Cold Evaluation Honesty Bundle

Runs a copied route-quality simulator and checks its all-B scorecard against the original code.

Contract

Does This component imports the real cold_eval.py route-quality simulator as an exact source copy. Running it over a synthetic workspace inspects the all-B scorecard shape, source-module digest evidence, and scope limit checks without exporting body text in result records or turning the fixture into benchmark truth.

Scope limit verified cold-eval source body import only, not a live benchmark, navigation truth, source authority, external model access, whole-system equivalence, public sharing, or launch-scope decision

Run

Quick probe

microcosm batch10-cold-eval-honesty-capsule run --input fixtures/first_wave/batch10_cold_eval_honesty_capsule/input --out /tmp/plectis-public-runs/batch10_cold_eval_honesty_capsule

Validator command

microcosm batch10-cold-eval-honesty-capsule run --input fixtures/first_wave/batch10_cold_eval_honesty_capsule/input --out receipts/first_wave/batch10_cold_eval_honesty_capsule --acceptance-out receipts/acceptance/first_wave/batch10_cold_eval_honesty_capsule_fixture_acceptance.json

Evidence

ClassVerified source importevidence 5/5Copied source body

ai-safetyagent-evaluationred-teaming

Source

Source Design note · Source

Map Show on the map

Index Open compact index card

Paper module

Paper module Set 10 Cold Eval Honesty Bundle

Set 10 Cold Eval Honesty Bundle runs cold-eval over public fixtures without benchmark, navigation-truth, or launch-scope decision.

Read the full write-up

Back to Components index → Reference roster →