What problem does it solve? Turning written specifications into working code often loses traceability: developers skip tests, mix unrelated work into one branch, or commit without review checkpoints. This Skill executes already-specified work (user story tasks, maintenance items, test cases, features) in code with enforced TDD, per-unit confirmation, and a persistent progress log. ## Core Features & Use Cases - Four implementation types: technical tasks (TK-XXX) under a user story, maintenance work items (WI-XXX) such as bugs and refactors, test case automation (TC-XXX), and feature test automation (FT-XXX), each with its own flow in references/. - Mandatory TDD cycle: Red-Green-Refactor per behavior unit, with tests scoped strictly to affected files and the full test suite deferred to the quality-check phase. - Progress tracking and guardrails: maintains progress.md as the only writable log, mirrors the plan in the agent's to-do list, supports git worktrees and parallel subagent execution, and escalates after a configurable number of failed attempts. - Use Case: A developer asks to implement the Ready tasks of US-042. The Skill creates the feature branch, presents the task queue, implements one task at a time with tests first, pauses for confirmation between tasks, and hands off to integration or PR creation when done. ## Quick Start Ask the agent to implement the Ready tasks of a specific user story, for example: implement all Ready tasks of US-042 one by one with tests.