Write and validate agent documentation skills using a RED-GREEN-REFACTOR process.
npx skills add https://github.com/XCmiaow/cc-deploy --skill writing-skills-xcmiaow
Please help me install this Agent Skill. Skill: writing-skills Source: https://github.com/XCmiaow/cc-deploy/tree/main/skills/writing-skills Command: npx skills add https://github.com/XCmiaow/cc-deploy --skill writing-skills-xcmiaow
This Skill prevents agent instructions from becoming untestable, ambiguous guidance by enforcing a test-driven loop for documentation-based skills.
Use the writing-skills Skill when you need to create or edit a documentation skill and you want to verify the agent complies by running a baseline “without-skill” pressure scenario first.
High-intent search queries and answers about installing and using this skill.
Test agent documentation compliance by running baseline pressure scenarios without the skill, then apply a RED-GREEN-REFACTOR loop to verify instructions change agent behavior under time, sunk-cost, authority, or exhaustion pressure.
Test-driven development for skill authoring adapts the RED-GREEN-REFACTOR process to documentation, ensuring skills enforce discipline rather than providing untestable, ambiguous guidance that agents rationalize away.
Close rationalization loopholes in agent instructions by iteratively refactoring documentation to add explicit negations, foundational principles, rationalization tables, and red-flag stop conditions that address observed violations.
Verify a skill changes agent behavior by defining discipline-verification scenarios with subagents that capture real failure modes, requiring a baseline test without the skill before making minimal documentation changes.
Yes, test-driven skill engineering applies when creating, editing, or verifying skills, using the RED-GREEN-REFACTOR loop to iteratively close new loopholes while maintaining existing compliance.
obraApply RED-GREEN-REFACTOR cycles to create and test AI skills.
charly3pinsTest skills with subagents under pressure scenarios and log rationalizations.
AlabamaMikeTest skills with subagents under combined pressure scenarios across RED-GREEN-REFACTOR cycles.
buchananwillSummarize UE C++ lint hook rules for auto-caught versus manual review items.
