What problem does it solve? Writing documentation-style skills for AI agents often produces instructions that agents ignore or rationalize away under pressure. This Skill applies the RED-GREEN-REFACTOR cycle of test-driven development to skill authoring, so every rule is verified against real baseline agent failures before deployment. ## Core Features & Use Cases - TDD-based skill authoring: Run baseline pressure scenarios with subagents, document verbatim rationalizations, write the minimal skill that addresses them, then close loopholes iteratively. - Discovery optimization (SDO): Guidance on writing trigger-focused descriptions, keyword coverage, naming conventions, and token-efficient structure so agents actually find and load the skill. - Bulletproofing toolkit: Rationalization tables, red-flag lists, persuasion principles, and form-matching rules to make discipline-enforcing skills resist workarounds. - Use Case: You want to create a new skill enforcing a code-review rule. You first run pressure scenarios (time pressure, sunk cost, authority) without the skill, record how agents violate the rule, then write and iteratively harden the skill until agents comply under maximum pressure. ## Quick Start Ask the agent to create a new skill following the writing-skills TDD process, starting with baseline pressure scenarios before writing any SKILL.md content.
