4-stage evaluation framework for testing Claude Code plugin component triggering. Validates skills, agents, and commands activate correctly via programmatic detection and LLM judgment.

Benchmark harness for A/B testing Claude Code plugins against OOLONG long-context reasoning tasks. Compare truncation vs RLM-RS recursive chunking strategies. Features Claude Code hooks integration, SQLite persistence, and comprehensive scoring aligned with the OOLONG paper methodology.

Testing framework for Codify plugins with complete lifecycle testing, IPC communication, and cross-platform support

🚀 Automate the evaluation of Claude Code plugin components to ensure accurate triggering of skills, agents, commands, and hooks.