Component · spec & evidence

Agent Completion Faithfulness Audit

Runs real git and pytest on a sample repo so wrap-up claims state only what the evidence proves.

Contract

Does Runs a public fixture repo through real git and pytest subprocesses, then checks that completion claims only say what the evidence supports: commit object exists, ledger cap exists, and pytest pass is claimed only after exit-zero status was checked.

Scope limit verified means the referenced evidence object exists or a pytest span ran; it does not imply the span passed unless exit-zero status was explicitly checked

Run

plectis agent-closeout-faithfulness-audit run --input fixtures/first_wave/agent_closeout_faithfulness_audit/input --out receipts/first_wave/agent_closeout_faithfulness_audit

Evidence

ClassExternal tool runevidence 4/5Real runtime resultRuns real tools

ai-safetyagent-evaluationred-teaming

Source

Source Design note · Source

Map Show on the map

Index Open compact index card

Paper module