Evaluate SKILL.md design quality across D1-D8 dimensions and generate improvement plans.
npx skills add https://github.com/sandalsoft/agent-skills --skill skill-judge-sandalsoft
Please help me install this Agent Skill. Skill: skill-judge Source: https://github.com/sandalsoft/agent-skills/tree/main/skill-judge Command: npx skills add https://github.com/sandalsoft/agent-skills --skill skill-judge-sandalsoft
The Skill Judge identifies, quantifies, and reports on the quality of a Skill's design by comparing it against official specifications and best practices. It facilitates consistent evaluation, governance, and actionable improvement suggestions for Skill units.
Run the Skill Judge on a target Skill's SKILL.md to produce a complete evaluation and improvement plan.
High-intent search queries and answers about installing and using this skill.
Evaluate SKILL.md quality by comparing it against official specifications and best practices using multi-dimensional analysis across D1-D8 dimensions to generate a structured report with actionable improvement recommendations.
Common anti-patterns in skill design include missing expert knowledge, lacking actionable domain procedures, and poor structural formatting, which are identified through pattern recognition during a comprehensive skill evaluation.
Audit skill packages for production readiness by validating the SKILL.md file to uncover genuine expert knowledge, actionable procedures, and anti-patterns, resulting in a prioritized improvement plan.
Skill evaluation supports quality assurance by applying multi-dimensional assessment across the D1-D8 dimensions, revealing specific design flaws and providing structured guidance for skill governance.
Improve skill design quality by applying the prioritized improvement plan generated from the evaluation report, which transforms identified anti-patterns into actionable guidance for production-ready skill creation.
Validate frontmatter and structural formatting during a skill evaluation to ensure the SKILL.md file aligns with official specifications and meets the required quality assessment standards.
Ikram-AlamEvaluate agent outputs against rubrics and ground truth with multi-dimensional scoring.
pianzhuEvaluates submitted ML models and returns structured JSON with predictions, confidence scores, and metadata.
nakano1122Validate Claude Code skills against best practices and quality standards.
tomas-uGuide security architecture design with STRIDE threat modeling and compliance assessments.
