Skillpack Maintenance

概要

This skill provides a rigorous framework for refreshing and improving existing Claude Code skill packs, ensuring they remain effective and accurate through behavioral validation. It moves beyond simple syntax checks by using subagent-driven gauntlet testing to identify where agents might rationalize or fail in real-world scenarios. By integrating domain analysis, structural audits, and coherence checks, the skill helps curators identify gaps, remove redundancies, and align packs with their intended domain boundaries while enforcing the 'Iron Law' that no skill changes occur without proven behavioral success.

主な機能

• Subagent-driven behavioral gauntlet testing for quality validation
• Structural review of pack organization and router accuracy
• Coherence checks for terminology and cross-skill consistency
• 5 GitHub stars
• Systematic domain coverage analysis and gap identification
• Impact-based version management and standardized commit workflows

ユースケース

• Refreshing outdated skill packs with modern domain knowledge
• Validating the effectiveness of existing guidance through adversarial stress-testing
• Identifying and filling gaps in skill pack coverage through systematic audits