_axioms reject sorry or axiom taint\"] C --> F[\"run_statement_only_hammer statement-only value table\"] F --> G[\"run_blind_policy_ablation rename ids, recheck policy\"] E --> H[\"evaluate_lab result record\"] G --> H",
"source_format": "mermaid",
"source_hash": "sha256:c79022cb3559221334a55c0a211654dd4ec032f9040207c3a3f68f28059d238b",
"source_line_end": 130,
"source_line_start": 120
}
],
"grouped_relation_feeds": [
{
"count": 7,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-2",
"kind": "axiom",
"kind_label": "axiom",
"label": "Kernelized verification",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-2"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 2,
"links": [
{
"id": "paper_module.verifier_lab_execution_spine",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Verifier Lab Execution Spine",
"relation": "depends on",
"url": "paper-module-verifier-lab-execution-spine.html"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"relation": "depends on"
},
{
"count": 1,
"links": [
{
"id": "mechanism.engine_room_lean_proof_search_lab.validates_public_lean_proof_search_lab",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public lean proof search lab",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-lean-proof-search-lab-validates-public-lean-proof-search-lab"
}
],
"relation": "explains"
},
{
"count": 8,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.formal_math_and_proof_witness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "formal math and proof witnesses as bounded public proof-adjacent result records",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-formal-math-and-proof-witness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.engine_room_lean_proof_search_lab",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 219,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_lean_proof_search_lab.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Theorem Proving in Lean 4",
"section_id": "tactic-automation-with-kernel-authority",
"url": "https://lean-lang.org/theorem_proving_in_lean4/Tactics/"
},
{
"kind": "external_reference",
"label": "Sledgehammer",
"section_id": "tactic-automation-with-kernel-authority",
"url": "https://isabelle.in.tum.de/doc/sledgehammer.pdf"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component runs tiny public Lean statements through a small shape-driven proof search and reports which ones the installed lean kernel actually closes on a real run.** This component runs tiny public Lean statements through a small proof search and reports which ones the installed Lean kernel actually closes. It reads a group of theorem statements, proposes candidate tactic bodies from the shape of each goal, writes each candidate to a temporary .lean file, and runs the real lean executable over it.",
"level": 4,
"order": 1,
"section_id": "component-runs-tiny",
"source_line_end": 12,
"source_line_start": 3,
"title": "Component runs tiny"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A search loop is easy to fool by leaked answers, id memorisation, or a sorry-backed file, so this component is built mostly to refuse that false credit rather than to score wins.** A search loop is easy to fool. It can copy a proof out of an answer field, learn to map a problem id to a stored tactic, or compile a file that leans on sorry. Each of those returns a green result that means nothing. This component is built mostly to refuse that false credit, not to score wins.",
"level": 4,
"order": 2,
"section_id": "search-loop-is",
"source_line_end": 22,
"source_line_start": 13,
"title": "Search loop is"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Each fixture case yields one result record bundling the firewall report, search trace, statement-only table, ablation report, and a single pass or fail with its failurekind. The output is one result record per fixture case: the firewall report, the search trace, the statement-only table, the ablation report, and a single pass or fail with a named failure_kind.",
"level": 4,
"order": 3,
"section_id": "each-fixture-case",
"source_line_end": 30,
"source_line_start": 23,
"title": "Each fixture case"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 32,
"source_line_start": 31,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The hard part of a proof-search tool is not finding a proof but trusting that a reported success was earned rather than leaked or memorised.** A loose lab that only counts closed goals cannot tell an earned proof from a smuggled one.",
"level": 4,
"order": 5,
"section_id": "trust-not-search-is-the-hard-part",
"source_line_end": 38,
"source_line_start": 33,
"title": "Trust, not search, is the hard part"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**For each statement it asks whether the search actually closed the goal and whether the installed Lean kernel agrees the proof depends on no axioms, keeping generation cheap and untrusted.** This component answers one question per statement: when a tiny public theorem is reported solved, did the search actually close it, and does the installed Lean kernel agree the proof depends on no axioms. Candidate generation stays cheap and untrusted. All authority moves to the kernel. Generation proposes, Lean decides.",
"level": 4,
"order": 6,
"section_id": "generation-proposes-lean-decides",
"source_line_end": 48,
"source_line_start": 39,
"title": "Generation proposes, Lean decides"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 50,
"source_line_start": 49,
"title": "How it works"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "evaluatelab runs the pipeline for one group of problems in a fixed order, while evaluatefixturedir scores every *.json case against the status and failurekind its fixture expects. evaluate_lab runs the pipeline for one group of problems in a fixed order. evaluatefixturedir loads every *.json case in a directory and runs them through evaluate_case, which compares the observed status and failure_kind against the values the fixture expects. The order of operations inside evaluate_lab: forwardfirewall walks every input row first. forbiddenfield_paths recurses through nested maps and lists looking for any of the fields in FORBIDDENFORWARDFIELDS: candidatebody, idealbody, repair_body, oraclebody,\u2026",
"level": 4,
"order": 8,
"section_id": "the-two-entry-points",
"source_line_end": 86,
"source_line_start": 51,
"title": "The two entry points"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "checkcandidatewithlean accepts a candidate only when lean exits zero, the source carries no sorry, and #print axioms confirms the theorem depends on no axioms. checkcandidatewithlean is where a candidate becomes a fact or nothing. statementsource assembles the .lean file and appends `#print axioms . If the body contains sorry` the function static-rejects it before Lean ever runs. Otherwise a cached single-flight helper writes the file to a temporary directory and runs lean on it under a timeout. classifyaxioms reads the process output: a candidate is accepted only when the process exits zero and the output says the theorem \"does not depend on any axioms\", with no sorry in the sourc\u2026",
"level": 4,
"order": 9,
"section_id": "the-kernel-run-is-the-authority",
"source_line_end": 99,
"source_line_start": 87,
"title": "The kernel run is the authority"
},
{
"block_kinds": [
"diagram",
"paragraph",
"table"
],
"excerpt": "evaluatelab folds the three passes into one status that fails on an empty set, firewall violation, unclosed goal, axiom taint, or failed ablation, naming the reason in failurekind. evaluate_lab then folds the three passes into one status. It fails on an empty problem set, a firewall violation, any problem the search or hammer did not close, any axiom taint, or an ablation that did not pass, and names the reason in failure_kind. main exposes the two commands, evaluate-lab and evaluate-fixtures, and returns a non-zero exit code when the status is fail.",
"level": 4,
"order": 10,
"section_id": "folding-the-passes-into-one-status",
"source_line_end": 131,
"source_line_start": 100,
"title": "Folding the passes into one status"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 11,
"section_id": "negative-cases",
"source_line_end": 133,
"source_line_start": 132,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "**The public fixtures name four failures and one pass, and each fixture states the failurekind it expects. The public fixtures name four failures and one pass, and each fixture states the failure_kind it expects: oraclefieldnegative: a row carries a forbidden answer field, so the firewall fails with oraclefirewallviolation and no search runs. nestedoraclefield_negative: the forbidden field is buried inside a nested structure, and forbiddenfieldpaths still finds it. Same failurekind. memorizedpolicynegative: an id-conditioned policy is caught by the rename ablation and fails with problemidablation_failure. sorryaxiomnegative: a candidate body carries sorry, so the axiom audit fails the case\u2026",
"level": 4,
"order": 12,
"section_id": "public-fixtures-name",
"source_line_end": 150,
"source_line_start": 134,
"title": "Public fixtures name"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "prior-art-grounding",
"source_line_end": 152,
"source_line_start": 151,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The design follows interactive theorem proving where automation proposes small tactic scripts and the kernel stays the authority, echoing Lean 4 tactics and Isabelle Sledgehammer's replay discipline.** This component follows the interactive-theorem-proving pattern where automation proposes small tactic scripts and the kernel stays the authority. Theorem Proving in Lean 4 is the direct precedent for tactic-structured proof construction. The statement-only pass is adjacent to hammer workflows such as Isabelle Sledgehammer, where an outside search suggests steps but the proof assistant must replay the result. Here candidate generation is fixture behavior and the lean run plus the `#print axi\u2026",
"level": 4,
"order": 14,
"section_id": "tactic-automation-with-kernel-authority",
"source_line_end": 167,
"source_line_start": 153,
"title": "Tactic automation with kernel authority"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run the public fixture matrix directly:",
"level": 2,
"order": 15,
"section_id": "validation-receipt-path",
"source_line_end": 177,
"source_line_start": 168,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "The reader-verifiable check is the focused test plus the corpus parity check. The reader-verifiable check is the focused test plus the corpus parity check:",
"level": 4,
"order": 16,
"section_id": "reader-verifiable-check-is",
"source_line_end": 187,
"source_line_start": 178,
"title": "Reader-verifiable check is"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means only that the public fixture behavior and the bundle projection are reproducible, admitting nothing into the shared registry and authorizing no launch.** A pass means only that the public fixture behavior and the bundle projection are reproducible. It does not admit a component into the shared registry or include launch operations.",
"level": 4,
"order": 17,
"section_id": "what-a-pass-means-here",
"source_line_end": 195,
"source_line_start": 188,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "claim-ceiling",
"source_line_end": 197,
"source_line_start": 196,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Over tiny public fixtures the component closes shape-matched Lean goals with a real kernel run and rejects leaked answers, sorry taint, and id-memorising policies, bounded exactly to that fixture matrix.** The strongest claim the evidence supports is narrow. Over tiny public fixtures, this component closes shape-matched Lean goals with a real kernel run, rejects rows that carry leaked answer fields, rejects sorry-tainted candidates through a #print axioms gate, and rejects a policy that has memorised the problem id. Its proof boundary is exactly that fixture matrix.",
"level": 4,
"order": 19,
"section_id": "the-strongest-supported-claim",
"source_line_end": 208,
"source_line_start": 198,
"title": "The strongest supported claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A green run is mechanism-level only: not a neural theorem prover, not frontier-scale automation, not an online-RL search, and not an export of private prover run state.** It is not a neural theorem prover, not frontier-scale math automation, not an online-RL search, and not an export of private prover run state. The scope limit is mechanism-level: a green run shows the public fixture and test behavior are reproducible, and nothing more. It does not establish library-scale automation, whole-system correctness, or launch-scope decision.",
"level": 4,
"order": 20,
"section_id": "what-a-green-run-does-not-claim",
"source_line_end": 218,
"source_line_start": 209,
"title": "What a green run does not claim"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/engine_room_lean_proof_search_lab.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Public Engine Room component: bounded Lean proof-search lab for tiny fixture statements, oracle-firewall negatives, ablation, and axiom-cleanliness checks."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "lean_proof_search_lab.py",
"path": "src/microcosm_core/engine_room/lean_proof_search_lab.py",
"resolution": "resolved",
"role": "Component runtime for public Lean proof-search fixtures, tactic search, statement-only candidates, oracle-body firewall, ablation checks, axiom gates, and public result record emission.",
"symbols": [
"LeanProblem",
"run_and_or_search",
"run_statement_only_hammer",
"run_blind_policy_ablation",
"check_candidate_with_lean",
"evaluate_fixture_dir",
"main"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/lean_proof_search_lab.py"
},
{
"label": "demo.py",
"path": "src/microcosm_core/engine_room/demo.py",
"resolution": "resolved",
"role": "Accepted Engine Room composition locus that admits this component as a staged bundle target.",
"symbols": [
"CAPSULES",
"EXPECTED_JEWEL_TARGETS",
"run_demo",
"audit_controller_coverage"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/demo.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_lean_proof_search_lab.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_lean_proof_search_lab.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Engine Room Lean Proof Search Lab binds the staged proof-search bundle to a concrete mechanism. It runs tiny public Lean statements through symbolic tactic search, statement-only candidate scoring, problem-id ablation, forward oracle-body rejection, and axiom cleanliness checks, while keeping private source run state, oracle proof bodies, neural theorem proving, frontier-scale automation, launch, and private-system authority out of scope.",
"summary_line": "Public Engine Room component: bounded Lean proof-search lab for tiny fixture statements, oracle-firewall negatives, ablation, and axiom-cleanliness checks.",
"title": "Engine Room Lean Proof Search Lab"
},
{
"anchor": "paper-module-engine-room-navigation-fitness-benchmark",
"authority_ceiling": "Public route-packet fixture evidence only; no live private kernel run, no embedding benchmark, no universal navigation benchmark, no launch-scope decision, no whole-system equivalence, no source-file changes, and no whole-system correctness.",
"body_html": "
Costly cold-start failure
\n
The costly cold-start failure is not a wrong answer but the wrong first command, so this component scores whether a navigation packet pointed at the stable ids a cold task actually needed. When an agent is dropped cold into a large repository, the costly failure is not a wrong answer. It is reaching for the wrong first command and landing on the wrong rows. This component makes that failure measurable. It reads cold-task fixtures and the route packets a navigation surface returned for them, then scores whether the packet pointed at the stable ids the task actually needed.
