validate-skills

The `validate-skills` AI skill is designed to ensure that all skills within a repository adhere to established specifications and best practices. It primarily addresses the challenge of maintaining consistency and quality across multiple skill modules by automatically validating each skill against the agentskills.io specification and Claude Code best practices. By performing structured checks, this skill helps developers catch errors, inconsistencies, and potential issues early in the development workflow, reducing the risk of misconfigured or poorly documented skills.

The skill provides a comprehensive validation process for each skill directory. It checks the `name` field format, ensures it matches the directory name, verifies the description length, and validates optional fields such as `license`, `metadata`, and `compatibility` if present. Additionally, it enforces best practices from Claude Code, including proper description formatting, maintaining concise and non-redundant skill documentation, limiting body length, ensuring links use markdown syntax, and preventing nested reference chains. The validation process is structured to provide clear, actionable feedback for each skill, indicating which checks passed and which failed, allowing developers to quickly identify and correct issues.

This skill is particularly useful for developers, technical writers, and quality assurance teams involved in creating, maintaining, or auditing AI skills. Use cases include routine repository audits, pre-release quality checks, and onboarding new contributors to ensure they follow consistent skill-authoring standards. By automating these validation tasks, teams can maintain higher code quality, improve documentation consistency, and streamline the review process, ultimately leading to a more reliable and professional skill ecosystem.