testing-skills-with-subagents

Test-driven development for skills applies the RED-GREEN-REFACTOR cycle: first verify the skill fails without it, then iteratively improve it until it withstands real-world pressure scenarios combining time constraints, sunk costs, and authority. This methodology ensures skills remain effective when rationalization tempts violations.

Pressure scenario testing creates realistic conditions—time pressure, sunk cost, authority, exhaustion—to verify skills resist rationalization in high-stakes situations. It satisfies compliance requirements by capturing verbatim rationalization attempts and documenting how skills perform when discipline is hardest to maintain.

Yes. This skill applies test-driven development specifically to compliance scenarios, running baseline tests without the skill, then with it across RED, GREEN, and REFACTOR phases. It generates pressure scenarios, captures rationalization attempts, and produces documentation-driven verification that the skill enforces discipline under real constraints.

Design pressure scenarios that combine time limits, sunk-cost reasoning, and authority challenges, then run the skill through each phase of the RED-GREEN-REFACTOR cycle. Document how the skill responds to rationalization attempts, ensuring it maintains compliance even when pressured to skip or rationalize violations.

Generic testing checks if a skill works in ideal conditions; pressure scenario testing deliberately introduces time constraints, sunk costs, authority, and exhaustion to reveal whether the skill resists rationalization. This approach catches failures that only surface when discipline is under maximum strain.