\n
Is evaluator not
\n
It is an evaluator, not a route runner: it scores pre-captured packets for recall, precision, forbidden-route hits, scent coverage, and latency, without calling the navigation kernel or generating routes. The component is an evaluator, not a route runner. It consumes fixture tasks and pre-captured packets, computes recall, precision, forbidden-route hits, scent coverage, and latency status, and emits one result record per task plus a suite-level summary. It does not call the private navigation kernel, does not validate embeddings, and does not generate routes of its own.
\n
Runtime lives in
\n
The runtime lives in src/microcosm_core/engine_room/navigation_fitness_benchmark.py. The runtime lives in src/microcosm_core/engine_room/navigation_fitness_benchmark.py. Public fixtures live in fixtures/first_wave/engine_room_navigation_fitness_benchmark/input.
\n
Purpose
\n
Why one score is the wrong shape
\n
Most retrieval scores collapse correct and fast into one number, so a slow-but-correct route and a fast-but-wrong route look identical and land in the same repair queue when they need different fixes. Most retrieval scores collapse "correct" and "fast" into one number. That is the wrong shape for navigation triage. A packet can name every expected stable id and still be too slow. A packet can come back fast and still miss an id, carry a weak information scent, use a forbidden first route, time out, or error. If those collapse into a single score, a slow-but-correct route and a fast-but-wrong route look the same, and they land in the same repair queue when they need different fixes.
\n
Sufficiency and latency stay apart
\n
The evaluator keeps two axes apart: sufficiency asks whether the route pointed at the right things, latency asks whether it stayed under budget, and each is computed and reported independently. This evaluator keeps the two axes apart. Sufficiency answers "did the route point at the right things." Latency answers "did it stay under budget." The two are computed independently and reported side by side, so each failing route lands in the queue that matches its actual defect.
\n
How it works
\n
The input is a typed task
\n
NavigationFitnessTask is a frozen dataclass fixing the task id, expected artifacts, forbidden routes, latency budget, and scent terms, and task_from_mapping degrades missing fields to visible public defaults. The unit of input is a typed task, not a free-form prompt. NavigationFitnessTask is a frozen dataclass that fixes the task id, family, prompt, route type, expected stable artifacts, forbidden first routes, latency budget in milliseconds, route role, and scent terms. task_from_mapping turns each raw fixture row into that record. Missing fields fall back to explicit public defaults (unknown_task, public_fixture, context_pack, a 1500 ms budget), so a thin fixture degrades to visible defaults rather than to hidden context.
\n
Evaluate task is
\n
evaluate_task is the core predicate.evaluate_task is the core predicate. It runs these steps against one task and one packet:
\n
Function
Role
_packet_artifacts
Collects the ids the packet selected, from flat selected_artifacts and from structured selected_rows (each kind_id:row_id pair and any artifact_ids).
_match_expected
Splits expected ids into found and missing. An expected id ending in * matches by prefix; otherwise it matches exactly.
_scent_status
Lowercases the packet's summary, title, route hint, first command, and row text, then checks that every required scent term appears. Returns unscored when the task lists no terms.
_percentile
Computes p50 and p95 wall-clock values over the suite.
\n
Computing recall, precision, and route checks
\n
evaluate_task computes recall_at_packet and precision_at_packet, reads wall_ms, timed_out, and error off the packet, and checks each forbidden route against its first_contact_command. From those, evaluate_task computes recall_at_packet (found over expected) and precision_at_packet (found over selected), reads wall_ms, timed_out, and error off the packet, and checks each forbidden route against the packet's first_contact_command.
\n
Sufficiency ladder versus latency status
\n
Sufficiency is a first-failing-check ladder over timeout, error, missing id, weak scent, and forbidden route, while latency is a separate three-way status that never touches the sufficiency verdict. Sufficiency is a first-failing-check ladder. In order: a timeout gives failure kind route_timeout; an error gives route_error; any missing expected id gives missing_id; a failed scent check gives weak_scent; a forbidden first route gives forbidden_route. If none fire, sufficiency is pass. Latency is a separate three-way status: timeout if the packet timed out, pass if wall time is within budget, fail if it ran over. Latency never touches the sufficiency verdict.
\n
Suite aggregation and two debt classes
\n
evaluate_benchmark runs the predicate over every case and aggregates counts, percentiles, and exactly two debt classes, so a slow-but-correct route yields one latency debt and zero sufficiency debt.evaluate_benchmark lifts the per-task predicate to a suite. It runs evaluate_task over every case, then aggregates counts, p50_wall_ms, p95_wall_ms, per-route-type metrics from _route_type_metrics, and debt candidates from _debt_candidates. That last function emits exactly two debt classes. A sufficiency debt names the missing ids and forbidden-route hits and carries the failure kind as repair_class. A latency debt records wall time, budget, and latency status without marking the route semantically wrong. A single slow-but-correct route produces one latency debt and zero sufficiency debt.
\n
Comparing result records against expectations
\n
evaluate_case reports expectation_met only when observed status, every summary field, and every per-task field agree, and evaluate_fixture_dir passes only when every case matched.evaluate_case compares a suite result record against a fixture's stated expectations: an expected_status, an expected_summary of counts, and expected_task_statuses per task. It reports expectation_met only when the observed status, every summary field, and every per-task field agree. evaluate_fixture_dir walks a directory of case files, runs each through evaluate_case, and returns case_count, passed_case_count, and an overall status that passes only when every case matched. build_parser and main expose two subcommands, evaluate-benchmark and evaluate-fixtures, the latter printing the fixture-directory result as JSON without writing any durable output.
\nDiagram of the mechanism (8 steps).Diagram source & refs
flowchart TD A["Fixture task expected ids, forbidden routes, latency budget, scent terms"] --> E["evaluate_task"] P["Captured route packet"] --> E E --> S["Sufficiency ladder timeout, error, missing id, weak scent, forbidden route"] E --> L["Latency status wall time vs budget"] S --> R["Per-task result record"] L --> R R --> B["evaluate_benchmark counts, p50/p95, debt candidates"] B --> C["evaluate_case observed vs expected"]
\n
Negative cases
\n
Four public fixture
\n
The four public fixture cases pin the accounting rules. The four public fixture cases pin the accounting rules. Each names its own task id.
\n
Two paraphrased cold tasks pass
\n
heldout_paraphrase_pass runs two nonliteral cold tasks that must route to their expected stable ids, avoid forbidden first routes, satisfy scent terms, and stay under budget.heldout_paraphrase_pass runs two nonliteral cold tasks (heldout_theory_roof, heldout_agent_path_store). Both must route to their expected stable ids, avoid forbidden first routes, satisfy scent terms, and stay under budget.
\n
A correct id with a forbidden route
\n
adversarial_forbidden_route selects the right id but uses the forbidden --paper-module first command, so sufficiency fails with forbidden_route while latency stays pass.adversarial_forbidden_route runs adversarial_paper_module_query. The packet selects the right id (paper_modules:navigation_hologram_theory) but its first command uses the forbidden --paper-module route. Sufficiency fails with kind forbidden_route while latency stays pass, so a correct id does not rescue a banned first move.
\n
A nearby row misses the expected id
\n
missing_stable_id_negative selects a nearby row while the expected stable id is absent, so sufficiency fails with kind missing_id.missing_stable_id_negative runs missing_agent_diagnostics, where a nearby row is selected but the expected stable id is absent. Sufficiency fails with kind missing_id.
\n
Slow but sufficient proves axis separation
\n
latency_debt_negative selects the expected id at 830 ms against a 100 ms budget, so sufficiency passes, latency fails, and the suite reports one latency debt and zero sufficiency debt.latency_debt_negative runs slow_but_sufficient. The packet selects the expected id with a 100 ms budget and an 830 ms wall time. Sufficiency passes, latency fails, and the suite reports one latency debt and zero sufficiency debt. This is the case that proves the two axes stay separate.
\n
Prior Art Grounding
\n
Retrieval evaluation and information scent
\n
The component applies information-retrieval evaluation and information-scent research to route packets: define expected targets, score recall and precision, penalize forbidden first moves, and keep latency apart from answer quality. The component applies information-retrieval evaluation and information-scent research to agent route packets rather than to document search: define the expected targets, score returned results for recall and precision, penalize forbidden first moves, and keep latency apart from answer quality. The retrieval-benchmark tradition of relevance judgments, precision, and recall is anchored by TREC. The information-scent framing follows Pirolli and Card's Information Foraging (Psychological Review, 1999). This is a fixture-scored evaluator, not a universal benchmark or a live-kernel proof.
A pass means the public fixture behavior and the bundle projection remain reproducible; it admits no component, unblocks no Atlas owner lane, and excludes launch. A pass means the public fixture behavior and the bundle projection remain reproducible. It does not admit an accepted component, unblock the Atlas owner lane, or include launch operations.
\n
Scope boundary
\n
Scope limit
\n
The strongest honest claim
\n
Over the public fixtures the evaluator scores route packets for stable-id recall and precision, catches forbidden routes, checks scent terms, and separates latency debt from sufficiency debt, matching the four cases exactly. The strongest honest claim: over the public fixtures, this evaluator scores route packets for stable-id recall and precision, catches forbidden first routes, checks scent terms, and separates latency debt from sufficiency debt, and its accounting matches the four fixture cases exactly. That is what the tests prove.
