writing-skills

Writing reusable agent skills with TDD involves following a structured, test-driven methodology that standardizes documentation, enforces frontmatter validation, and scaffolds baseline test scenarios to ensure consistency.

Progressive disclosure in skill documentation is an approach that loads extended references only when needed, keeping the core SKILL.md content concise and search-friendly while reducing speculative edits during skill creation.

Validating agent skill frontmatter before deployment requires running a pre-deployment testing workflow that enforces required name and description fields, checks baseline scenarios, and identifies anti-patterns.

Creating test scaffolding for agent skills involves using TDD-driven templates to establish baseline scenarios and anti-patterns, reducing speculative edits and ensuring the skill behaves predictably before deployment.

Test-driven development applies to editing existing agent skills by using pre-deployment verification workflows and baseline tests to validate changes, ensuring updates maintain consistency without breaking established functionality.

You should not use a TDD methodology for skill writing when your documentation requires highly dynamic, non-deterministic outputs that cannot be validated against baseline test scenarios or standard frontmatter rules.