Audit Agent Skills against a 10-rule quality rubric and generate compliant scaffolds.
npx skills add https://github.com/Liu-PenPen/skill-review --skill skill-reviewer-liu-penpen
Please help me install this Agent Skill. Skill: skill-reviewer Source: https://github.com/Liu-PenPen/skill-review/tree/main Command: npx skills add https://github.com/Liu-PenPen/skill-review --skill skill-reviewer-liu-penpen
Agent Skills often suffer from inconsistent quality, missing critical best practices like progressive disclosure, confirmation nodes for irreversible actions, and proper trigger keyword density in descriptions. These gaps lead to context bloat, missed auto-triggers, unexpected behavior during execution, and poor user experience for both developers and end users of the Skill.
Ask the AI to use the skill-reviewer skill to audit the Agent Skill located at .cursor/skills/my-design-skill and receive a graded quality report with actionable fix recommendations.
High-intent search queries and answers about installing and using this skill.
You can audit Cursor Agent Skills by checking them against a 10-rule quality rubric to identify compliance gaps. This process outputs P0-P3 graded findings with concrete evidence and fix suggestions for issues causing context bloat or missed auto-triggers.
Agent Skills miss auto-triggers and cause context bloat due to missing critical best practices like proper trigger keyword density in descriptions and progressive disclosure. Auditing the Skill against a quality rubric identifies these exact structural gaps.
To generate a pre-validated Agent Skill scaffold, use a rubric-based tool to create a new directory structure that already includes required checklists, frontmatter fields, and workflow patterns. This ensures the scaffold complies with quality rules out of the box.
Yes, you can run an Agent Skill code review in GitHub Actions or pre-commit hooks using a zero-dependency lint script. This automates Skill quality checks as part of your standard CI/CD workflows without requiring external packages.
A rubric-based approach for Agent Skill quality audits is limited to structural and metadata checks like progressive disclosure, confirmation nodes, and trigger keyword density. It does not evaluate the runtime execution logic or semantic accuracy of the Skill's internal code.
No, you do not need to install external dependencies to lint Agent Skills for missing confirmation nodes. The lint script operates with zero external dependencies, enforcing pre-delivery self-checks and confirming irreversible actions directly from the command line.
albertdobmeyerEvaluate SKILL.md files for structural validity, quality scoring, and publish readiness.
bordenetAutomate progressive code review before commits and pushes.
lbussellGuide SKILL.md authoring with YAML frontmatter and Markdown structure.
bernajaberReview git diffs for UX, visual, accessibility, security, and constitution violations.
