Skill Granularity Convention

概要

This skill provides a comprehensive framework for designing Claude Code skills with optimal granularity, focusing on the core principle of 'one skill, one purpose.' It offers actionable guidance on when to split complex tasks into smaller skills and when to merge undersized skills to maintain practical utility. By establishing clear metrics for responsibility scope, workflow phases, and documentation size, it helps developers create highly reusable, low-overhead skills that improve AI performance and system maintainability.

主な機能

• Establishes merging guidelines for skills that are too narrow for practical use.
• 0 GitHub stars
• Offers a comprehensive checklist for auditing skill granularity and design quality.
• Defines the 'One Skill, One Purpose' principle for modular design.
• Provides specific criteria for splitting skills based on complexity and scope.
• Includes quantitative judgment metrics for description length and workflow phases.

ユースケース

• Auditing a skill library to ensure consistent granularity and performance optimization.
• Refactoring monolithic skills into smaller, more reusable components.
• Establishing design standards when creating new Claude Code skills.