subagent-driven-development

You can execute a development plan by dispatching a fresh subagent per task, which provides isolated context for each implementation step and enables rapid iteration. The system manages task dispatch and provides progress updates for traceability.

Context isolation means each task is handled by a fresh subagent with its own separate context, preventing information bleed between tasks. This approach maintains high-velocity iterations and ensures strict reviews improve overall code quality.

The two-stage review workflow enforces spec compliance first, then evaluates code quality. This structured review process ensures implementation tasks meet plan requirements before assessing the technical quality of the code produced.

Use subagent-driven development for plans with multiple independent tasks where strict traceability is required. It is ideal when rapid iteration and strict reviews are needed to improve quality across isolated implementation contexts.

Yes, you need to provide your implementation plan to initiate subagent task dispatch. The system uses this plan to automatically dispatch a fresh subagent for each independent task within the plan.

Yes, the subagent-driven development system provides progress updates during task execution. These updates ensure traceability throughout the implementation process as fresh subagents complete their assigned tasks.