Skill Evaluator is a specialized tool for developers building Claude Code skills, providing a rigorous framework for auditing skill quality. It classifies skills into four distinct types—Reference, Workflow, Interactive, or Principle—and performs a multi-dimensional assessment covering token efficiency, structural integrity, and emotional tone. By generating detailed scored reports, concrete rewrite recommendations, and specific test scenarios, it helps developers refine their agentic capabilities from basic prompts into highly reliable, professional-grade tools.
主な機能
01Detailed markdown report generation with prioritized improvement recommendations
02Generation of edge-case test scenarios for reliability verification
03Automated skill classification based on structure and usage patterns
04Multi-dimensional scoring against universal and type-specific rubrics
052 GitHub stars
06Emotional tone analysis and alignment for enhanced user interaction
