What problem does it solve? Skills degrade over time: descriptions stop triggering, prose bloats with redundant lines, and domain terminology drifts from current documentation. This Skill statically reviews an existing SKILL.md, ranks concrete findings by severity, and applies the fixes the user approves. ## Core Features & Use Cases - Multi-dimension audit: Reviews triggering, scope, prose leanness, structure, frontmatter validity, and repo-specific authoring conventions, with each finding tied to specific lines. - Domain fidelity checks: Verifies terminology and patterns in domain-specific skills against current documentation via the ctx7 CLI, WebFetch, or WebSearch, citing sources for every finding. - Critique-then-edit workflow: Presents findings ranked as Blocking, Important, or Nit, then offers to apply all, some, or a user-picked subset of fixes. - Use Case: You drafted a skill that never seems to trigger. Point this Skill at its SKILL.md and it flags that the description lacks the recommended "Use when ..." phrasing and overlaps with a neighboring skill, then rewrites it after you approve. ## Quick Start Review the skill at skills/agent-docs/review-skill/SKILL.md and list the findings by severity.
