What problem does it solve?
Analyzes a tested Skill's SKILL.md to perform static analysis, extract structure and references, and conduct semantic evaluation across five dimensions; it also gathers evaluation context from the user and outputs skill-analysis.yaml for downstream steps.
Core Features & Use Cases
- Static analysis: extract structure, token counts, API references, API coverage, and reference integrity.
- Semantic analysis: assess intent clarity, instruction ambiguity, scenario coverage, API chain logic, and edge case gaps.
- Context collection: prompt the user to fill in target_user, known_failures, success_criteria, and forbidden_behaviors, feeding the resulting user_context.
- Output: generates eval/evaluations/{skill-name}/skill-analysis.yaml to guide the rest of the evaluation pipeline.
- Use case: serves as the first phase in the end-to-end skill evaluation workflow, ensuring the tested skill is well-defined before test generation.
Quick Start
Run the analyze command on your skill directory to generate eval/evaluations/{skill-name}/skill-analysis.yaml.