completion-proof — anomaly A11 detector

Package scope and prerequisites

This catalog ships instructions, not the executable runtime. Python modules, scripts, local services and optional tools named below must be installed separately from the upstream project. Check their availability before running commands; if they are missing, explain the prerequisite instead of claiming execution or results. The host agent may use model tokens while following these instructions.

Commands, APIs and runtime behavior described below are conditional examples for a compatible upstream installation; they are not evidence of execution in this package.

The upstream detector is described by A11 of the upstream docs/local-analysis-policy.md (not bundled here): a completion claim without proof is an anomaly. Proof is one of:

• structured proof.test_id / proof.cmd_output_ref / proof.artifact_hash
• the same keys at the report root
• an agent_state completion whose validations contain a passed item with a non-empty reference
• lexical evidence in text reports (N passed, exit_code=0, pytest, snapshot=, sha256:)

python -m skills.completion_proof.cli [path]        # human-readable report
python -m skills.completion_proof.cli .botte --json # compact JSON
python -m skills.completion_proof.test_completion_proof

Scans .json / .jsonl plus report-named .md/.txt/.html files. The documented upstream mode is report-only; verify the installed command exit status and CI integration. Pairs with code_complexity (A1) and code_duplication (A6).