概要
This skill provides a rigorous framework for developing AI-native documentation by treating knowledge as code. By implementing a Red-Green-Refactor cycle for skill creation, it ensures that every instruction added to an agent's library is empirically verified to work and addresses specific failure modes. It includes advanced guidelines for Claude Search Optimization (CSO), token efficiency, and standardized directory structures, allowing developers to build a high-performance library of reusable techniques that Claude can discover and execute reliably.