What problem does it solve?
This Skill addresses the challenge of maintaining clean task context and rapid iteration when executing plan-driven work. By dispatching a new subagent for each task, it prevents cross-task context leakage and enables structured progress with automatic per-task reviews.
Core Features & Use Cases
- Per-task subagent dispatch: A new subagent is created for each task to preserve isolation and context accuracy.
- Two-stage reviews: After each task, a spec compliance review confirms alignment with goals, followed by a code quality review to ensure implementation quality.
- Iterative task execution within a session: Supports many tasks in a single session without context contamination, while preserving traceability.
- Self-review and traceability: Subagents perform self-review before handing back results and logging changes for audit.
Use Case: You have a feature plan with multiple independent tasks. Dispatch a subagent for each task; after completion, run the spec and quality reviews; repeat until all tasks pass.
Quick Start
- Provide your implementation plan to the Skill in plan form.
- The system will dispatch an implementer subagent for each task using the prompts stored in scripts/.
- Each task will undergo a spec compliance review and a code quality review, with fixes looped until approved.
- When all tasks are approved, the Skill completes the plan.