\n
What it refuses to claim
\n
It is not a live private kernel.py run, an embedding benchmark, a universal navigation benchmark, or launch-scope decision, and a live-kernel claim would need packets from the real route runner. It refuses more. It is not a live private kernel.py run, not an embedding benchmark, not a universal navigation benchmark, and not launch-scope decision. Its result records say nothing about unseen-task navigation quality, browser or provider state, or accepted-component admission. A live-kernel claim needs packets captured from the real route runner. The proof boundary is the fixture set and the scope limit is mechanism-level.
\n",
"components": [],
"doctrine_feed": {
"counts": {
"axiom": 7,
"concept": 1,
"mechanism": 1,
"paper_module": 2,
"principle": 8
},
"link_count": 19,
"links": [
{
"id": "mechanism.engine_room_navigation_fitness_benchmark.validates_public_navigation_fitness_benchmark",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public navigation fitness benchmark",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-navigation-fitness-benchmark-validates-public-navigation-fitness-benchmark"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-7",
"kind": "principle",
"kind_label": "principle",
"label": "Track known unknowns without claiming the unknown is mapped",
"relation": "governed by",
"url": "doctrine.html#dcard-p-7"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-11",
"kind": "principle",
"kind_label": "principle",
"label": "Bind volatile facts to refresh routes",
"relation": "governed by",
"url": "doctrine.html#dcard-p-11"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
},
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[31:paper_module.engine_room_demo].depends_on",
"url": "paper-module-engine-room-demo.html"
}
],
"paper_module_id": "paper_module.engine_room_navigation_fitness_benchmark",
"reader_packet": {
"abstract": "Public Engine Room component: route-packet benchmark evaluator for stable-id recall, precision, forbidden first routes, latency, and debt candidates.",
"authority_ceiling": "Public route-packet fixture evidence only; no live private kernel run, no embedding benchmark, no universal navigation benchmark, no launch-scope decision, no whole-system equivalence, no source-file changes, and no whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.engine_room_navigation_fitness_benchmark.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-engine-room-navigation-fitness-benchmark-diagram-1",
"html_url": "docs/paper-module-engine-room-navigation-fitness-benchmark.html#paper-module-engine-room-navigation-fitness-benchmark-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Fixture task expected ids, forbidden routes, latency budget, scent terms\"] --> E[\"evaluate_task\"] P[\"Captured route packet\"] --> E E --> S[\"Sufficiency ladder timeout, error, missing id, weak scent, forbidden route\"] E --> L[\"Latency status wall time vs budget\"] S --> R[\"Per-task result record\"] L --> R R --> B[\"evaluate_benchmark counts, p50/p95, debt candidates\"] B --> C[\"evaluate_case observed vs expected\"]",
"source_format": "mermaid",
"source_hash": "sha256:d02c1921dc906a7a43773119d8fb0fb0d370663750320daeed8b10e002ff1689",
"source_line_end": 141,
"source_line_start": 131
}
],
"grouped_relation_feeds": [
{
"count": 7,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-6",
"kind": "axiom",
"kind_label": "axiom",
"label": "Open-world epistemics",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-6"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-10",
"kind": "axiom",
"kind_label": "axiom",
"label": "Temporal validity and freshness contracts",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-10"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 2,
"links": [
{
"id": "paper_module.navigation_hologram_route_plane",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Navigation Hologram Route Plane",
"relation": "depends on",
"url": "paper-module-navigation-hologram-route-plane.html"
},
{
"id": "paper_module.engine_room_demo",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Engine Room Demo",
"relation": "depends on",
"url": "paper-module-engine-room-demo.html"
}
],
"relation": "depends on"
},
{
"count": 1,
"links": [
{
"id": "mechanism.engine_room_navigation_fitness_benchmark.validates_public_navigation_fitness_benchmark",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public navigation fitness benchmark",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-engine-room-navigation-fitness-benchmark-validates-public-navigation-fitness-benchmark"
}
],
"relation": "explains"
},
{
"count": 9,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-7",
"kind": "principle",
"kind_label": "principle",
"label": "Track known unknowns without claiming the unknown is mapped",
"relation": "governed by",
"url": "doctrine.html#dcard-p-7"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-11",
"kind": "principle",
"kind_label": "principle",
"label": "Bind volatile facts to refresh routes",
"relation": "governed by",
"url": "doctrine.html#dcard-p-11"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.architecture_and_navigation_route_contract_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "architecture and navigation route contracts as executable doctrine boundaries",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-architecture-and-navigation-route-contract-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.engine_room_navigation_fitness_benchmark",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 247,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_navigation_fitness_benchmark.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "TREC",
"section_id": "retrieval-evaluation-and-information-scent",
"url": "https://trec.nist.gov/"
},
{
"kind": "external_reference",
"label": "Information Foraging",
"section_id": "retrieval-evaluation-and-information-scent",
"url": "https://cir.nii.ac.jp/crid/1363951795634897280?lang=en"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The costly cold-start failure is not a wrong answer but the wrong first command, so this component scores whether a navigation packet pointed at the stable ids a cold task actually needed.** When an agent is dropped cold into a large repository, the costly failure is not a wrong answer. It is reaching for the wrong first command and landing on the wrong rows. This component makes that failure measurable. It reads cold-task fixtures and the route packets a navigation surface returned for them, then scores whether the packet pointed at the stable ids the task actually needed.",
"level": 4,
"order": 1,
"section_id": "costly-cold-start-failure",
"source_line_end": 13,
"source_line_start": 3,
"title": "Costly cold-start failure"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is an evaluator, not a route runner: it scores pre-captured packets for recall, precision, forbidden-route hits, scent coverage, and latency, without calling the navigation kernel or generating routes.** The component is an evaluator, not a route runner. It consumes fixture tasks and pre-captured packets, computes recall, precision, forbidden-route hits, scent coverage, and latency status, and emits one result record per task plus a suite-level summary. It does not call the private navigation kernel, does not validate embeddings, and does not generate routes of its own.",
"level": 4,
"order": 2,
"section_id": "is-evaluator-not",
"source_line_end": 24,
"source_line_start": 14,
"title": "Is evaluator not"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The runtime lives in src/microcosmcore/engineroom/navigationfitnessbenchmark.py.** The runtime lives in src/microcosmcore/engineroom/navigationfitnessbenchmark.py. Public fixtures live in fixtures/firstwave/engineroomnavigationfitness_benchmark/input.",
"level": 4,
"order": 3,
"section_id": "runtime-lives-in",
"source_line_end": 31,
"source_line_start": 25,
"title": "Runtime lives in"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 4,
"section_id": "purpose",
"source_line_end": 33,
"source_line_start": 32,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Most retrieval scores collapse correct and fast into one number, so a slow-but-correct route and a fast-but-wrong route look identical and land in the same repair queue when they need different fixes.** Most retrieval scores collapse \"correct\" and \"fast\" into one number. That is the wrong shape for navigation triage. A packet can name every expected stable id and still be too slow. A packet can come back fast and still miss an id, carry a weak information scent, use a forbidden first route, time out, or error. If those collapse into a single score, a slow-but-correct route and a fast-but-wrong route look the same, and they land in the same repair queue when they need different fixes.",
"level": 4,
"order": 5,
"section_id": "why-one-score-is-the-wrong-shape",
"source_line_end": 45,
"source_line_start": 34,
"title": "Why one score is the wrong shape"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The evaluator keeps two axes apart: sufficiency asks whether the route pointed at the right things, latency asks whether it stayed under budget, and each is computed and reported independently.** This evaluator keeps the two axes apart. Sufficiency answers \"did the route point at the right things.\" Latency answers \"did it stay under budget.\" The two are computed independently and reported side by side, so each failing route lands in the queue that matches its actual defect.",
"level": 4,
"order": 6,
"section_id": "sufficiency-and-latency-stay-apart",
"source_line_end": 54,
"source_line_start": 46,
"title": "Sufficiency and latency stay apart"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 56,
"source_line_start": 55,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**NavigationFitnessTask is a frozen dataclass fixing the task id, expected artifacts, forbidden routes, latency budget, and scent terms, and taskfrommapping degrades missing fields to visible public defaults.** The unit of input is a typed task, not a free-form prompt. NavigationFitnessTask is a frozen dataclass that fixes the task id, family, prompt, route type, expected stable artifacts, forbidden first routes, latency budget in milliseconds, route role, and scent terms. taskfrommapping turns each raw fixture row into that record. Missing fields fall back to explicit public defaults (unknown_task, publicfixture, contextpack, a 1500 ms budget), so a thin fixture degrades to visible default\u2026",
"level": 4,
"order": 8,
"section_id": "the-input-is-a-typed-task",
"source_line_end": 69,
"source_line_start": 57,
"title": "The input is a typed task"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "evaluatetask is the core predicate. evaluatetask is the core predicate. It runs these steps against one task and one packet:",
"level": 4,
"order": 9,
"section_id": "evaluate-task-is",
"source_line_end": 81,
"source_line_start": 70,
"title": "Evaluate task is"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluatetask computes recallatpacket and precisionatpacket, reads wallms, timedout, and error off the packet, and checks each forbidden route against its firstcontactcommand. From those, evaluatetask computes recallatpacket (found over expected) and precisionatpacket (found over selected), reads wallms, timedout, and error off the packet, and checks each forbidden route against the packet's firstcontactcommand.",
"level": 4,
"order": 10,
"section_id": "computing-recall-precision-and-route-checks",
"source_line_end": 90,
"source_line_start": 82,
"title": "Computing recall, precision, and route checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Sufficiency is a first-failing-check ladder over timeout, error, missing id, weak scent, and forbidden route, while latency is a separate three-way status that never touches the sufficiency verdict.** Sufficiency is a first-failing-check ladder. In order: a timeout gives failure kind route_timeout; an error gives routeerror; any missing expected id gives missingid; a failed scent check gives weakscent; a forbidden first route gives forbiddenroute. If none fire, sufficiency is pass. Latency is a separate three-way status: timeout if the packet timed out, pass if wall time is within budget, fail if it ran over. Latency never touches the sufficiency verdict.",
"level": 4,
"order": 11,
"section_id": "sufficiency-ladder-versus-latency-status",
"source_line_end": 102,
"source_line_start": 91,
"title": "Sufficiency ladder versus latency status"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "evaluatebenchmark runs the predicate over every case and aggregates counts, percentiles, and exactly two debt classes, so a slow-but-correct route yields one latency debt and zero sufficiency debt. evaluatebenchmark lifts the per-task predicate to a suite. It runs evaluate_task over every case, then aggregates counts, p50wallms, p95wallms, per-route-type metrics from routetypemetrics, and debt candidates from debt_candidates. That last function emits exactly two debt classes. A sufficiency debt names the missing ids and forbidden-route hits and carries the failure kind as repair_class. A latency debt records wall time, budget, and latency status without marking the route semantically wrong.\u2026",
"level": 4,
"order": 12,
"section_id": "suite-aggregation-and-two-debt-classes",
"source_line_end": 116,
"source_line_start": 103,
