What problem does it solve? Executing a multi-task implementation plan in one AI session leads to context pollution, skipped reviews, and lost progress after compaction. This Skill orchestrates plan execution by delegating each task to an isolated subagent with its own brief, enforcing a review gate after every task, and tracking progress in a persistent ledger. ## Core Features & Use Cases - Per-task subagent dispatch: Each plan task gets a fresh implementer subagent with an extracted task brief, a report file contract, and an explicit model choice, keeping the controller's context clean for coordination. - Two-stage review loop: Every task passes a spec-compliance and code-quality review, with a bounded five-round fix loop, scoped re-reviews, and adjudication rules for findings that survive the cap. - Ledger-based recovery: A per-plan workspace under .superpowers/sdd/ stores briefs, reports, review packages, and a progress ledger so work resumes correctly after context compaction. - Use Case: Given a written implementation plan with eight independent tasks, run this Skill to implement, test, review, and commit each task in a git worktree, then finish with a whole-branch review before merging. ## Quick Start Use the subagent-driven-development skill to execute the implementation plan at docs/superpowers/plans/feature-plan.md task by task with reviews.
