What problem does it solve? Writing reusable skill documentation without verification leads to instructions that agents misinterpret or ignore under pressure. This Skill enforces a TDD discipline for documentation: test with subagents first, watch baseline failures, then write and iterate until the skill is bulletproof. ## Core Features & Use Cases - RED-GREEN-REFACTOR for docs: Run pressure scenarios with subagents before writing, capture verbatim rationalizations, then write the minimal skill that counters them. - Claude Search Optimization (CSO): Guidance on rich when_to_use fields, keyword coverage, naming conventions, and token budgets so future agents can actually discover the skill. - Skill type testing strategies: Distinct test approaches for discipline-enforcing, technique, pattern, and reference skills, plus rationalization tables and red-flag lists to close loopholes. - Use Case: After discovering a useful debugging technique, run baseline subagent scenarios without documentation, write a SKILL.md addressing the observed failures, and re-test until agents comply under time, sunk-cost, and authority pressure. ## Quick Start Ask the agent to help you create a new skill by first running baseline subagent test scenarios, then drafting the SKILL.md and iterating until the tests pass.
