subagent-driven-development

Automating implementation plan execution involves breaking down complex features into independent tasks and dispatching specialized subagents to implement them. This approach uses a two-stage review process to ensure spec compliance and code quality for each delegated task.

Subagent-driven development is a workflow technique that dispatches a fresh subagent for each implementation task. Using a new context per task prevents context pollution, ensuring the delegate remains focused solely on its specific step without interference from previous operations.

To enforce TDD principles during automated task delegation, the execution workflow integrates testing checks within its two-stage review loop. This ensures each dispatched implementation step passes both spec compliance and code quality reviews before proceeding.

Yes, automated code review works with iterative fixes by utilizing a review loop. If a dispatched task fails the initial spec compliance or code quality check, the system facilitates iterative corrections until the implementation meets the required standards.

Yes, you can use automated task delegation for managing API endpoints and user models. The system autonomously creates user models, implements API endpoints, and writes tests by breaking detailed feature plans into manageable, independently reviewed implementation steps.

A limitation of using fresh subagents for each implementation step is that delegates lack shared memory of previous tasks, requiring explicit context passing. The system manages this by handling subagent questions and maintaining task granularity throughout the workflow.