plugin-audit

Auditing Vercel plugin performance involves extracting tool calls from Claude Code conversation logs, testing hook matching against actual inputs, identifying pattern coverage gaps, and checking plugin cache staleness. This process ensures optimal plugin skill injection and execution.

Plugin hook matching tests Vercel plugin hooks against actual conversation inputs to ensure correct skill execution. By validating hook patterns with real-world data, you can verify that the right skills trigger for the right tasks and identify any pattern coverage gaps.

Yes, you need a Node.js environment to check Vercel plugin cache staleness. The cache audit compares the installed plugin cache against the development version, ensuring your project uses up-to-date plugin versions for optimal performance and avoiding stale cache issues.

Pattern coverage gaps occur when Vercel plugin hooks fail to match actual conversation inputs extracted from Claude Code logs. Auditing these logs identifies missing patterns, allowing you to update hook definitions and improve overall plugin matching efficiency.

Yes, you can test Vercel plugin hooks against actual Claude Code conversation logs. The audit process extracts tool calls from these logs, validates hook matching against real inputs, and highlights discrepancies between expected and actual skill execution patterns.

The best way to investigate plugin skill injection is by running an automated audit that extracts conversation log tool calls, tests hook matching, identifies pattern gaps, and checks cache versions. This provides a comprehensive performance overview for Vercel plugins.