reviewing-agent-effectiveness

You can audit agent enforcement mechanisms by generating a per-task report that details which rules, hooks, and skills activated. This diagnostic produces a FIRED, NOT FIRED, or NOT APPLICABLE status for each mechanism.

An agent enforcement firing report is used to evaluate whether rules, skills, and hooks gripped as intended during non-trivial implementations. It identifies enforcement gaps across L2, L3, L5, and pre-commit layers to guide improvements.

You should run an audit of agent rules and hook performance after completing non-trivial implementations, refactors, or audits. This timing ensures you accurately assess whether Always-Apply Rules and Pre-Commit Hooks performed as expected.

Yes, you can check pre-commit hook and stop hook performance across tasks. The audit evaluates these specific enforcement layers alongside Always-Apply Rules and Skills, reporting their activation status.

Agent enforcement rules might not fire due to gaps in the enforcement configuration across L2, L3, L5, or pre-commit layers. An audit report identifies which mechanisms are NOT FIRED, guiding configuration improvements.