What problem does it solve?
This Skill helps you verify that a Vercel plugin behaves correctly in real Claude Code sessions, catching hook failures, missing injections, deduplication bugs, and coverage gaps before they reach users.
Core Features & Use Cases
- Live Session Evaluation: Launch and observe real agent sessions so you can confirm hook execution in the same environment where the plugin will actually run.
- Debug Log Inspection: Monitor session logs in real time to validate session start, prompt submission, tool usage, post-tool validation, and session end behavior.
- Deduplication and Coverage Reporting: Compare injection events against atomic claim files and compile a structured report of coverage, issues, and validation outcomes.
- Use Case: A maintainer wants to test whether a new plugin release injects the right skills exactly once per scenario and produces a clean coverage summary across multiple runs.
Quick Start
Run a live Claude Code eval session with the plugin installed, then inspect the debug logs and claim files to confirm hook firing, deduplication, and coverage.