How do I trigger a custom skill in Claude Code?

Skills are activated via the UserPromptSubmit hook when a user's prompt matches specific keywords or regex intent patterns defined in the skill-rules.json file.

What is the 500-line rule for Claude Skills?

It is a core best practice to keep SKILL.md files under 500 lines. This ensures the model remains focused and efficient, while deeper details are offloaded to external reference files.

How can I test if my skill triggers correctly?

You can use the SKILL_DEBUG=true environment variable or specialized yarn workspace commands to inspect the matching logic and see exactly why a skill was or was not suggested.

What is progressive disclosure in skill writing?

Progressive disclosure is a pattern where the main SKILL.md provides high-level guidance, while more complex or lengthy documentation is stored in separate reference files to save context space.

Claude Skill Authoring

Name: Claude Skill Authoring
Author: Microck

byMicrock

•

개발자 도구

Guides the creation and management of custom Claude Code skills using Anthropic best practices and standardized architectural patterns.

This skill provides a comprehensive framework for developing custom Claude Code capabilities, ensuring they adhere to the '500-line rule' and progressive disclosure patterns. It guides users through configuring trigger patterns (keywords and regex intents), managing YAML frontmatter, and implementing hooks like UserPromptSubmit for proactive skill activation. Whether you are debugging skill matching logic or setting up new domain-specific guidance, this tool ensures your AI extensions are performant, valid, and highly effective for your specific development environment.

주요 기능

01Automated guidance for creating SKILL.md files with proper YAML frontmatter

02Standardized debugging and validation workflows for skill activation logic

03Support for hook system mechanics and UserPromptSubmit logic documentation

04Configuration management for skill-rules.json trigger patterns and intent matching

05Implementation of the Anthropic 500-line rule for optimal context management

0681 GitHub stars

사용 사례

01Debugging and refining skill triggers to prevent false positives in AI activation

02Restructuring legacy documentation into high-performance progressive disclosure patterns

03Building custom domain-specific skills for proprietary codebases or specialized frameworks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add microck/ordinary-claude-skills writing-skills

For use in Claude.ai and ChatGPT

Download Skill