skill-audit

You can audit custom AI skills by running an automated structural compliance check across your skill ecosystem to detect dead cross-references, missing frontmatter fields, and tool permission mismatches that degrade agent workflow reliability.

Functional gaps in AI agent skills are unenforced promises, silent failure risks, and context leaks not caught by structural checks. You detect them by running deep functional gap analysis with advanced checks across your skill batches.

Yes, you can limit audits to specific skill groups such as MAD Core or Review/Lint. Batch auditing allows you to target configurable skill groups for focused structural checks and functional analysis.

Yes, skill audit supports projects using Claude Code or similar agent frameworks with custom skills stored in .claude/skills, applying both quick structural checks and deep functional analysis to maintain workflow reliability.

You can automatically apply fixes for common structural issues like missing documentation sections and version bumps by enabling the optional automated remediation feature during your skill ecosystem audit.

You should run a skill ecosystem compliance check before a release or major workflow update to catch broken references, stale versions, and missing documentation that would otherwise cause agent failures in production.