subagent-driven-development

Subagent-driven development dispatches fresh subagents per task, applies test-driven development, and enforces inter-area code reviews. Load your plan from docs/plans, divide it into feature areas, then trigger a new subagent for each task with automated review before moving to the next area.

Fresh subagents per task eliminate fragmentation by isolating execution scope. Each subagent implements one task with TDD-first practice, reports results, and triggers a reviewer subagent to verify constitution compliance before the next area begins.

Yes. After each feature area completes, dispatch a dedicated reviewer subagent to conduct inter-area code reviews and ensure quality gates are met before proceeding to the next task in your plan.

Test-driven development is enforced as part of this workflow. Each subagent implements with TDD-first practice and verification runs npm tests against your project context constraints to confirm quality.

Use it for plans with clearly grouped features and independent tasks where rapid iteration and quality gates are priorities. Fresh subagents enable faster execution, automated reviews, and clear task isolation compared to monolithic approaches.

After implementation, a reviewer subagent conducts post-area code review to check constitution compliance and quality. This review gates progression to the next task, ensuring each feature area meets standards before moving forward.