implement-spec

# Implement spec Implement a specification end to end on one integration branch. The tickets are not a list of steps — they are a **task graph**, so there is always a **frontier** of tickets ready to be grabbed. ## Roles - **Exploration subagent.** Before implementation, explore the codebase and save notes outside the repository so implementer subagents focus on implementation rather than discovery. - **Implementer subagents.** Each runs in its own worktree on its own branch, based on the integration branch (reset onto it if not), builds with [test-driven-development](../test-driven-development/SKILL.md), and merges the integration tip into its own branch before reporting done, so its merge is a fast-forward. - **Merger subagents.** Land finished work on the integration branch. - **Final review.** Run [two-axis-code-review](../two-axis-code-review/SKILL.md) on the integration branch, then fix all issues in a single implementer subagent. ## Communication Keep it sparse. Communicate through **context pointers**, not duplicated content: point at the spec, the ticket and the exploration notes rather than restating them. ## Goal and cleanup The goal is the **integration branch**, not a pull request; a draft pull request opens only after the first merge. Completing a merge may change the frontier — that is the scheduler, and it is what kicks off more implementers. Clean up every worktree at the end.