"title": "Suite aggregation and two debt classes"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "evaluatecase reports expectation_met only when observed status, every summary field, and every per-task field agree, and evaluatefixturedir passes only when every case matched. evaluatecase compares a suite result record against a fixture's stated expectations: an expectedstatus, an expectedsummary of counts, and expectedtaskstatuses per task. It reports expectation_met only when the observed status, every summary field, and every per-task field agree. evaluatefixturedir walks a directory of case files, runs each through evaluatecase, and returns casecount, passedcasecount, and an overall status that passes only when every case matched. build_parser and main expose two subcommands, evaluate\u2026",
"level": 4,
"order": 13,
"section_id": "comparing-receipts-against-expectations",
"source_line_end": 142,
"source_line_start": 117,
"title": "Comparing result records against expectations"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 14,
"section_id": "negative-cases",
"source_line_end": 144,
"source_line_start": 143,
"title": "Negative cases"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The four public fixture cases pin the accounting rules. The four public fixture cases pin the accounting rules. Each names its own task id.",
"level": 4,
"order": 15,
"section_id": "four-public-fixture",
"source_line_end": 149,
"source_line_start": 145,
"title": "Four public fixture"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "heldoutparaphrase_pass runs two nonliteral cold tasks that must route to their expected stable ids, avoid forbidden first routes, satisfy scent terms, and stay under budget. heldoutparaphrase_pass runs two nonliteral cold tasks (heldouttheoryroof, heldoutagentpath_store). Both must route to their expected stable ids, avoid forbidden first routes, satisfy scent terms, and stay under budget.",
"level": 4,
"order": 16,
"section_id": "two-paraphrased-cold-tasks-pass",
"source_line_end": 158,
"source_line_start": 150,
"title": "Two paraphrased cold tasks pass"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "adversarialforbidden_route selects the right id but uses the forbidden --paper-module first command, so sufficiency fails with forbidden_route while latency stays pass. adversarialforbidden_route runs adversarialpapermodule_query. The packet selects the right id (papermodules:navigationhologram_theory) but its first command uses the forbidden --paper-module route. Sufficiency fails with kind forbidden_route while latency stays pass, so a correct id does not rescue a banned first move.",
"level": 4,
"order": 17,
"section_id": "a-correct-id-with-a-forbidden-route",
"source_line_end": 168,
"source_line_start": 159,
"title": "A correct id with a forbidden route"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "missingstableidnegative selects a nearby row while the expected stable id is absent, so sufficiency fails with kind missingid. missingstableid_negative runs missingagentdiagnostics, where a nearby row is selected but the expected stable id is absent. Sufficiency fails with kind missing_id.",
"level": 4,
"order": 18,
"section_id": "a-nearby-row-misses-the-expected-id",
"source_line_end": 175,
"source_line_start": 169,
"title": "A nearby row misses the expected id"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "latencydebt_negative selects the expected id at 830 ms against a 100 ms budget, so sufficiency passes, latency fails, and the suite reports one latency debt and zero sufficiency debt. latencydebtnegative runs slowbut_sufficient. The packet selects the expected id with a 100 ms budget and an 830 ms wall time. Sufficiency passes, latency fails, and the suite reports one latency debt and zero sufficiency debt. This is the case that proves the two axes stay separate.",
"level": 4,
"order": 19,
"section_id": "slow-but-sufficient-proves-axis-separation",
"source_line_end": 184,
"source_line_start": 176,
"title": "Slow but sufficient proves axis separation"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 20,
"section_id": "prior-art-grounding",
"source_line_end": 186,
"source_line_start": 185,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The component applies information-retrieval evaluation and information-scent research to route packets: define expected targets, score recall and precision, penalize forbidden first moves, and keep latency apart from answer quality.** The component applies information-retrieval evaluation and information-scent research to agent route packets rather than to document search: define the expected targets, score returned results for recall and precision, penalize forbidden first moves, and keep latency apart from answer quality. The retrieval-benchmark tradition of relevance judgments, precision, and recall is anchored by TREC. The information-scent framing follows Pirolli and Card's Informati\u2026",
"level": 4,
"order": 21,
"section_id": "retrieval-evaluation-and-information-scent",
"source_line_end": 201,
"source_line_start": 187,
"title": "Retrieval evaluation and information scent"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "You can also replay the fixtures directly:",
"level": 2,
"order": 22,
"section_id": "validation-receipt-path",
"source_line_end": 215,
"source_line_start": 202,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A pass means the public fixture behavior and the bundle projection remain reproducible; it admits no component, unblocks no Atlas owner lane, and excludes launch.** A pass means the public fixture behavior and the bundle projection remain reproducible. It does not admit an accepted component, unblock the Atlas owner lane, or include launch operations.",
"level": 4,
"order": 23,
"section_id": "what-a-pass-means-here",
"source_line_end": 223,
"source_line_start": 216,
"title": "What a pass means here"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 24,
"section_id": "claim-ceiling",
"source_line_end": 225,
"source_line_start": 224,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Over the public fixtures the evaluator scores route packets for stable-id recall and precision, catches forbidden routes, checks scent terms, and separates latency debt from sufficiency debt, matching the four cases exactly.** The strongest honest claim: over the public fixtures, this evaluator scores route packets for stable-id recall and precision, catches forbidden first routes, checks scent terms, and separates latency debt from sufficiency debt, and its accounting matches the four fixture cases exactly. That is what the tests prove.",
"level": 4,
"order": 25,
"section_id": "the-strongest-honest-claim",
"source_line_end": 235,
"source_line_start": 226,
"title": "The strongest honest claim"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It is not a live private kernel.py run, an embedding benchmark, a universal navigation benchmark, or launch-scope decision, and a live-kernel claim would need packets from the real route runner.** It refuses more. It is not a live private kernel.py run, not an embedding benchmark, not a universal navigation benchmark, and not launch-scope decision. Its result records say nothing about unseen-task navigation quality, browser or provider state, or accepted-component admission. A live-kernel claim needs packets captured from the real route runner. The proof boundary is the fixture set and the scope limit is mechanism-level.",
"level": 4,
"order": 26,
"section_id": "what-it-refuses-to-claim",
"source_line_end": 246,
"source_line_start": 236,
"title": "What it refuses to claim"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/engine_room_navigation_fitness_benchmark.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "Public Engine Room component: route-packet benchmark evaluator for stable-id recall, precision, forbidden first routes, latency, and debt candidates."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "navigation_fitness_benchmark.py",
"path": "src/microcosm_core/engine_room/navigation_fitness_benchmark.py",
"resolution": "resolved",
"role": "Component runtime for public navigation-fitness fixtures, route-packet metrics, forbidden-route checks, latency status, debt candidate summaries, and public result record emission.",
"symbols": [
"NavigationFitnessTask",
"evaluate_task",
"evaluate_benchmark",
"evaluate_fixture_dir",
"main"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/navigation_fitness_benchmark.py"
},
{
"label": "demo.py",
"path": "src/microcosm_core/engine_room/demo.py",
"resolution": "resolved",
"role": "Accepted Engine Room composition locus that admits this component as a staged bundle target.",
"symbols": [
"CAPSULES",
"EXPECTED_JEWEL_TARGETS",
"run_demo",
"audit_controller_coverage"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/engine_room/demo.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_navigation_fitness_benchmark.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/engine_room_navigation_fitness_benchmark.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Engine Room Navigation Fitness Benchmark binds the staged navigation-fitness bundle to a concrete mechanism. It evaluates public route-packet fixtures for expected stable-id recall, precision, forbidden first routes, latency budgets, scent terms, and debt candidates without running the private kernel, validating embeddings, claiming universal benchmark authority, or upgrading launch/private-system status.",
"summary_line": "Public Engine Room component: route-packet benchmark evaluator for stable-id recall, precision, forbidden first routes, latency, and debt candidates.",
"title": "Engine Room Navigation Fitness Benchmark"
},
{
"anchor": "paper-module-cold-clone-probe",
"authority_ceiling": "Public source-root bootstrap mechanics and metadata-only result record refs only; no launch-scope decision, hosted-product readiness, external model access, source-file changes, whole-system equivalence, publishing-scope decision, or whole-system correctness.",
"body_html": "
Probe is first
\n
The probe is the first script a fresh clone runs, checking whether it can bootstrap the first-wave contract from source without leaking anything private.cold_clone_probe is the first script a reader runs after cloning the public checkout. It answers one question: can this clone bootstrap the bounded first-wave contract from source, without leaking anything private, before the reader installs the package or reads the long component inventory?
\n
Imports from src
\n
It imports from src/, checks fixture inputs, scans for secrets, replays the pattern-binding fixture, and writes one ignored local record; a pass proves only the first bounded action. The probe imports the package from src/, checks that the first-wave fixture inputs are present, runs a secret-exclusion scan, replays the pattern-binding fixture, and then writes one small result record. That record is ignored local state, not a tracked artifact. A pass means the first local action from a fresh clone works and stays inside the public boundary. It is not an install proof, a launch proof, or a statement about the rest of the system.
\n
Purpose
\n
Pulling the first-step check to the front
\n
A public checkout must prove its own first step, so the probe pulls the bootstrap check ahead of install or inventory into one command, one record, four typed failure modes. A public checkout has to prove its own first step. If the reader's first action is make install or a long tour of the component inventory, a broken or leaky bootstrap is discovered late and off to the side. The probe pulls that check to the front and makes it concrete: one command, one result record, four typed failure modes.
\n
Why a bare import-and-ok is not enough
\n
A probe that only imports and prints "ok" would pass on missing fixtures, an unavailable scan, or absent result records, so each failure gets a distinct blocked status and code instead. "Loose" here means a probe that only imports the package and prints "ok". That would pass on a clone that is missing its fixtures, or one whose secret scan is unavailable, or one where the pattern-binding result records never materialize. The probe refuses to call any of those a pass. Each is a distinct blocked status with its own code, so the reader learns which part of the bootstrap is wrong.
\n
How it works
\n
run_probe predicate and main wrapper
\n
The proof predicate run_probe runs a fixed sequence and never raises, while the CLI wrapper main parses flags, writes the --emit record, and sets the exit code. The runtime is one file, src/microcosm_core/cold_clone_probe.py. The proof predicate is run_probe(root, suite, emit_ref). It runs a fixed sequence over the checkout root and returns a result dict; it never raises. The CLI wrapper main parses --suite and --emit, calls run_probe on the current directory, writes the returned record, and exits 0 on pass or 1 otherwise. The one file that writes the --emit record is main, not run_probe.
\n
A reproducible command stamp before four gates
\n
run_probe stamps the base record with a reproducible ./bootstrap.sh command quoted through shlex.quote, then walks four gates in order.run_probe builds a metadata-only base record stamped with the reproducible command string from _bootstrap_command(suite, emit_ref), which returns ./bootstrap.sh --suite <suite> --emit <path> with both arguments quoted through shlex.quote. It then walks four gates in order:
\n
Function or constant
Role
SUPPORTED_SUITES
The only accepted suite is first-wave; anything else is rejected before fixture or scanner state is touched.
