skill-judge

Evaluates Agent Skill quality against official specifications with multi-dimensional scoring.

Updated Mar 31, 2026
One-click install
npx skills add https://github.com/SESELOVSKYDarian/Vase --skill skill-judge-seselovskydarian
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-judge
Source: https://github.com/SESELOVSKYDarian/Vase/tree/main/.agents/skills/skill-judge
Command: npx skills add https://github.com/SESELOVSKYDarian/Vase --skill skill-judge-seselovskydarian

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Most Skills waste tokens on knowledge the model already has, and poorly written descriptions prevent Skills from ever being activated. This Skill provides a systematic evaluation framework to measure a Skill's knowledge delta, structural quality, and usability against official specifications. ## Core Features & Use Cases - Eight-Dimension Scoring: Scores Skills across 120 points covering knowledge delta, mindset transfer, anti-patterns, specification compliance, progressive disclosure, freedom calibration, pattern recognition, and usability. - Failure Pattern Detection: Identifies 9 common failure patterns such as The Tutorial, The Dump, Orphan References, and The Invisible Skill. - Structured Reports: Produces a standardized evaluation report with grades (A-F), critical issues, and prioritized improvement suggestions. - Use Case: Before publishing a new SKILL.md, run an evaluation to verify the description contains WHAT, WHEN, and trigger keywords, and that redundant content is eliminated. ## Quick Start Evaluate the skill at skills/my-new-skill/SKILL.md and generate a full scoring report with 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 the quality of a SKILL.md file?

Score the Skill across eight dimensions totaling 120 points, with knowledge delta weighted highest at 20 points. Classify each section as Expert, Activation, or Redundant content, then check frontmatter validity, structure, and usability to produce a graded report.

What makes a good Skill description for agent activation?

A good description answers three questions: WHAT the Skill does, WHEN it should be used, and which KEYWORDS trigger it. Agents only see descriptions before loading, so vague descriptions mean the Skill never gets activated regardless of content quality.

What is knowledge delta in Skill design?

Knowledge delta is the gap between what a Skill provides and what the model already knows. Good Skills contain expert-only knowledge like decision trees, trade-offs, and anti-patterns, while deleting redundant explanations of basic concepts that waste context tokens.

How long should a SKILL.md file be?

SKILL.md should stay under 500 lines, ideally under 300. Heavy content belongs in references/ files loaded on demand through explicit triggers embedded in workflow steps, following the three-layer progressive disclosure model.

What are the most common Skill design failure patterns?

Nine failure patterns are documented, including The Tutorial (explaining basics), The Dump (800+ line files), Orphan References (files never loaded), and The Invisible Skill (great content but poor description). Each pattern has a specific root cause and fix.