Testing Skills With Subagents

关于

This skill introduces a rigorous Test-Driven Development (TDD) framework for creating and validating AI instructions. It guides the creation of 'bulletproof' skills by forcing agents through high-pressure scenarios—incorporating time constraints, sunk costs, and social pressures—to identify where they might rationalize breaking rules. By documenting these failures in the RED phase and iteratively plugging holes in the REFACTOR phase, users can ensure their domain-specific guidelines are followed consistently, even when the agent is tempted to prioritize speed over quality.

主要功能

• Multi-pressure testing scenarios including time, economic, and sunk cost constraints
• 0 GitHub stars
• Red-Green-Refactor cycle for AI documentation
• Verbatim rationalization capture and counter-strategy mapping
• Meta-testing methodology to verify instruction clarity and organization
• Foundational principles for enforcing the spirit of rules over simple compliance

使用场景

• Developing discipline-enforcing coding standards and TDD skills
• Validating skill compliance before deployment in production environments
• Hardening process documentation against AI rationalizations and shortcuts