What problem does it solve? Code written without tests drifts into regressions, and AI coding agents routinely skip test-first discipline unless it is enforced. This Skill installs a rigorous TDD methodology—RED-GREEN-REFACTOR with mode selection, mutation gates, and commit evidence—so every feature, bugfix, and refactor follows a verifiable test-first process. ## Core Features & Use Cases - Three TDD modes: strict (Iron Law: no production code without a failing test), souple (RED-GREEN-REFACTOR without mutation gates), and exploratory (throwaway spikes with declared deletion dates), auto-detected from file paths or explicit markers. - Mutation testing gate: when a mutation runner like Stryker is detected in the project config, strict mode requires running it and killing surviving mutants before completion. - Companion hook enforcement: a PreToolUse hook blocks edits to production code unless a sibling test file or a declared E2E coverage file exists. - Use Case: While implementing a retry utility in a TypeScript service, the Skill forces you to write a failing backoff test first, watch it fail, write minimal code to pass, run the mutation suite, and produce a commit history showing the RED-GREEN-REFACTOR progression. ## Quick Start Ask the agent to implement a new feature or fix a bug using test-driven development in strict mode, and it will write the failing test before any production code.
