What problem does it solve? Executing a multi-task implementation plan in one session risks context pollution, skipped reviews, and lost progress after compaction. This Skill orchestrates the whole plan: a fresh implementer subagent per task, a spec-and-quality review after each, a bounded fix loop, and a final whole-branch review, with a ledger file tracking every decision. ## Core Features & Use Cases - Per-task subagent dispatch: Each task gets a fresh implementer with an extracted brief file, isolated context, and an explicit model choice, so no task inherits session history. - Two-gate review loop: Every task passes a spec-compliance and code-quality review, with up to five fix rounds, model escalation on rounds 4-5, and adjudication of residual findings at the cap. - Ledger-based recovery: A per-plan workspace (.superpowers/sdd/<plan>/) holds briefs, reports, review packages, and a progress ledger, so work resumes correctly even after context compaction. - Use Case: You have a written implementation plan with eight tasks. Run this Skill to dispatch each task to an implementer, review every diff, fix findings through the loop, and finish with a whole-branch review before merging. ## Quick Start Use subagent-driven-development to execute the implementation plan at docs/plans/feature-plan.md task by task with reviews.
