What problem does it solve? Governance rules, hooks, and skills accumulate over time without visibility into which ones still fire, which are dead weight, and which user corrections should be codified. This Skill runs a monthly, read-only metric scan that turns raw telemetry into concrete retire, upgrade, and codify proposals. ## Core Features & Use Cases - Metric Harvesting: Reads opt-in telemetry logs (hook-fires, skill-invokes, user-corrections, metric-snapshots) through a canonical runtime resolver that rejects symlink and path-escape attempts. - Auto-Propose Engine: Flags hooks with zero fires in 6 months as retire candidates, rules firing over 50 times as Meta-Pattern upgrade candidates, and accumulated user corrections as codify candidates. - Health Report Generation: Produces a month-over-month trend report covering instruction line counts, hook fires, skill invokes, benchmark freshness, and infra-reference integrity. - Use Case: A repository maintainer asks whether any governance rules should be retired; the Skill scans six months of hook-fire telemetry, finds three dead hooks, and proposes them for the knowledge-prune queue without mutating any policy files. ## Quick Start Ask the assistant to run a governance health scan and report which hooks or rules should be retired or upgraded.
