What problem does it solve? When an approved implementation plan contains multiple independent slices, running them one at a time wastes wall-clock time. This Skill parallelizes slice execution by computing dependency waves from the plan, dispatching isolated workers per slice, and gating each wave with a build, typecheck, and test integration check. ## Core Features & Use Cases - Wave computation: Parses Scope and Depends-on fields from IMPLEMENTATION_PLAN.md, builds the slice DAG, and groups slices into parallelizable waves with pairwise-disjoint file scopes. - Worktree-isolated workers: Dispatches one git worktree worker per slice, each executing the implement-approved-slice contract confined to its declared scope files. - Integration gate: Merges each wave's worktrees and runs a mandatory build, typecheck, and test pass before opening the next wave. - Use Case: A task has an approved plan with six slices, three of which touch disjoint file sets. The Skill runs those three concurrently in one wave, merges them, verifies the build, then proceeds to the dependent slices. ## Quick Start Run implement-fleet on the active task to execute the approved multi-slice plan in parallel waves.
