What is the RED phase in skill testing?

The RED phase involves running a scenario without the skill active to watch the agent fail and document its exact rationalizations for failing.

How does the REFACTOR phase work for documentation?

In the REFACTOR phase, you identify new loopholes or excuses the agent found during testing and add explicit negations and counter-rules to the skill.

When should I not use this skill?

You don't need to test pure reference skills like API documentation or syntax guides where there are no specific rules for an agent to violate.

Why are pressure scenarios important?

Pressure scenarios simulate real-world constraints like deadlines and sunk costs, which reveal if an agent will 'rationalize away' your skill's rules.

Skill Testing with Subagents

Name: Skill Testing with Subagents
Author: AlabamaMike

byAlabamaMike

보안 및 테스팅

Applies Test-Driven Development (TDD) principles to verify and bulletproof Claude Code skills against agent rationalization.

소개

This skill implements a rigorous RED-GREEN-REFACTOR cycle specifically for process documentation and agent instructions. By simulating high-pressure scenarios—such as tight deadlines, sunk costs, and authority overrides—developers can identify exactly where an AI agent might naturally fail. The skill guides the iterative process of capturing these failures, drafting minimal rules to address them, and building rationalization tables to ensure the final skill is followed consistently, even in demanding real-world coding environments.

주요 기능

High-pressure scenario simulation templates
Meta-testing frameworks for iterative skill refinement
0 GitHub stars
Rationalization capture and verbatim negation mapping
RED-GREEN-REFACTOR workflow for process documentation
Multi-pressure testing (Time, Sunk Cost, Authority, Social)

사용 사례

Verifying compliance-heavy coding standards under tight deadlines
Iteratively refining team-specific workflow documentation for consistency
Hardening architectural guidelines against 'pragmatic' bypasses

소개

주요 기능

High-pressure scenario simulation templates
Meta-testing frameworks for iterative skill refinement
0 GitHub stars
Rationalization capture and verbatim negation mapping
RED-GREEN-REFACTOR workflow for process documentation
Multi-pressure testing (Time, Sunk Cost, Authority, Social)

사용 사례

Verifying compliance-heavy coding standards under tight deadlines
Iteratively refining team-specific workflow documentation for consistency
Hardening architectural guidelines against 'pragmatic' bypasses