What problem does it solve?
Custom skill files for coding agents often have structural gaps, missing compliance requirements, or behavioral inconsistencies that break when deployed. This Skill eliminates manual, time-consuming reviews of skill files by automating compliance checks, behavioral validation, and category-specific rubric evaluation.
Core Features & Use Cases
- Static Structural Linter: Runs 7 automated compliance checks on skill files to catch missing frontmatter fields, absent phase structures, and missing collaborative protocol language.
- Behavioral Spec Verifier: Evaluates skill instructions against predefined test cases to confirm they produce expected behavior for given project fixtures.
- Category Rubric Evaluator: Checks skills against domain-specific quality metrics for their assigned category (e.g., gate, review, utility).
- Coverage Audit Tool: Generates a full report of testing status across all skills and agents in a project, highlighting untested high-priority items.
Use Case: A team maintaining a library of 50+ custom Claude Code skills can run this tool to identify non-compliant skills, fill testing gaps, and ensure all skills meet quality standards before sharing with the team.
Quick Start
Invoke the skill-test skill to run a full static compliance audit of all skill files in your project's skills directory.
