What problem does it solve? It is hard to know which Skills are actually being used, which ones fire automatically when they should, and which descriptions or trigger keywords need improvement. This Skill parses Claude Code transcript logs and turns them into concrete usage statistics. ## Core Features & Use Cases - Invocation Counting: Separately counts automatic (model-invoked) and explicit (slash command) Skill activations per Skill. - Trigger Hit-Rate Analysis: Matches user messages against declared trigger keywords and measures how often the matching Skill auto-fires afterward. - Flexible Filtering: Filter by date range, Skill name, or project, with per-project breakdowns and Markdown or JSON output. - Use Case: Run a 90-day audit across all projects to find Skills with low hit rates, then rewrite their descriptions or trigger keywords to improve automatic activation. ## Quick Start Ask the assistant to run the skill-stats analysis over the last 90 days of Claude Code transcripts and show the per-Skill hit-rate table.