What problem does it solve? Manually running test-driven development cycles is repetitive and error-prone: writing failing tests, iterating on production code, running typecheck and lint gates, and structuring commits and PRs. This Skill automates the entire TDD workflow from failing test suite to merged-ready pull request. ## Core Features & Use Cases - Autonomous Red-Green-Refactor Loop: Writes a failing test suite from acceptance criteria, commits it, then iterates up to 20 times — running tests, forming hypotheses, and editing production code until all tests pass. - Quality Gates and Guardrails: Runs typecheck, lint, and a full regression suite with up to 5 fix iterations, enforces forbidden edits (no deleting tests, no mocking the unit under test), and stops cleanly on hard caps. - Structured Commits and PRs: Produces exactly two conventional commits (test: then feat:) and opens a PR embedding the full TDD iteration log for reviewer transparency. - Use Case: Ask the agent to TDD a new accessible Tabs component with a numbered list of acceptance criteria; it writes the failing Vitest suite, iterates on the implementation, passes gates, and opens a PR with the iteration log. ## Quick Start Ask the agent to TDD a new feature by providing a feature description with a numbered list of acceptance criteria on a clean feature branch.