REQUIRED_INPUTS
The four fixture files the first-wave suite needs; each must be a readable file under the root.
validate_secret_exclusion_scan
Imported scan that must return status == "pass" before any pattern replay runs.
validate_pattern_binding
Imported validator that replays the first-wave fixture into a scratch directory under .microcosm/.
_mirror_missing_pattern_receipts
Copies the five PATTERN_RECEIPTS from the scratch run into their canonical clone-root slots.
DEFAULT_EMIT_REF
The default result path, .microcosm/cold_clone_probe.json, which stays ignored local state.
\n
Gate order and result record mirroring
\n
The gates run suite, inputs, secret scan, then replay so the boundary check gates fixture work, after which _mirror_missing_pattern_receipts copies missing result records into their canonical slots. The order matters. run_probe checks the suite first, then REQUIRED_INPUTS, then the secret scan, and only then the pattern-binding replay. The scan runs before replay on purpose: the boundary check gates the fixture work, not the other way around. After validate_pattern_binding writes into the scratch directory, _mirror_missing_pattern_receipts(root_path, source_dir) copies each canonical result record that is absent at the root but present in the scratch output. It skips destinations that already exist and sources that are not files, using _path_exists and _path_is_file, two probes that swallow OSError and report missing rather than crashing on a fresh clone with odd permissions. When it copies the validation-result result record it rewrites that file's receipt_paths field to the canonical PATTERN_RECEIPTS list, so the mirrored record points at the stable slots.
\n
What a pass record carries
\n
A pass record lists the emit path, the five result record refs, and the secret-scan summary with no private bodies, also aliased under private_state_scan for older readers. A pass record lists the emit path first, then the five pattern-binding result record refs, and carries the secret-exclusion scan summary with no private bodies copied in. The scan summary is also exposed under private_state_scan as a compatibility alias for older readers.
\nDiagram of the mechanism (12 steps).Diagram source & refs
Source refs
MISSING_FIXTURE_INPUT
blocked_dependency_missing
Validate secret exclusion scan
validate_secret_exclusion_scan
COMMAND_UNAVAILABLE
blocked_command_unavailable
SECRET_EXCLUSION_SCAN_BLOCKED
blocked_secret_exclusion
Validate pattern binding
validate_pattern_binding + mirror
MISSING_PATTERN_BINDING_RECEIPT
blocked_dependency_missing
flowchart TD A["Fresh public checkout"] --> B["run_probe(root, suite, emit_ref)"] B --> C{"suite in SUPPORTED_SUITES?"} C -- no --> X1["blocked_invalid_input UNKNOWN_COLD_CLONE_SUITE"] C -- yes --> D{"REQUIRED_INPUTS present?"} D -- no --> X2["blocked_dependency_missing MISSING_FIXTURE_INPUT"] D -- yes --> E["validate_secret_exclusion_scan"] E -- raises --> X3["blocked_command_unavailable COMMAND_UNAVAILABLE"] E -- status not pass --> X4["blocked_secret_exclusion SECRET_EXCLUSION_SCAN_BLOCKED"] E -- pass --> F["validate_pattern_binding + mirror"] F -- fail or missing --> X5["blocked_dependency_missing MISSING_PATTERN_BINDING_RECEIPT"] F -- all present --> P["status pass emit ref + five result record refs"]
\n
Negative cases
\n
Run probe never
\n
run_probe never raises.run_probe never raises. Every failure is a typed record the reader can read:
\n
blocked_invalid_input with code UNKNOWN_COLD_CLONE_SUITE: the requested suite is not first-wave. It also returns the supported list.
blocked_dependency_missing with code MISSING_FIXTURE_INPUT: one or more of REQUIRED_INPUTS is absent. The record names the missing files.
blocked_command_unavailable with code COMMAND_UNAVAILABLE: the secret scan raised, so the boundary cannot be checked. The exception text is recorded.
blocked_secret_exclusion with code SECRET_EXCLUSION_SCAN_BLOCKED: the scan ran but did not pass. The scan summary is attached, without leaking findings.
blocked_dependency_missing with code MISSING_PATTERN_BINDING_RECEIPT: the pattern-binding validator did not pass, or one of the five PATTERN_RECEIPTS is still absent after mirroring. The record names the missing result records.
\n
Prior Art Grounding
\n
Reproducible builds, twelve-factor, clean-checkout CI
\n
The shape borrows from reproducible builds, the twelve-factor dependency rule, and clean-checkout CI, applied to one bounded first action from source. The shape borrows from three public practices. Reproducible builds recreate outputs from declared sources and a constrained environment (reproducible-builds.org). The Twelve-Factor dependency rule declares dependencies rather than relying on ambient system packages (12factor.net). Clean-checkout CI on hosted runners is the common way to prove a fresh clone builds and smoke-tests (GitHub Actions). The probe keeps the clean-checkout, declared-input, gated shape and applies it to one bounded first action from source.
\n
Validation Result record Path
\n
Run ./bootstrap.sh from the public root. Use --dry-run first to print the exact command and the ignored result target without writing anything, then run it for real and check the focused tests:
The normal run writes an ignored result record, and a green pytest line means the predicate, wrapper, and public-entry docs agree, though a pass installs nothing and certifies no launch. The normal run writes ignored .microcosm/cold_clone_probe.json evidence with status=pass, the five public result record refs, and the secret-scan summary. Pass --emit <path> only to refresh an intentional local target; the default stays ignored local state. A green pytest line means the predicate, the root wrapper, and the public-entry docs all agree with the current checkout. These are the validation result records for this module; a pass here does not install the package, certify launch, or aggregate any wider coverage.
\n
Scope boundary
\n
Scope limit
\n
The bounded bootstrap claim and its ceiling
\n
The strongest claim is that a fresh clone can bootstrap the first-wave suite from src/ with typed failure codes; the boundary stops there, authorizing no launch or deployment. The strongest claim the evidence supports: a fresh public clone can bootstrap the first-wave suite from src/, with the fixture inputs present, the secret-exclusion scan passing, the pattern-binding fixture replaying, and all five pattern-binding result records materialized, and it records that with typed failure codes when any step is wrong. The proof boundary stops there. The probe excludes launch, hosted deployment, public sharing, external model access, secret export, package distribution, whole-system equivalence, or whole-system correctness. Its result record is ignored local state, not a tracked claim, and the scope limit for this module is the source-root bootstrap check plus its metadata-only result record. The generated Mermaid and Atlas views are projections of the source record, not independent authority.
\n",
"components": [],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"paper_module": 1,
"principle": 9
},
"link_count": 18,
"links": [
{
"id": "mechanism.cold_clone_probe.validates_public_source_root_bootstrap",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public source root bootstrap",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-cold-clone-probe-validates-public-source-root-bootstrap"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.entry_and_reveal_route_readiness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "entry and reveal routes as bounded cold-reader readiness evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-entry-and-reveal-route-readiness-bundle"
},
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [],
"paper_module_id": "paper_module.cold_clone_probe",
"reader_packet": {
"abstract": "The cold-clone probe validates the first public source-root bootstrap path: src import, secret-exclusion scan, first-wave pattern-binding fixture replay, public result record refs, and ignored local result record emission.",
"authority_ceiling": "Public source-root bootstrap mechanics and metadata-only result record refs only; no launch-scope decision, hosted-product readiness, external model access, source-file changes, whole-system equivalence, publishing-scope decision, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.cold_clone_probe.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-cold-clone-probe-diagram-1",
"html_url": "docs/paper-module-cold-clone-probe.html#paper-module-cold-clone-probe-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Fresh public checkout\"] --> B[\"run_probe(root, suite, emit_ref)\"] B --> C{\"suite in SUPPORTED_SUITES?\"} C -- no --> X1[\"blocked_invalid_input UNKNOWN_COLD_CLONE_SUITE\"] C -- yes --> D{\"REQUIRED_INPUTS present?\"} D -- no --> X2[\"blocked_dependency_missing MISSING_FIXTURE_INPUT\"] D -- yes --> E[\"validate_secret_exclusion_scan\"] E -- raises --> X3[\"blocked_command_unavailable COMMAND_UNAVAILABLE\"] E -- status not pass --> X4[\"blocked_secret_exclusion SECRET_EXCLUSION_SCAN_BLOCKED\"] E -- pass --> F[\"validate_pattern_binding + mirror\"] F -- fail or missing --> X5[\"blocked_dependency_missing MISSING_PATTERN_BINDING_RECEIPT\"] F -- all present --> P[\"status pass emit ref + five result record refs\"]",
"source_format": "mermaid",
"source_hash": "sha256:624463d0462763d03fa33d5cb1b9d9da3e88fdec4528139e69145c7c86ffa9c8",
"source_line_end": 121,
"source_line_start": 108
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.pattern_binding_contract",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Pattern Binding Contract",
"relation": "depends on",
"url": "paper-module-pattern-binding-contract.html"
}
],
"relation": "depends on"
},
{
"count": 1,
"links": [
{
"id": "mechanism.cold_clone_probe.validates_public_source_root_bootstrap",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public source root bootstrap",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-cold-clone-probe-validates-public-source-root-bootstrap"
}
],
"relation": "explains"
},
{
"count": 10,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.entry_and_reveal_route_readiness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "entry and reveal routes as bounded cold-reader readiness evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-entry-and-reveal-route-readiness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.cold_clone_probe",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 208,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/cold_clone_probe.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "reproducible-builds.org",
"section_id": "reproducible-builds-twelve-factor-clean-checkout-ci",
"url": "https://reproducible-builds.org/docs/"
},
{
"kind": "external_reference",
"label": "12factor.net",
"section_id": "reproducible-builds-twelve-factor-clean-checkout-ci",
"url": "https://12factor.net/dependencies"
},
{
"kind": "external_reference",
"label": "GitHub Actions",
"section_id": "reproducible-builds-twelve-factor-clean-checkout-ci",
"url": "https://github.com/features/actions"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The probe is the first script a fresh clone runs, checking whether it can bootstrap the first-wave contract from source without leaking anything private.** coldcloneprobe is the first script a reader runs after cloning the public checkout. It answers one question: can this clone bootstrap the bounded first-wave contract from source, without leaking anything private, before the reader installs the package or reads the long component inventory?",
"level": 4,
"order": 1,
"section_id": "probe-is-first",
"source_line_end": 11,
"source_line_start": 3,
"title": "Probe is first"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**It imports from src/, checks fixture inputs, scans for secrets, replays the pattern-binding fixture, and writes one ignored local record; a pass proves only the first bounded action.** The probe imports the package from src/, checks that the first-wave fixture inputs are present, runs a secret-exclusion scan, replays the pattern-binding fixture, and then writes one small result record. That record is ignored local state, not a tracked artifact. A pass means the first local action from a fresh clone works and stays inside the public boundary. It is not an install proof, a launch proof, or a statement about the rest of the system.",
"level": 4,
"order": 2,
"section_id": "imports-from-src",
"source_line_end": 23,
"source_line_start": 12,
"title": "Imports from src"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 3,
"section_id": "purpose",
"source_line_end": 25,
"source_line_start": 24,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A public checkout must prove its own first step, so the probe pulls the bootstrap check ahead of install or inventory into one command, one record, four typed failure modes.** A public checkout has to prove its own first step. If the reader's first action is make install or a long tour of the component inventory, a broken or leaky bootstrap is discovered late and off to the side. The probe pulls that check to the front and makes it concrete: one command, one result record, four typed failure modes.",
