skills.rest
Thedougler
ThedouglerCOMMUNITY
Explore all skills by this author →
GitHub

skill-judge

Evaluate Agent Skill design quality against official specifications and best practices.

Updated Feb 16, 2026
Search Tags:#quality assurance#prompt engineering#agent development#skill evaluation#ai skill design#knowledge delta
One-click install
npx skills add https://github.com/Thedougler/agent-template --skill skill-judge-thedougler
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-judge
Source: https://github.com/Thedougler/agent-template/tree/main/.github/skills/skill-judge
Command: npx skills add https://github.com/Thedougler/agent-template --skill skill-judge-thedougler

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill addresses the inefficiency and token waste in poorly designed AI Skills by providing a rigorous evaluation framework. It helps identify redundant information, structural flaws, and areas for improvement, ensuring Skills add genuine expert value.

Core Features & Use Cases

  • Skill Evaluation: Scores Skills across 8 dimensions (Knowledge Delta, Mindset, Anti-Patterns, etc.) for a total of 120 points.
  • Pattern Identification: Recognizes and assesses adherence to 5 official Skill design patterns (Mindset, Navigation, Philosophy, Process, Tool).
  • Failure Pattern Detection: Identifies 9 common pitfalls like "The Tutorial" or "The Invisible Skill".
  • Use Case: Before deploying a new Skill, use Skill Judge to audit its quality, ensuring it's concise, valuable, and adheres to best practices, thereby maximizing its effectiveness and minimizing token usage.

Quick Start

Evaluate the skill at .github/skills/my-new-skill/SKILL.md.

Frequently Asked Questions about skill-judge

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

FAQPage Schema
How do I evaluate AI agent skill quality before deployment?▼

To evaluate AI agent skill quality, audit the skill against official specifications to score dimensions like knowledge delta, mindset transfer, and anti-pattern adherence, generating a structured report with critical issues and improvement recommendations.

What are common failure patterns in AI skill design?▼

Common failure patterns in AI skill design include pitfalls like "The Tutorial" or "The Invisible Skill," which signal redundant information or structural flaws that cause token waste and reduce practical usability.

How can I audit prompt engineering skills for specification compliance?▼

Audit prompt engineering skills for specification compliance by analyzing pattern recognition, progressive disclosure, and freedom calibration across 5 official design patterns, yielding a multi-dimensional score out of 120 points.

Does skill evaluation help reduce token waste in agent development?▼

Skill evaluation reduces token waste in agent development by identifying redundant information and structural flaws, ensuring the skill adds genuine expert value and adheres to best practices for maximum effectiveness.

What is the best way to score agent skills against best practices?▼

The best way to score agent skills against best practices is to assess adherence to 5 official design patterns—Mindset, Navigation, Philosophy, Process, and Tool—across 8 dimensions, providing actionable improvement suggestions.

Related Skills

Explore More Software Engineering Skills→
galligangalligan
1

skills-check

Validate and review SKILL.md files against the Agent Skills specification.

Community
Basic
opencodosopencodos
14

skill-judge

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

Official
Advanced
fabioc-alohafabioc-aloha
1

Architecture Refinement Skill

Document and consolidate learned patterns and process improvements into structured markdown files.

Community
Intermediate
gohypergiantgohypergiant
21

skill-judge

Evaluates METAMIR SKILLs across eight dimensions including correctness, HITL readiness, and meta-skill abuse detection.

Official
Advanced