testing-skills-with-subagents

Test-Driven Development for skills applies RED-GREEN-REFACTOR cycles to documentation, identifying failures before deployment. This skill validates that agents follow rules consistently, even under pressure, by running baseline scenarios with subagents and capturing rationalizations to close behavioral loopholes.

Pressure scenarios simulate realistic constraints—tight deadlines, competing priorities—that expose when agents skip steps or find excuses to bypass skill rules. Testing under pressure reveals weaknesses and ensures your skill documentation is robust enough to prevent agents from rationalizing noncompliance.

Yes. This skill processes documentation directly through TDD principles, running baseline tests and pressure scenarios with subagents to verify your skill meets structured requirements. It yields actionable outcomes showing where documentation fails to enforce consistent AI behavior.

This skill captures rationalizations—common excuses agents use to bypass rules—during pressure testing. By comparing baseline scenarios against pressure-test results, you surface inconsistencies in agent behavior and explicitly counter loopholes in your skill documentation.

Prepare baseline test scenarios, pressure scenarios that introduce constraints, and documented skill rules. The skill applies these inputs through subagent validation cycles, identifying failures and rationalizations to ensure your skill is bulletproof before deployment.

TDD for skills systematically closes gaps in documentation by writing tests that expose failures first, then refactoring documentation to prevent them. This structured approach captures agent rationalizations under pressure and guarantees consistent behavior across scenarios, reducing deployment risk.