skills.rest
pianzhu
pianzhuCOMMUNITY
Explore all skills by this author →
GitHub

skill-judge

Evaluates submitted ML models and returns structured JSON with predictions, confidence scores, and metadata.

1|Updated Nov 21, 2025
Search Tags:#quality-assurance#frontmatter#progressive-disclosure#skill-evaluation#loading-triggers#dimension-score
One-click install
npx skills add https://github.com/pianzhu/my-claude-skills --skill skill-judge-pianzhu
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-judge
Source: https://github.com/pianzhu/my-claude-skills/tree/main/skills/skill-judge
Command: npx skills add https://github.com/pianzhu/my-claude-skills --skill skill-judge-pianzhu

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a rigorous, repeatable framework for evaluating Claude Skills against official specifications and best practices. It helps auditors identify gaps between current SKILL.md documentation, structure, and activation behavior, enabling faster improvement cycles.

Core Features & Use Cases

  • Multi-dimensional scoring: D1–D8 evaluation dimensions (Knowledge Delta, Mindset/Procedures, Anti-Patterns, Specification Compliance, Progressive Disclosure, Freedom Calibration, Pattern Recognition, Practical Usability) to quantify Skill quality.
  • Structured evaluation protocol: Step-by-step process from frontmatter validity to detailed dimension scoring and a final verdict.
  • Activation-ready output: Provides a report template with dimension scores, critical issues, top improvements, and recommended patterns for Skill design.
  • Self-evaluation capability: The Skill can be used to calibrate other Skills and as a tutorial for building higher-quality Skills.
  • Use cases: Auditing own Skills, comparing SKILL.md across packages, guiding improvements, training new Skill authors.

Quick Start

To begin evaluating a target Skill, provide its SKILL.md to this Skill and request a structured evaluation report. The Skill will return dimension scores, a verdict, and actionable improvement suggestions.

Frequently Asked Questions about skill-judge

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

FAQPage Schema
How do I evaluate Claude SKILL.md files for specification compliance?▼

You can audit Claude Skills for quality by evaluating SKILL.md files against official specifications across eight dimensions. This surfaces knowledge gaps and provides actionable improvement suggestions for frontmatter, loading triggers, and usability.

What is the progressive disclosure pattern in Claude Skill design?▼

Progressive disclosure in Claude Skill design is an evaluation dimension measuring how effectively a Skill reveals information incrementally. It checks whether the Skill structure avoids overwhelming the user while maintaining necessary activation context.

How do I score Claude Skills across multiple quality dimensions?▼

You score Claude Skills by applying a structured evaluation protocol across eight dimensions, including Knowledge Delta, Anti-Patterns, Specification Compliance, and Practical Usability. This generates a quantified report with a final verdict.

Can I use this skill evaluation framework to audit my own custom Skills?▼

Yes, you can audit your own custom Skills. The framework supports self-evaluation capabilities, allowing authors to calibrate their SKILL.md documentation, identify anti-patterns, and guide improvements for higher-quality Skill design.

What are the limitations of automated SKILL.md frontmatter validation?▼

Automated SKILL.md frontmatter validation focuses strictly on structural compliance and loading triggers. It does not assess subjective creative quality or external integration logic, limiting its scope to specification adherence and practical usability metrics.

Related Skills

Explore More Design & Creative Skills→
mgajewskikmgajewskik
6

skill-judge

Evaluate Skills across eight dimensions and produce a formal Skill Evaluation Report.

Community
Advanced
opencodosopencodos
14

skill-judge

Evaluate AI skills across an 8-dimension rubric and return structured scores.

Official
Advanced
LusciiLuscii

mermaid-diagrams

Embed Mermaid diagrams in Markdown documents for visualizations.

Official
Basic
christopherlouetchristopherlouet
5

dev-refactor

Detect code smells and generate refactoring suggestions for JavaScript files.

Community
Intermediate