subagent-driven-development

Automate task dispatch and code reviews by assigning dedicated subagents to execute independent tasks within a single session. This enforces a strict two-stage review pipeline—checking spec compliance then code quality—after each task completes.

A two-stage review pipeline enforces spec compliance first, followed by a dedicated code quality review for every task. A fresh subagent executes each task to isolate context, and a final development-branch review coordinates all completed work.

Yes, you can dispatch specialized subagents to sequentially execute clearly defined tasks. Each task undergoes mandatory plan compliance and code quality reviews before integrating with the final development-branch workflow.

Prevent cross-task context leakage by dispatching a fresh subagent for each independent task. This isolation ensures that context from one task does not interfere with another, while built-in reviews maintain overall code quality.

Avoid subagent-driven development for sessions lacking clearly defined tasks that can be parallelized or sequentially executed. It requires tasks to be strictly independent to prevent parallel conflicting work during the review pipeline.

Yes, you need to provide a predefined plan before dispatching specialized subagents. The workflow relies on plan-driven implementation to assign tasks and execute the two-stage spec compliance and code quality reviews.