Agent Skills: skill-quality

Rates Claude Code skills with numerical scores (1-5) across 6 quality dimensions. Use when evaluating skill quality, scoring skills, rating customizations, or comparing skill effectiveness. Provides weighted scores and quality tier assessment.

UncategorizedID: philoserf/claude-code-setup/skill-quality

Install this agent skill to your local

pnpm dlx add-skill https://github.com/philoserf/claude-code-setup/tree/HEAD/skills/skill-quality

Skill Files

Browse the full folder contents for skill-quality.

Download Skill

Loading file tree…

skills/skill-quality/SKILL.md

Skill Metadata

Name
skill-quality
Description
Rates Claude Code skills with numerical scores (1-5) across 6 quality dimensions. Use when evaluating skill quality, scoring skills, or comparing skill effectiveness. Produces weighted scores, quality tier assessment, and verification results.

Reference Files

Detailed scoring guidance:


Purpose

This skill provides objective quality assessment of Claude Code skills using a standardized 6-dimension scoring system. Unlike cc-lint (structural validation) or cc-check (functional testing), this skill focuses on quality measurement with numerical scores.

Quality Dimensions (Weighted)

| Dimension | Weight | Focus | | -------------------- | ------ | ----------------------------------------------- | | Effectiveness | 28% | Does it achieve its stated purpose? | | Clarity | 22% | Is documentation clear and understandable? | | Best Practices | 17% | Progressive disclosure, context economy, safety | | Documentation | 15% | Completeness and organization of docs | | Verification | 10% | Can you confirm the output is correct? | | Trigger Coverage | 8% | Will users discover and invoke it? |

Quality Tiers

Based on weighted average score:

| Range | Tier | Meaning | | ------- | -------------------- | ------------------------------------------- | | 4.5-5.0 | Production Ready | Excellent quality, no improvements needed | | 3.5-4.4 | Good | Solid quality, minor improvements possible | | 2.5-3.4 | Needs Work | Functional but requires attention | | 1.5-2.4 | Poor | Significant issues, major rework needed | | 1.0-1.4 | Unusable | Fundamental problems, likely non-functional |

Evaluation Process

  1. Locate the skill - Find SKILL.md and any reference files
  2. Read all content - Examine main file and supporting documentation
  3. Score each dimension - Apply rubric criteria (see scoring-rubric.md)
  4. Calculate weighted average - Compute overall score
  5. Determine quality tier - Map score to tier
  6. Generate report - Include dimension scores and evidence (see report-template.md)

Scoring Guidelines

For each dimension:

  1. Review the dimension criteria in quality-dimensions.md
  2. Apply the 1-5 rubric from scoring-rubric.md
  3. Document specific evidence supporting the score
  4. Note any borderline cases or scoring rationale

Scoring principles:

  • Be specific - Cite exact text, files, or patterns as evidence
  • Be fair - Consider the skill's intended scope and purpose
  • Be consistent - Apply the same standards across all skills
  • Document reasoning - Explain why each score was given

Tools Used

This skill uses read-only tools for analysis:

  • Read - Examine SKILL.md and reference files
  • Glob - Find all files in the skill directory
  • Grep - Search for patterns and keywords
  • Bash - Run wc and other read-only commands

No files are modified during evaluation.

Usage Examples

Evaluate a specific skill:

Rate the quality of the cc-lint skill

Compare multiple skills:

Score cc-lint and cc-check, which is higher quality?

Quick quality check:

What's the quality tier for vc-ship?

Output

Reports include:

  • Individual dimension scores with evidence
  • Weighted average score
  • Quality tier classification
  • Dimension-specific observations
  • Optional comparison to other skills

See report-template.md for the complete output format.

Relationship to Other Tools

| Tool | Purpose | | ------------------- | --------------------------------- | | cc-lint | Structural validation (pass/fail) | | cc-check | Functional testing (works/broken) | | skill-quality | Quality scoring (1-5 scale) | | skill-improve | Improvement recommendations |

Run cc-lint first for structural validation, then skill-quality for subjective quality scoring.

skill-quality Skill | Agent Skills