What problem does it solve?
Over time, Claude skill libraries accumulate outdated, redundant, or low-quality skills that bloat your workspace, reduce AI response quality, and create confusion about which tools to use. This skill automates the end-to-end audit process to categorize every skill, so you can focus on maintaining a lean, effective library.
Core Features & Use Cases
- Dual Audit Modes: Run a Quick Scan to evaluate only recently changed skills (5–10 minutes) or a Full Stocktake to review your entire skill library (20–30 minutes) with no manual tracking.
- Holistic Quality Evaluation: Uses a standardized checklist to assess content overlap, technical freshness, usage frequency, and actionability, with AI-powered judgment to assign clear verdicts for each skill.
- Usage Tracking & Persistent Results: Automatically pulls 7-day and 30-day usage metrics for each skill, and caches audit results to avoid redundant work on future runs.
- Use Case: If you have 50+ custom Claude skills and are unsure which are still useful, run this skill to get a prioritized list of actions (keep, improve, update, retire, or merge) to clean up your library efficiently.
Quick Start
Use the skill-stocktake skill to audit your global and project-level Claude skills, then review the generated summary table of verdicts and actionable recommendations for each skill.
