writing-skills

To test AI skills with TDD, adapt the RED-GREEN-REFACTOR cycle to documentation by writing failing pressure scenarios first, then building the skill to pass those tests, and refactoring to enforce rules against rationalization.

A rationalization table is a countermeasure used in AI skill development to explicitly address common excuses agents use to bypass rules, ensuring skills remain robust and effective under realistic pressure scenarios.

Pressure test an AI skill by applying realistic scenarios with multiple pressures like time, sunk cost, and authority to uncover weaknesses, ensuring the skill enforces explicit rules and remains effective under stress.

TDD for AI skills applies to all skill types, including discipline-enforcing, technique, pattern, and reference skills, ensuring each variant is robust, discoverable, and effective against rule rationalization.

AI agents bypass skill rules under pressure due to rationalization, using common excuses to skip steps when facing time, sunk cost, or authority pressures, which is why explicit rule enforcement and rationalization countermeasures are needed.

The best way to create robust AI skills is using a test-driven framework that employs pressure scenarios and rationalization tables, ensuring the skills are effective, discoverable, and enforce explicit rules under stress.