What problem does it solve?
This Skill provides a systematic, TDD-driven approach to creating high-quality, bulletproof AI skills. It ensures that new skills are effective, discoverable, and resistant to AI rationalization, preventing the deployment of untested or ineffective process documentation.
Core Features & Use Cases
- TDD for Documentation: Applies the RED-GREEN-REFACTOR cycle to skill creation, starting with baseline testing to identify AI failures.
- Skill Structure & Content: Guides on organizing
SKILL.md with rich when_to_use, clear overviews, and concise examples.
- Claude Search Optimization (CSO): Provides critical techniques for making skills easily discoverable by future Claude instances through keyword coverage, descriptive naming, and token efficiency.
- Bulletproofing Against Rationalization: Teaches how to explicitly counter AI rationalizations with specific negations, rationalization tables, and red flags.
- Use Case: When you identify a new best practice or tool worth documenting for your AI, use this skill to systematically create, test, and refine the skill until it reliably guides your AI's behavior under pressure.
Quick Start
To create a new skill, first identify a technique or pattern. Then, before writing the skill, run a baseline test (RED phase) with a subagent to see how it naturally behaves without the skill. Document its failures and rationalizations. Only then, write your skill (GREEN phase) to address those specific failures.
