How do I debug why a skill isn't activating?

You can use the environment variable SKILL_DEBUG=true to see the matching logic or run the specific validation command 'yarn lint:skill' to check your configuration.

What is the 500-line rule for Claude skills?

It is an Anthropic best practice to keep SKILL.md files under 500 lines to ensure the AI processes the context efficiently. Detailed information should be moved to reference files.

What is progressive disclosure in this context?

It is a structural pattern where the main SKILL.md provides a high-level summary and links to specialized reference files for deeper technical details.

What is the UserPromptSubmit hook?

This is a proactive hook that runs before Claude sees the user's prompt. It checks for triggers and injects relevant skill reminders into the context.

Can I use regular expressions for skill triggers?

Yes, you can define 'intentPatterns' in your skill-rules.json using regex to match specific user intents rather than just static keywords.

Claude Skill Writing & Management

Name: Claude Skill Writing & Management
Author: Microck

byMicrock

•

開発者ツール

Standardizes the creation and management of specialized Claude Code skills following Anthropic's official best practices.

概要

This skill provides a comprehensive framework for developing, debugging, and maintaining custom skills within the Claude Code environment. It guides users through the implementation of the 500-line rule, progressive disclosure patterns, and complex trigger configurations like keyword matching and regex-based intent patterns. By automating the suggestion of relevant domain knowledge through the UserPromptSubmit hook, it ensures that Claude operates with the most pertinent context while maintaining project performance and structural integrity.

主な機能

Automated skill activation via keyword and regex intent triggers
Proactive context injection via UserPromptSubmit hook mechanics
Implementation of the Anthropic 500-line rule for optimal performance
Integrated debugging tools using SKILL_DEBUG and validation flags
Comprehensive management of skill-rules.json and YAML frontmatter
81 GitHub stars

ユースケース

Debugging and refining skill activation patterns to reduce false positives
Architecting large documentation sets using the progressive disclosure pattern
Developing custom domain-specific skills for private libraries or workflows

概要

主な機能

Automated skill activation via keyword and regex intent triggers
Proactive context injection via UserPromptSubmit hook mechanics
Implementation of the Anthropic 500-line rule for optimal performance
Integrated debugging tools using SKILL_DEBUG and validation flags
Comprehensive management of skill-rules.json and YAML frontmatter
81 GitHub stars

ユースケース

Debugging and refining skill activation patterns to reduce false positives
Architecting large documentation sets using the progressive disclosure pattern
Developing custom domain-specific skills for private libraries or workflows