What are the different severity levels in the validation report?

The validator categorizes issues as Critical (must fix), Error (should fix), Warning (consider fixing), or Suggestion (optional improvement).

How do I fix a structure error for a long SKILL.md file?

The validator recommends splitting large files into modular components like WORKFLOW.md or EXAMPLES.md to improve readability and AI performance.

What is the main purpose of Skill Validator?

It automates the review process for Claude Code skills, ensuring they meet specific standards for structure, security, and readability before deployment.

Can I ignore specific validation rules?

Yes, you can use the --ignore flag in the command line or add inline comments like # skill-validator: ignore to skip specific checks that may be false positives.

Does it check for security vulnerabilities?

Yes, it specifically scans for security risks such as the use of eval/exec functions and undocumented constants that could pose a safety threat.

Skill Validator

Name: Skill Validator
Author: joaquimscosta

byjoaquimscosta

•

Developer Tools

Ensures Claude Code skills adhere to Anthropic's best practices for structure, content quality, and security standards.

Skill Validator is a specialized utility designed to audit and refine Claude Code skills before they are deployed or shared. It automatically checks for compliance across multiple domains, including YAML frontmatter integrity, markdown structure, writing style consistency, and security vulnerabilities like undocumented constants or dangerous execution patterns. By providing clear severity levels and actionable fixes, it helps developers maintain high-quality, secure, and user-friendly AI capabilities while ensuring that skills remain performant and easy for the LLM to process.

Key Features

01Security auditing to detect dangerous eval/exec calls and undocumented constants

02Automated validation of YAML frontmatter and required metadata fields

03Flexible reporting with severity levels ranging from critical errors to optional suggestions

04Content linting to ensure imperative writing style and terminology consistency

053 GitHub stars

06Structural analysis of SKILL.md to enforce progressive disclosure and line limits

Use Cases

01Performing security reviews to ensure skills do not contain hidden execution risks

02Refactoring large SKILL.md files into modular components based on structural recommendations

03Auditing a new Claude Code skill for compliance before publishing to a public repository

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add joaquimscosta/arkhe-claude-plugins skill-validator

For use in Claude.ai and ChatGPT

Download Skill