What problem does it solve? Skill authors often ship directories with malformed frontmatter, mismatched names, or broken progressive-disclosure structures that prevent agents from discovering and loading skills correctly. This Skill audits every skill directory in a repository against the agentskills.io spec and Claude Code authoring guidelines so defects are caught before publication. ## Core Features & Use Cases - Spec Compliance Checks: Verifies name format (1-64 chars, lowercase alphanumeric with hyphens), name-to-directory matching, description length (1-1024 chars), and validity of optional fields like license and metadata. - Best Practice Linting: Enforces third-person descriptions, body length under 500 lines, one-level-deep reference loading, markdown link formatting, and conciseness rules. - Structured Reporting: Produces per-skill PASS/FAIL results in a consistent report format. - Use Case: Before publishing a new skill to a shared repository, run the validation to confirm the SKILL.md frontmatter matches its directory name and that every reference file is reachable directly from SKILL.md. ## Quick Start Ask the agent to run /validate-skills to audit all skill directories in the skills folder and report any spec or best-practice violations.
