What problem does it solve?
Prevents context pollution and manual coordination when executing multi-step implementation plans by dispatching isolated subagents per task and enforcing structured review gates, so work progresses safely and predictably within the same session.
Core Features & Use Cases
- Fresh subagent per task: Each task runs in an isolated context to avoid leaking session state or history.
- Two-stage review pipeline: Enforce spec compliance review first, then a separate code quality review before marking tasks complete.
- Implementer lifecycle & statuses: Implementers must ask clarifying questions, run tests, self-review, and report precise statuses (DONE, DONE_WITH_CONCERNS, NEEDS_CONTEXT, BLOCKED) to drive correct re-dispatching.
- Model selection guidance: Use lower-power models for mechanical tasks and stronger models for integration, architecture, or troubleshooting.
- Use case: Execute a development plan by extracting tasks to a TodoWrite, dispatch implementer subagents for each task, and iterate reviews until the branch is ready to finish.
Quick Start
Dispatch this skill to run your implementation plan by extracting tasks, running an implementer subagent per task, and iterating spec compliance and code-quality reviews until all tasks are approved.
