coder

What problem does it solve? Coordinating feature implementation across branches, tests, and task tracking is error-prone when done ad hoc. This Skill gives an AI agent a disciplined, repeatable workflow for claiming coding tasks, writing code test-first in an isolated git worktree, and committing changes without ever touching the default branch. ## Core Features & Use Cases - Task-Driven Implementation: Claims pending tasks from markdown task files or a GitHub Projects board, then works through a written plan with small, independently testable steps. - TDD in Git Worktrees: Creates a dedicated worktree branched off the feature branch, writes unit tests first (with all external I/O mocked), runs the full test suite, and commits with conventional or co-authored commit styles. - Safe Handoff: Marks tasks done with a summary of changes and stops — never pushes, merges, or cleans up, leaving review and integration to the Coordinator. - Use Case: A coordinator agent dispatches a task to add a MIDI clock divider to a sequencer engine. The coder agent claims the task, creates a worktree, writes failing unit tests, implements the divider, runs the full suite, commits, and reports back for code review. ## Quick Start Ask the agent to pick up the next pending coder task from the workflow directory and implement it using test-driven development in a new git worktree.