"level": 4,
"order": 4,
"section_id": "pulling-the-first-step-check-to-the-front",
"source_line_end": 36,
"source_line_start": 26,
"title": "Pulling the first-step check to the front"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A probe that only imports and prints \"ok\" would pass on missing fixtures, an unavailable scan, or absent result records, so each failure gets a distinct blocked status and code instead.** \"Loose\" here means a probe that only imports the package and prints \"ok\". That would pass on a clone that is missing its fixtures, or one whose secret scan is unavailable, or one where the pattern-binding result records never materialize. The probe refuses to call any of those a pass. Each is a distinct blocked status with its own code, so the reader learns which part of the bootstrap is wrong.",
"level": 4,
"order": 5,
"section_id": "why-a-bare-import-and-ok-is-not-enough",
"source_line_end": 47,
"source_line_start": 37,
"title": "Why a bare import-and-ok is not enough"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 6,
"section_id": "how-it-works",
"source_line_end": 49,
"source_line_start": 48,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "The proof predicate runprobe runs a fixed sequence and never raises, while the CLI wrapper main parses flags, writes the --emit record, and sets the exit code.** The runtime is one file, src/microcosmcore/coldclone_probe.py. The proof predicate is runprobe(root, suite, emitref). It runs a fixed sequence over the checkout root and returns a result dict; it never raises. The CLI wrapper main parses --suite and --emit, calls run_probe on the current directory, writes the returned record, and exits 0 on pass or 1 otherwise. The one file that writes the --emit record is main, not run_probe.",
"level": 4,
"order": 7,
"section_id": "run-probe-predicate-and-main-wrapper",
"source_line_end": 61,
"source_line_start": 50,
"title": "`run_probe` predicate and `main` wrapper"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "runprobe stamps the base record with a reproducible ./bootstrap.sh command quoted through shlex.quote, then walks four gates in order.** run_probe builds a metadata-only base record stamped with the reproducible command string from bootstrapcommand(suite, emit_ref), which returns ./bootstrap.sh --suite --emit with both arguments quoted through shlex.quote. It then walks four gates in order:",
"level": 4,
"order": 8,
"section_id": "a-reproducible-command-stamp-before-four-gates",
"source_line_end": 79,
"source_line_start": 62,
"title": "A reproducible command stamp before four gates"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The gates run suite, inputs, secret scan, then replay so the boundary check gates fixture work, after which mirrormissingpatternreceipts copies missing result records into their canonical slots.** The order matters. runprobe checks the suite first, then REQUIREDINPUTS, then the secret scan, and only then the pattern-binding replay. The scan runs before replay on purpose: the boundary check gates the fixture work, not the other way around. After validatepatternbinding writes into the scratch directory, mirrormissingpatternreceipts(rootpath, sourcedir) copies each canonical result record that is absent at the root but present in the scratch output. It skips destinations that already exist a\u2026",
"level": 4,
"order": 9,
"section_id": "gate-order-and-receipt-mirroring",
"source_line_end": 97,
"source_line_start": 80,
"title": "Gate order and result record mirroring"
},
{
"block_kinds": [
"diagram",
"paragraph"
],
"excerpt": "**A pass record lists the emit path, the five result record refs, and the secret-scan summary with no private bodies, also aliased under privatestatescan for older readers.** A pass record lists the emit path first, then the five pattern-binding result record refs, and carries the secret-exclusion scan summary with no private bodies copied in. The scan summary is also exposed under privatestatescan as a compatibility alias for older readers.",
"level": 4,
"order": 10,
"section_id": "what-a-pass-record-carries",
"source_line_end": 122,
"source_line_start": 98,
"title": "What a pass record carries"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 11,
"section_id": "negative-cases",
"source_line_end": 124,
"source_line_start": 123,
"title": "Negative cases"
},
{
"block_kinds": [
"list",
"paragraph"
],
"excerpt": "runprobe never raises. runprobe never raises. Every failure is a typed record the reader can read: blockedinvalidinput with code UNKNOWNCOLDCLONE_SUITE: the requested suite is not first-wave. It also returns the supported list. blockeddependencymissing with code MISSINGFIXTUREINPUT: one or more of REQUIRED_INPUTS is absent. The record names the missing files. blockedcommandunavailable with code COMMAND_UNAVAILABLE: the secret scan raised, so the boundary cannot be checked. The exception text is recorded. blockedsecretexclusion with code SECRETEXCLUSIONSCAN_BLOCKED: the scan ran but did not pass. The scan summary is attached, without leaking findings. blockeddependencymissing with code MISSI\u2026",
"level": 4,
"order": 12,
"section_id": "run-probe-never",
"source_line_end": 141,
"source_line_start": 125,
"title": "Run probe never"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 13,
"section_id": "prior-art-grounding",
"source_line_end": 143,
"source_line_start": 142,
"title": "Prior Art Grounding"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The shape borrows from reproducible builds, the twelve-factor dependency rule, and clean-checkout CI, applied to one bounded first action from source.** The shape borrows from three public practices. Reproducible builds recreate outputs from declared sources and a constrained environment (reproducible-builds.org). The Twelve-Factor dependency rule declares dependencies rather than relying on ambient system packages (12factor.net). Clean-checkout CI on hosted runners is the common way to prove a fresh clone builds and smoke-tests (GitHub Actions). The probe keeps the clean-checkout, declared-input, gated shape and applies it to one bounded first action from source.",
"level": 4,
"order": 14,
"section_id": "reproducible-builds-twelve-factor-clean-checkout-ci",
"source_line_end": 157,
"source_line_start": 144,
"title": "Reproducible builds, twelve-factor, clean-checkout CI"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 15,
"section_id": "validation-receipt-path",
"source_line_end": 159,
"source_line_start": 158,
"title": "Validation Result record Path"
},
{
"block_kinds": [
"code",
"paragraph"
],
"excerpt": "Run ./bootstrap.sh from the public root. Use --dry-run first to print the exact command and the ignored result target without writing anything, then run it for real and check the focused tests:",
"level": 4,
"order": 16,
"section_id": "run-bootstrap-sh",
"source_line_end": 175,
"source_line_start": 160,
"title": "Run bootstrap sh"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The normal run writes an ignored result record, and a green pytest line means the predicate, wrapper, and public-entry docs agree, though a pass installs nothing and certifies no launch.** The normal run writes ignored .microcosm/coldcloneprobe.json evidence with status=pass, the five public result record refs, and the secret-scan summary. Pass --emit only to refresh an intentional local target; the default stays ignored local state. A green pytest line means the predicate, the root wrapper, and the public-entry docs all agree with the current checkout. These are the validation result records for this module; a pass here does not install the package, certify launch, or aggregate an\u2026",
"level": 4,
"order": 17,
"section_id": "the-receipt-and-what-green-pytest-means",
"source_line_end": 188,
"source_line_start": 176,
"title": "The result record and what green pytest means"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 18,
"section_id": "claim-ceiling",
"source_line_end": 190,
"source_line_start": 189,
"title": "Scope limit"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The strongest claim is that a fresh clone can bootstrap the first-wave suite from src/ with typed failure codes; the boundary stops there, authorizing no launch or deployment.** The strongest claim the evidence supports: a fresh public clone can bootstrap the first-wave suite from src/, with the fixture inputs present, the secret-exclusion scan passing, the pattern-binding fixture replaying, and all five pattern-binding result records materialized, and it records that with typed failure codes when any step is wrong. The proof boundary stops there. The probe excludes launch, hosted deployment, public sharing, external model access, secret export, package distribution, whole-system equivale\u2026",
"level": 4,
"order": 19,
"section_id": "the-bounded-bootstrap-claim-and-its-ceiling",
"source_line_end": 207,
"source_line_start": 191,
"title": "The bounded bootstrap claim and its ceiling"
}
],
"source": {
"body_available": true,
"body_source_ref": "microcosm-substrate/paper_modules/cold_clone_probe.md",
"capsule_source_ref": "microcosm-substrate/core/paper_module_capsules.json"
},
"teleology": {
"section_id": "",
"source": "capsule_summary_line",
"text": "The cold-clone probe validates the first public source-root bootstrap path: src import, secret-exclusion scan, first-wave pattern-binding fixture replay, public result record refs, and ignored local result record emission."
}
},
"source_drilldown": {
"source_manifests": [],
"source_modules": [
{
"label": "cold_clone_probe.py",
"path": "src/microcosm_core/cold_clone_probe.py",
"resolution": "resolved",
"role": "Runtime and result record-writing locus for source-root bootstrap validation, first-wave pattern-binding fixture replay, public result record-ref mirroring, secret-exclusion scan enforcement, and ignored local result record emission.",
"symbols": [
"REQUIRED_INPUTS",
"PATTERN_RECEIPTS",
"DEFAULT_EMIT_REF",
"_mirror_missing_pattern_receipts",
"_bootstrap_command",
"run_probe",
"main"
],
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/src/microcosm_core/cold_clone_probe.py"
}
],
"source_records": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/cold_clone_probe.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
]
},
"source_links": [
{
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/cold_clone_probe.md"
},
{
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
}
],
"summary": "Cold Clone Probe is the public source-root bootstrap membrane. It binds bootstrap.sh, src/microcosm_core/cold_clone_probe.py, the first-wave pattern-binding fixture, the secret-exclusion scan, public relative result record refs, and focused tests so a fresh checkout has one bounded proof of first-run mechanics before install, CI, hosted launch, or full component inventory review.",
"summary_line": "The cold-clone probe validates the first public source-root bootstrap path: src import, secret-exclusion scan, first-wave pattern-binding fixture replay, public result record refs, and ignored local result record emission.",
"title": "Cold Clone Probe"
},
{
"anchor": "paper-module-first-screen-composition-root",
"authority_ceiling": "Public first-screen card composition and focused validation only; no launch-scope decision, hosted-publishing-scope decision, external model access authority, source-file changes, non-public data equivalence, score-based progress, reader-success certification, or whole-system correctness.",
"body_html": "
Component builds card
\n
This component builds the one card a new reader sees before choosing a deeper route into the public Plectis repository, hosting nothing, calling no provider, and claiming no finished repository. The first-screen composition root builds the one card a new reader sees before they pick a deeper route into the public Plectis repository. It is one Python component, microcosm_core.first_screen_composition, with a small command-line wrapper. It does not host anything, call a model provider, or claim the repository is finished.
\n
Output is card
\n
Its output is one card in three shapes, full JSON, budgeted compact JSON, and terminal text, all built from a standard and checked back against it so drift becomes a failed check. Its output is a single card in three shapes: a full JSON contract, a compact JSON card under a character budget, and a terminal-sized text card. Every shape carries the same shared first command, the same six reader branches, the same evidence accounting, and the same scope limit. The component builds the card from a standard file and then checks the built card back against that same standard, so a renamed field or a dropped branch shows up as a failed check instead of silent drift.
