Writing Claude Code Skills
Provides comprehensive guidance and best practices for creating, managing, and debugging auto-activating skills within the Claude Code environment.
This skill serves as a meta-guide for developers building specialized capabilities into Claude Code, specifically within the HASH knowledge graph repository. It streamlines the creation of new domain-specific skills by enforcing Anthropic's official best practices, such as the 500-line rule and the progressive disclosure pattern. By leveraging the UserPromptSubmit hook system, it enables automated skill activation through sophisticated keyword and intent matching, ensuring that relevant architectural patterns and documentation are surfaced to the AI precisely when they are most needed.
Key Features
01Step-by-step guidance for creating skills with YAML frontmatter and proper folder structures.
02Enforcement of Anthropic best practices including the 500-line rule and progressive disclosure.
031,387 GitHub stars
04Built-in validation and debugging tools for testing skill triggers and hook mechanics.
05Configuration management for domain-specific guidance and advisory enforcement levels.
06Automated skill activation management via keyword and regex-based intent matching.
Use Cases
01Establishing standardized architectural patterns and best practices for new repository contributors.
02Configuring proactive AI suggestions based on specific developer intent or keywords.
03Debugging complex skill activation logic and refining trigger patterns to reduce false positives.