skills.rest
shawn-sandy
shawn-sandyCOMMUNITY
Explore all skills by this author →
Website𝕏@shawnsandyGitHub

reviewing-skills

Scores SKILL.md files against Anthropic authoring best practices across five quality dimensions.

Updated Jun 8, 2026
Search Tags:#best practices#claude code#skill.md#skill audit#quality scoring#frontmatter validation#regression check
One-click install
npx skills add https://github.com/shawn-sandy/agentics-kit --skill reviewing-skills-shawn-sandy
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: reviewing-skills
Source: https://github.com/shawn-sandy/agentics-kit/tree/main/kit/plugins/skill-reviewer/skills/reviewing-skills
Command: npx skills add https://github.com/shawn-sandy/agentics-kit --skill reviewing-skills-shawn-sandy

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Skill authors lack a structured way to verify that their SKILL.md files follow Anthropic's official authoring best practices, leading to poorly discoverable, oversized, or malformed skills that fail to activate reliably. ## Core Features & Use Cases - Scored 5-Dimension Audit: Evaluates frontmatter validity, body quality, structure and progressive disclosure, anti-pattern detection, and discoverability on a 10-point scale with a letter grade. - Regression Risk Check: Compares the current SKILL.md against its last git commit to flag breaking changes like renamed skills or removed trigger phrases. - Automated Fix Generation: Produces an optimized version of the file with corrected frontmatter and inline suggestion comments, writing to disk only after explicit confirmation. - Use Case: After drafting a new skill, ask for a review to receive a scored report listing errors, warnings, and suggestions, then accept a corrected version before publishing. ## Quick Start Ask the assistant to review and score the SKILL.md file at a specific path against skill authoring best practices.

Frequently Asked Questions about reviewing-skills

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I review a SKILL.md file for best practices?▼

Provide the path to your SKILL.md file and ask for a review. The skill reads the file, measures lines and word count, then scores five dimensions on a 10-point scale and outputs a graded report with errors, warnings, and suggestions.

What makes a good SKILL.md description field?▼

A good description is third person, under 1024 characters, contains a "Use when..." trigger phrase, and includes a capability statement with at least three searchable keywords. It must not contain XML tags or the reserved words anthropic or claude.

Can the skill detect breaking changes in a SKILL.md update?▼

Yes, the optional regression risk check compares the current file against its last git commit. It flags BREAKING changes like a renamed name field or removed trigger phrases, and WARNINGs for removed reference files or large content reductions.

Does this skill review CLAUDE.md files or commands?▼

No, the audit applies only to SKILL.md files. It explicitly does not review CLAUDE.md, commands, hooks, or general markdown documents.

Will the skill overwrite my SKILL.md file automatically?▼

No, writing to disk requires explicit confirmation after the optimized version is presented. Without a clear confirmation such as "yes, write it", no changes are made to the file.

Related Skills

Explore More Software Engineering Skills→
hidonguyenhidonguyen

ck:template-skill

Scaffold Claude skill directories with a starter SKILL.md and YAML frontmatter.

Community
Basic
vinhnguyenn020325-cmdvinhnguyenn020325-cmd

vercel-plugin-eval

Execute Claude Code sessions to validate Vercel plugin hooks and generate coverage reports.

Community
Intermediate
openaiopenai
3.6k

runtime-behavior-probe

Plan and execute runtime behavior investigations using disposable TypeScript probes.

Official
Advanced
DipakMandlikDipakMandlik

skill-tester

Validate Claude skills for structure, Python syntax, and quality scoring.

Community
Advanced