writing-skills

AI agent documentation uses a test-driven development workflow by applying red-green-refactor cycles to skill creation, ensuring clear and verifiable guidance through test scaffolds and structured SKILL.md layouts.

Progressive disclosure in skill documentation loads references, scripts, and background guides only when needed to minimize context load. Use it when organizing self-contained SKILL.md files with optional asset directories.

To validate frontmatter requirements before deployment, check that your SKILL.md enforces YAML frontmatter with a name and description, constraining length per specification to ensure readiness across AI agents.

You can use this test-driven framework to edit existing skills, create new ones, and validate readiness before deployment. It applies red-green-refactor cycles to both skill creation and maintenance.

The best way to organize skill documentation for cross-skill reuse is linking to existing skills to reduce duplication, using a self-contained SKILL.md with optional directories for scripts, references, and assets.

You do not need external dependencies to maintain skill documentation with this framework. It operates independently with no required dependencies, enforcing frontmatter discipline and progressive disclosure natively.