\n
Purpose
\n
The first-screen problem is not depth
\n
The repository already has deep surfaces; the first-screen problem is that depth lands badly when a cold reader first meets a long command list or a raw JSON dump. The repository already has deep surfaces: route maps, component status, standards, result records, and the localhost observatory. The first-screen problem is not missing depth. It is that depth lands badly when the first thing a reader meets is a long command list or a raw JSON dump.
\n
Prose has no contract, so it drifts
\n
A hand-written welcome card drifts because prose has no contract, so the composition root replaces it with a fixed structure answering what must fit on one screen and in what order. A loose, hand-written welcome card fails in a specific way. It drifts. Someone renames a branch, adds a command, or removes the scope limit, and nothing catches it, because prose has no contract. The composition root replaces prose with a fixed structure and answers one question: when a cold reader lands here, what has to fit on one screen, and in what order.
\n
The fixed slots on the card
\n
The fixed slots are a goal-entry line, a shared first command, one behavior proof, six reader branches, an evidence-count legend, a link to the wider system, a deeper-route list, and the scope limit. The fixed slots are the goal-entry line plectis comprehend --first-action "<your goal>", the shared first command plectis hello <project>, one shared behavior proof plectis tour --card <project>, six reader branches, an evidence-count legend read as accounting rather than a score, a link from the local command to the wider source-open system, a short list of the deeper routes the card names instead of copying, and the scope limit.
\n
How it works
\n
The order below is the order first_screen_composition_card runs.
\n
Function
Role
first_screen_composition_card
Loads the standard, assembles every card packet, runs both check passes, stamps status.
_validation_checks
Checks the built card is internally consistent.
_standard_backed_first_screen_scan
Checks the built card against the standard it was built from.
first_screen_compact_card
Projects the full card into a budgeted compact card with an omission list.
first_screen_text_card
Projects the card into a terminal-sized text screen, optionally for one reader.
\n
The builder assembles from the standard
\n
first_screen_composition_card loads the standard and assembles one large payload of shared commands, six reader routes, menus, proof fields, legend, counts, and a scope limit, free-handing none of it.first_screen_composition_card(root, project_label) is the builder. It loads standards/std_microcosm_first_screen_composition_root.json through _load_standard, then assembles one large payload: the shared commands, the six reader routes from _reader_routes, the copyable per-reader menu, the behavior-proof fields, the evidence-class legend, the scale counts, and the scope limit copied from the standard. It free-hands none of this. The authority_ceiling, anti_claim, and public_private_boundary are deep-copied straight from the standard.
\n
Internal consistency checks
\n
_validation_checks returns named booleans that are true only when each card surface satisfies its own contract, yielding a false check for any missing or malformed surface and never raising. Once the payload is built the component checks it twice. _validation_checks(payload) returns an ordered dictionary of named booleans. Each boolean is true only when one surface of the card satisfies its own contract: the six reader-route ids match REQUIRED_ROUTE_IDS, the reader menu and the landing packets cover those same ids, the evidence-class legend is complete, the scale counts are honest integers. A missing or malformed surface yields a false check. It never raises.
\n
Scanning the card against its standard
\n
_standard_backed_first_screen_scan confirms the card's routes, copyable commands, and six false denied-authority flags all match the standard, returning pass only when every check is true and never raising._standard_backed_first_screen_scan(payload, standard, check_ids) then compares the built card against the standard it came from. It confirms the standard reference and validator id match, that every reader route is present across all four route surfaces, that each reader carries the exact copyable commands plectis hello --reader <id> <label> and its text-projection form, and that the six denied-authority flags in DENIED_AUTHORITY_KEYS are false in both the standard and the card. Those six keys are launch, source-file changes, non-public data equivalence, external model access, score-based progress, and whole-system correctness. The scan returns a record whose status is "pass" only when every check is true. On any drift it returns "blocked" with the failing check named. It never raises.
\n
Folding the scan into status
\n
The builder folds the scan status into the check set, sets status to pass only when every check is true, and still returns a blocked card in full so the failing check is visible. The builder folds the scan status into the check set, sets validation.status to "pass" only when every check is true, and copies that to the top-level status. A blocked card is still returned in full, so a reader can see which check failed.
\n
Compact and text projections with budgets
\n
first_screen_compact_card produces a budgeted summary card with an omission list, and first_screen_text_card produces a 33-line terminal card that raises rather than truncating when it would overflow. Two projections render the same card for different media. first_screen_compact_card(payload) produces a summary-first card under COMPACT_JSON_CARD_MAX_CHARS (16000) characters, with an omission list naming the full-contract keys it dropped and the plectis first-screen --full command that restores them. first_screen_text_card(payload, reader_id) produces a terminal card of at most TEXT_CARD_MAX_LINES (33) lines over the same scope limit. Passing reader_id="all" shows every branch; an alias focuses one. It raises ValueError when the reader id is unknown or when the assembled card would exceed the line budget, so an over-long text card is a hard failure, not a truncated screen.
\n
The command-line wrapper
\n
The wrapper scripts/inspect/first_screen_composition_card.py exposes the card with --root, --project-label, --format, and --reader, returning exit code 0 only when the card status is pass. The command-line wrapper scripts/inspect/first_screen_composition_card.py exposes this without the full console. build_parser defines --root, --project-label, --format (json or text), and --reader. main builds the card, prints the JSON contract or the reader-focused text card, and returns exit code 0 only when the card status is "pass".
\nDiagram of the mechanism (6 steps).Diagram source & refs
flowchart TD A["Standard std_microcosm_first_screen_composition_root.json"] --> B["first_screen_composition_card assemble card packets"] B --> C["_validation_checks internal consistency"] B --> D["_standard_backed_first_screen_scan card vs standard"] C --> E["status: pass or blocked"] D --> E E --> F["first_screen_compact_card / first_screen_text_card JSON or terminal projection"]
\n
Negative cases
\n
What the checks force the card to reject
\n
Mismatched reader ids fail reader_route_parity, a wrong copyable command fails copyable_reader_commands, a non-false denied-authority flag fails denied_authority_flags_false, and an over-33-line text card raises instead of printing. The scan and the checks name what the card must reject. A card whose reader branches do not match the six required ids fails reader_route_parity. A card missing a copyable per-reader command, or carrying a command that does not match the expected plectis hello --reader <id> string, fails copyable_reader_commands. A card where any of the six denied-authority flags is not false fails denied_authority_flags_false. A text card that would run past 33 lines raises instead of printing. Each of these is a blocked result with the failing check named, never a quiet pass.
\n
Prior Art Grounding
\n
Command-line usability and progressive disclosure
\n
The card borrows from command-line usability practice, progressive disclosure, and W3C PROV: a concise shared command, a terminal card, a JSON contract, and evidence counts kept attached to their scope. The card borrows from command-line usability practice rather than inventing an onboarding genre. The Command Line Interface Guidelines argue for concise default help, worked examples, discoverable next commands, and machine-readable output. Those appear here as the one shared command, the terminal-sized card, and the JSON contract. The compression rule follows progressive disclosure: reveal enough to orient a reader without dumping the whole system, the pattern Nielsen Norman Group documents as progressive disclosure. W3C PROV informs the rule that evidence counts stay attached to their source and scope limit rather than becoming standalone scores.
\n
Validation Result record Path
\n
Reader-verifiable emitter commands, run from the microcosm-substrate/ public root:
A pass means the built card matched its standard and both projections stayed inside their budgets; the emitter writes no state and certifies no reader success or launch-scope decision. A pass means the built card matched its standard and the two projections stayed inside their budgets. The emitter writes no project state and prints the card to stdout. It does not certify that any reader will succeed or that the repository is ready to launch.
\n
Scope boundary
\n
Scope limit
\n
The narrow proof boundary
\n
The strongest claim is narrow: first-screen composition, reader-branch routing, the omission list, the text projection, and the scope limit are reproducible from public source and tests, and nothing more. The strongest claim this component supports is narrow: the first-screen card is built from a public standard and checked back against it, so first-screen composition, reader-branch routing, the omission list, the text projection, and the scope limit are reproducible from public source and tests. That is the proof boundary. Its scope limit stops there. It does not establish runtime correctness, replace the route map, certify launch or hosted public sharing, use external model services, change project source, turn counts into maturity scores, or cover the whole doctrine lattice. A blocked scan reports drift. It does not repair anything.
