What problem does it solve? Skill and command collections grow stale over time, accumulating outdated, overlapping, or low-quality entries. This Skill audits Claude skills and commands against a quality checklist combined with holistic AI judgment, producing actionable verdicts (Keep, Improve, Update, Retire, Merge) for each entry. ## Core Features & Use Cases - Two Audit Modes: Quick Scan re-evaluates only skills changed since the last run using mtime comparison against a cached results.json; Full Stocktake performs a complete inventory and evaluation. - Subagent Batch Evaluation: Inventories are evaluated in chunks of ~20 skills per general-purpose subagent, with intermediate progress saved and resume support for interrupted runs. - Structured Results Cache: Verdicts, reasons, and mtimes are persisted to results.json with accurate UTC timestamps, enabling incremental future scans. - Use Case: After months of accumulating custom skills, run a Full Stocktake to identify which skills duplicate each other, which reference outdated APIs, and which should be retired or merged, then confirm consolidation actions. ## Quick Start Run the skill-stocktake audit on my global and project-level Claude skills and give me a verdict table with consolidation recommendations.
