What problem does it solve? When a Claude Code plugin injects skills based on tool-call patterns, it is hard to know whether the matching rules actually fire on real-world usage. This Skill audits vercel-plugin performance by replaying actual conversation logs against the hook pipeline, revealing coverage gaps, dedup masking, and stale plugin caches. ## Core Features & Use Cases - Log Parsing: Extracts Read, Edit, Write, and Bash tool calls from Claude Code JSONL conversation logs. - Hook Match Testing: Imports the plugin's loadSkills and matchSkills functions directly to test pattern matching against real inputs without shelling out per call. - Gap & Staleness Analysis: Identifies path and Bash pattern gaps, dedup masking, budget drops, and diffs between the installed plugin cache and the dev version. - Use Case: After a long coding session on a Next.js project, run the batch-match script against the session's JSONL log to see which skills were injected, which were deduped, and which tool calls should have matched but did not. ## Quick Start Ask the AI to audit plugin skill injection on your project by analyzing the latest Claude Code conversation log and reporting match coverage gaps.
