Applies Test-Driven Development principles to author and refine high-quality documentation and specialized capabilities for Claude Code.
This skill provides a rigorous framework for creating and maintaining Claude Code skills by treating process documentation like production code. By following a specialized Red-Green-Refactor cycle, users first verify that an AI agent fails a specific task (baseline), author the necessary guidance, and then confirm compliance. It emphasizes Claude Search Optimization (CSO) to ensure skills are discoverable, token-efficient, and correctly triggered by specific symptoms or conditions rather than generic summaries.
Key Features
01Optimizes skill descriptions for Claude Search Optimization (CSO)
020 GitHub stars
03Includes guidance for progressive disclosure and skill cross-referencing
04Provides standardized templates for SKILL.md and directory structures
05Implements a Red-Green-Refactor cycle for process documentation
06Enforces token efficiency to preserve AI context window
Use Cases
01Optimizing large documentation sets for better AI discovery and recall
02Fixing AI agent non-compliance by closing documentation loopholes
03Creating new reusable domain-specific skills for development teams