What problem does it solve? Writing tests that survive refactoring is hard: tests coupled to implementation details break on every code change, tautological tests can never fail, and horizontal slicing locks in test structure before the implementation is understood. This Skill provides a reference guide for running the red-green TDD loop correctly so the resulting tests verify behavior through public interfaces. ## Core Features & Use Cases - Red-Green Loop Rules: Enforces writing a failing test first, then only the minimal code to pass, one vertical slice at a time. - Seam-Based Test Placement: Requires agreeing on public-interface seams with the user before writing any test, focusing effort on critical paths. - Anti-Pattern Detection: Identifies implementation-coupled tests, tautological assertions, and horizontal slicing, with good/bad examples in tests.md and mocking guidance in mocking.md. - Use Case: When fixing a bug in a checkout flow, write one failing test at the public checkout interface, implement the minimal fix, and repeat—mocking only the external payment API at the system boundary. ## Quick Start Ask the AI to build the next feature using test-driven development with the red-green cycle, agreeing on the test seams first.
