tdd

The red-green-refactor loop is a testing workflow where you write a failing test, implement the minimal code to pass it, and then refactor. This cycle validates software behavior through public interfaces with one test per behavior.

Write integration tests that interact exclusively with public APIs rather than internal implementation. This approach ensures you validate observable behavior end-to-end, keeping tests resilient during refactors and feature development.

Use end-to-end testing when you need to validate a feature works through its public surface across the entire stack. This Skill applies testing workflows to unit, integration, and acceptance testing for features, bug fixes, and refactors.

Test-driven development works for refactoring by using reproducible end-to-end tests to verify observable behavior through public interfaces. This ensures the software behavior remains valid before and after the code changes.

You can apply the red-green-refactor workflow to acceptance testing by writing a single end-to-end test for the first behavior, then implementing the minimal code to pass it. This vertical slice workflow repeats for each behavior.

Tests break during refactors when they validate internal implementation instead of observable behavior. Disciplined testing through public interfaces prevents this by ensuring reproducible end-to-end tests remain stable across code changes.