\n",
"components": [],
"doctrine_feed": {
"counts": {
"axiom": 6,
"concept": 1,
"mechanism": 1,
"paper_module": 1,
"principle": 9
},
"link_count": 18,
"links": [
{
"id": "mechanism.first_screen_composition_root.validates_public_first_screen_composition_root",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public first screen composition root",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-first-screen-composition-root-validates-public-first-screen-composition-root"
},
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
},
{
"id": "concept.entry_and_reveal_route_readiness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "entry and reveal routes as bounded cold-reader readiness evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-entry-and-reveal-route-readiness-bundle"
},
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
}
],
"schema": "microcosm_public_paper_module_doctrine_feed_v1",
"source_policy": "capsule_typed_relation_fields_only_no_prose_inference"
},
"omitted_fields": [
"code_loci",
"legacy_markdown_projection",
"strangler_note",
"generated_projections",
"source_authority"
],
"paper_module_backrefs": [
{
"id": "paper_module.public_reveal_walkthrough",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Public Reveal Walkthrough",
"relation": "depended on by",
"relation_id": "paper_module.depends_on.paper_module",
"source_ref": "core/paper_module_capsules.json::paper_modules[28:paper_module.public_reveal_walkthrough].depends_on",
"url": "paper-module-public-reveal-walkthrough.html"
}
],
"paper_module_id": "paper_module.first_screen_composition_root",
"reader_packet": {
"abstract": "The first-screen composition root validates the public one-screen entry card, reader branches, omission result records, evidence accounting frame, text projection, README order, and scope limit without becoming launch or hosted-publishing-scope decision.",
"authority_ceiling": "Public first-screen card composition and focused validation only; no launch-scope decision, hosted-publishing-scope decision, external model access authority, source-file changes, non-public data equivalence, score-based progress, reader-success certification, or whole-system correctness.",
"body_fallback": {
"available": true,
"field": "body_html",
"role": "rendered_html_fallback_not_machine_source"
},
"diagrams": [
{
"asset_id": "paper_module.first_screen_composition_root.diagram.1",
"cdn_dependency": false,
"html_anchor": "paper-module-first-screen-composition-root-diagram-1",
"html_url": "docs/paper-module-first-screen-composition-root.html#paper-module-first-screen-composition-root-diagram-1",
"kind": "mermaid",
"render_status": "build_time_svg_available",
"rendered_asset_role": "body_html_build_time_svg",
"source": "flowchart TD A[\"Standard std_microcosm_first_screen_composition_root.json\"] --> B[\"first_screen_composition_card assemble card packets\"] B --> C[\"_validation_checks internal consistency\"] B --> D[\"_standard_backed_first_screen_scan card vs standard\"] C --> E[\"status: pass or blocked\"] D --> E E --> F[\"first_screen_compact_card / first_screen_text_card JSON or terminal projection\"]",
"source_format": "mermaid",
"source_hash": "sha256:dfdd414c5d6b7a777e8eadb9677e8354718f1281b8800303ae972f452077ccc6",
"source_line_end": 153,
"source_line_start": 145
}
],
"grouped_relation_feeds": [
{
"count": 6,
"links": [
{
"id": "AX-1",
"kind": "axiom",
"kind_label": "axiom",
"label": "Derivation before assertion",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-1"
},
{
"id": "AX-4",
"kind": "axiom",
"kind_label": "axiom",
"label": "Content-addressed determinism",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-4"
},
{
"id": "AX-5",
"kind": "axiom",
"kind_label": "axiom",
"label": "Fail-closed monotone lattice",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-5"
},
{
"id": "AX-7",
"kind": "axiom",
"kind_label": "axiom",
"label": "Typed partiality and refusal",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-7"
},
{
"id": "AX-8",
"kind": "axiom",
"kind_label": "axiom",
"label": "Provenance propagation and non-interference",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-8"
},
{
"id": "AX-11",
"kind": "axiom",
"kind_label": "axiom",
"label": "Executable grammar before doctrine authority",
"relation": "abides by",
"url": "doctrine.html#dcard-ax-11"
}
],
"relation": "abides by"
},
{
"count": 1,
"links": [
{
"id": "paper_module.cold_reader_route_map",
"kind": "paper_module",
"kind_label": "paper module",
"label": "Cold-Reader Route Map",
"relation": "depends on",
"url": "paper-module-cold-reader-route-map.html"
}
],
"relation": "depends on"
},
{
"count": 1,
"links": [
{
"id": "mechanism.first_screen_composition_root.validates_public_first_screen_composition_root",
"kind": "mechanism",
"kind_label": "mechanism",
"label": "validates public first screen composition root",
"relation": "explains",
"url": "rules-and-ideas.html#lattice-mechanism-mechanism-first-screen-composition-root-validates-public-first-screen-composition-root"
}
],
"relation": "explains"
},
{
"count": 10,
"links": [
{
"id": "P-1",
"kind": "principle",
"kind_label": "principle",
"label": "Recompute, do not echo",
"relation": "governed by",
"url": "doctrine.html#dcard-p-1"
},
{
"id": "P-2",
"kind": "principle",
"kind_label": "principle",
"label": "Lower claim strength to checker strength",
"relation": "governed by",
"url": "doctrine.html#dcard-p-2"
},
{
"id": "P-3",
"kind": "principle",
"kind_label": "principle",
"label": "Concentrate trust in small checkers",
"relation": "governed by",
"url": "doctrine.html#dcard-p-3"
},
{
"id": "P-5",
"kind": "principle",
"kind_label": "principle",
"label": "Cache by content, not by name",
"relation": "governed by",
"url": "doctrine.html#dcard-p-5"
},
{
"id": "P-6",
"kind": "principle",
"kind_label": "principle",
"label": "Status fails closed",
"relation": "governed by",
"url": "doctrine.html#dcard-p-6"
},
{
"id": "P-8",
"kind": "principle",
"kind_label": "principle",
"label": "Refuse inadmissible computations with typed reasons",
"relation": "governed by",
"url": "doctrine.html#dcard-p-8"
},
{
"id": "P-9",
"kind": "principle",
"kind_label": "principle",
"label": "Preserve provenance across every boundary",
"relation": "governed by",
"url": "doctrine.html#dcard-p-9"
},
{
"id": "P-12",
"kind": "principle",
"kind_label": "principle",
"label": "Make doctrine executable before authoritative",
"relation": "governed by",
"url": "doctrine.html#dcard-p-12"
},
{
"id": "P-15",
"kind": "principle",
"kind_label": "principle",
"label": "Keep projections below source authority",
"relation": "governed by",
"url": "doctrine.html#dcard-p-15"
},
{
"id": "concept.entry_and_reveal_route_readiness_bundle",
"kind": "concept",
"kind_label": "concept",
"label": "entry and reveal routes as bounded cold-reader readiness evidence",
"relation": "governed by",
"url": "rules-and-ideas.html#lattice-concept-concept-entry-and-reveal-route-readiness-bundle"
}
],
"relation": "governed by"
}
],
"math_latex": [],
"module_id": "paper_module.first_screen_composition_root",
"omissions": {
"body_html_omitted_from_packet": true,
"omission_reason": "reader_packet_carries_structured_public_slots_with_body_html_as_page_fallback",
"omitted_sections": [
{
"reason": "public_projection_internal_section_denylist",
"section_id": "json-capsule-binding",
"source_line_start": 237,
"title": "JSON Bundle Binding"
}
]
},
"references": [
{
"kind": "source_link",
"label": "Design note",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/paper_modules/first_screen_composition_root.md"
},
{
"kind": "source_link",
"label": "Source registry",
"url": "https://github.com/wcook04/practical-formal-verification/blob/main/core/paper_module_capsules.json"
},
{
"kind": "external_reference",
"label": "Command Line Interface Guidelines",
"section_id": "command-line-usability-and-progressive-disclosure",
"url": "https://clig.dev/"
},
{
"kind": "external_reference",
"label": "progressive disclosure",
"section_id": "command-line-usability-and-progressive-disclosure",
"url": "https://www.nngroup.com/articles/progressive-disclosure/"
},
{
"kind": "external_reference",
"label": "W3C PROV",
"section_id": "command-line-usability-and-progressive-disclosure",
"url": "https://www.w3.org/TR/prov-overview/"
}
],
"schema": "microcosm_public_module_reader_packet_v1",
"sections": [
{
"block_kinds": [
"paragraph"
],
"excerpt": "**This component builds the one card a new reader sees before choosing a deeper route into the public Plectis repository, hosting nothing, calling no provider, and claiming no finished repository.** The first-screen composition root builds the one card a new reader sees before they pick a deeper route into the public Plectis repository. It is one Python component, microcosmcore.firstscreen_composition, with a small command-line wrapper. It does not host anything, call a model provider, or claim the repository is finished.",
"level": 4,
"order": 1,
"section_id": "component-builds-card",
"source_line_end": 12,
"source_line_start": 3,
"title": "Component builds card"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**Its output is one card in three shapes, full JSON, budgeted compact JSON, and terminal text, all built from a standard and checked back against it so drift becomes a failed check.** Its output is a single card in three shapes: a full JSON contract, a compact JSON card under a character budget, and a terminal-sized text card. Every shape carries the same shared first command, the same six reader branches, the same evidence accounting, and the same scope limit. The component builds the card from a standard file and then checks the built card back against that same standard, so a renamed field or a dropped branch shows up as a failed check instead of silent drift.",
"level": 4,
"order": 2,
"section_id": "output-is-card",
"source_line_end": 24,
"source_line_start": 13,
"title": "Output is card"
},
{
"block_kinds": [],
"excerpt": null,
"level": 2,
"order": 3,
"section_id": "purpose",
"source_line_end": 26,
"source_line_start": 25,
"title": "Purpose"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The repository already has deep surfaces; the first-screen problem is that depth lands badly when a cold reader first meets a long command list or a raw JSON dump.** The repository already has deep surfaces: route maps, component status, standards, result records, and the localhost observatory. The first-screen problem is not missing depth. It is that depth lands badly when the first thing a reader meets is a long command list or a raw JSON dump.",
"level": 4,
"order": 4,
"section_id": "the-first-screen-problem-is-not-depth",
"source_line_end": 35,
"source_line_start": 27,
"title": "The first-screen problem is not depth"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**A hand-written welcome card drifts because prose has no contract, so the composition root replaces it with a fixed structure answering what must fit on one screen and in what order.** A loose, hand-written welcome card fails in a specific way. It drifts. Someone renames a branch, adds a command, or removes the scope limit, and nothing catches it, because prose has no contract. The composition root replaces prose with a fixed structure and answers one question: when a cold reader lands here, what has to fit on one screen, and in what order.",
"level": 4,
"order": 5,
"section_id": "prose-has-no-contract-so-it-drifts",
"source_line_end": 45,
"source_line_start": 36,
"title": "Prose has no contract, so it drifts"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "**The fixed slots are a goal-entry line, a shared first command, one behavior proof, six reader branches, an evidence-count legend, a link to the wider system, a deeper-route list, and the scope limit.** The fixed slots are the goal-entry line plectis comprehend --first-action \"\", the shared first command plectis hello , one shared behavior proof plectis tour --card , six reader branches, an evidence-count legend read as accounting rather than a score, a link from the local command to the wider source-open system, a short list of the deeper routes the card names instead of copying, and the scope limit.",
"level": 4,
"order": 6,
"section_id": "the-fixed-slots-on-the-card",
"source_line_end": 56,
"source_line_start": 46,
"title": "The fixed slots on the card"
},
{
"block_kinds": [
"paragraph",
"table"
],
"excerpt": "The order below is the order firstscreencomposition_card runs.",
"level": 2,
"order": 7,
"section_id": "how-it-works",
"source_line_end": 68,
"source_line_start": 57,
"title": "How it works"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "firstscreencompositioncard loads the standard and assembles one large payload of shared commands, six reader routes, menus, proof fields, legend, counts, and a scope limit, free-handing none of it. `firstscreencompositioncard(root, project_label)` is the builder. It loads standards/stdmicrocosmfirstscreencomposition_root.json through loadstandard, then assembles one large payload: the shared commands, the six reader routes from readerroutes, the copyable per-reader menu, the behavior-proof fields, the evidence-class legend, the scale counts, and the scope limit copied from the standard. It free-hands none of this. The scope_limit, anticlaim, and publicprivate_boundary are deep-copied straig\u2026",
"level": 4,
"order": 8,
"section_id": "the-builder-assembles-from-the-standard",
"source_line_end": 82,
"source_line_start": 69,
"title": "The builder assembles from the standard"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "validationchecks returns named booleans that are true only when each card surface satisfies its own contract, yielding a false check for any missing or malformed surface and never raising.** Once the payload is built the component checks it twice. validationchecks(payload) returns an ordered dictionary of named booleans. Each boolean is true only when one surface of the card satisfies its own contract: the six reader-route ids match REQUIREDROUTEIDS, the reader menu and the landing packets cover those same ids, the evidence-class legend is complete, the scale counts are honest integers. A missing or malformed surface yields a false check. It never raises.",
"level": 4,
"order": 9,
"section_id": "internal-consistency-checks",
"source_line_end": 94,
"source_line_start": 83,
"title": "Internal consistency checks"
},
{
"block_kinds": [
"paragraph"
],
"excerpt": "standardbackedfirstscreen_scan confirms the card's routes, copyable commands, and six false denied-authority flags all match the standard, returning pass only when every check is true and never raising.** standardbackedfirstscreenscan(payload, standard, checkids) then compares the built card against the standard it came from. It confirms the standard reference and validator id match, that every reader route is present across all four route surfaces, that each reader carries the exact copyable commands plectis hello --reader