This skill should be used when the user asks to 'review a skill', 'audit skill quality', 'validate a skill', 'check my skill', 'improve my skill', 'skill feedback', or wants expert assessment of a SKILL.md against Anthropic best practices and the open Agent Skills specification. Performs structural validation, content quality analysis, triggering assessment, and token efficiency review. NOT for creating skills — use /skill-creator instead.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
This skill should be used when the user asks to 'review a skill', 'audit skill quality', 'validate a skill', 'check my skill', 'improve my skill', 'skill feedback', or wants expert assessment of a SKILL.md against Anthropic best practices and the open Agent Skills specification. Performs structural validation, content quality analysis, triggering assessment, and token efficiency review. NOT for creating skills — use /skill-creator instead.
/skill-content-reviewer {path-or-name} — review a specific skill
/skill-content-reviewer — no arg = ask user which skill to review
Workflow
Step 1: Locate & Read Skill
Resolve the skill path. Check in order:
Exact path provided
.claude/skills/{name}/
.github/skills/{name}/
~/.dotfiles/claude/project/*/skills/{name}/
~/.dotfiles/claude/global/skills/{name}/
Read SKILL.md and inventory all files in the skill directory
If not found, list available skills and ask user to pick
Step 2: Research the Skill's Domain Content
Verify the skill's domain advice is accurate and current using WebSearch:
Identify the skill's domain (e.g., React, SEO, Java entities, marketing copy)
Search for current best practices in that domain: {domain} best practices {year}
Search for any recent changes or deprecations: {domain} changes deprecated {year}
Synthesize findings to verify whether the skill's instructions and references contain accurate, up-to-date domain knowledge. Flag any outdated advice, missing current patterns, or incorrect recommendations.
Skip only if the skill is purely procedural with no domain advice (e.g., a skill that just locates files or manages skill directories). The skill's folder location doesn't matter — an architecture skill in meta/ still has domain content that needs verification.
Step 3: Structural Validation
Run every check in references/structural-checks.md against the skill. Any FAIL = automatic overall FAIL. Record each result in the output table.
Step 4: Content Quality Analysis
Score each of the 6 dimensions 1-5 using references/quality-rubric.md:
Description Quality — WHAT + WHEN + triggers, specificity balance