writing-skills

Test-driven skill documentation creates and validates AI skills using a RED-GREEN-REFACTOR workflow, ensuring skill docs are clear, verifiable, and resistant to rationalization under pressure before deployment.

Skill documentation requires YAML frontmatter with a name and description, followed by a concise three-section structure that integrates tested workflows, pressure scenarios, and rationalization tracking.

The RED-GREEN-REFACTOR workflow applies to skill development by writing tests first, creating the skill to pass tests, then refining the documentation to ensure it remains bulletproof against rationalization.

You need rationalization tracking in process documentation when verifying skills under pressure scenarios, ensuring the AI resists rationalization and maintains disciplined behavior during deployment.

The three-section structure is strictly limited to concise documentation with YAML frontmatter, meaning it may not accommodate complex skill designs requiring extensive multi-section explanations or nested sub-workflows.

Yes, this test-driven documentation approach works for skills without dependencies, focusing on frontmatter validation, rationalization tracking, and cross-referencing to ensure standalone skills remain bulletproof.