What is the minimum passing score for a skill?

A standard audit requires a minimum weighted average of 3.0 (Acceptable), with a target of 4.0 (Good) for production-ready status.

Does it fix issues automatically?

It includes a --fix flag for simple formatting and structural issues, though content-related improvements require manual intervention based on the audit report.

What are the four dimensions used for auditing?

The skill evaluates Clarity, Completeness, Accuracy, and Usefulness, with Accuracy carrying the highest weight (30%) to ensure safe and correct code.

Can this skill be used in a CI/CD pipeline?

Yes, it includes implementation patterns for GitHub Actions to automate quality gating during pull requests for skill repositories.

Quality Audit

Name: Quality Audit
Author: NickCrew

byNickCrew

•

보안 및 테스팅

Validates and audits Claude Code skills using structured rubrics to ensure high standards for clarity, accuracy, and usefulness.

소개

The Quality Audit skill provides a systematic framework for evaluating the performance and reliability of specialized Claude capabilities. By scoring skills across four key dimensions—Clarity, Completeness, Accuracy, and Usefulness—it ensures that any added functionality meets production-grade standards. Whether you are reviewing a new skill before registry submission or auditing existing tools for consistency, this skill generates actionable reports and scoring rubrics to maintain a high-quality ecosystem for AI-driven development.

주요 기능

7 GitHub stars
Interactive scoring mode for human-in-the-loop reviews
Four-dimension scoring framework (Clarity, Completeness, Accuracy, Usefulness)
Automated structure and YAML frontmatter validation
Production-ready report generation with actionable recommendations
CI/CD integration for automated quality gating

사용 사례

Pre-submission review for new Claude Cortex skills
Periodic auditing of existing skill libraries for accuracy
Standardizing documentation and code quality across team projects

소개

주요 기능

7 GitHub stars
Interactive scoring mode for human-in-the-loop reviews
Four-dimension scoring framework (Clarity, Completeness, Accuracy, Usefulness)
Automated structure and YAML frontmatter validation
Production-ready report generation with actionable recommendations
CI/CD integration for automated quality gating

사용 사례

Pre-submission review for new Claude Cortex skills
Periodic auditing of existing skill libraries for accuracy
Standardizing documentation and code quality across team projects