subagent-driven-development

To execute implementation plans without context pollution, dispatch a fresh, isolated subagent for each independent task to prevent cross-task contamination and maintain clean execution state.

Automated spec compliance and code quality review works by enforcing a strict two-stage sequence where a spec reviewer validates requirements first, then a code quality reviewer checks standards, looping until both approve.

Yes, you can use TDD and commit-based handoffs for task orchestration. The implementer follows a TDD-first workflow, passing completed work via commits to the next review stage for validation.

The best way to orchestrate independent coding tasks in a single session is dispatching isolated subagents for each task, ensuring automated implementation, spec review, and code-quality verification occur sequentially.

Multi-task feature implementations fail spec compliance when the two-stage review order is violated. You must run spec compliance first, then code quality, using re-review loops until approvals are satisfied.

You should not use subagent-driven development for dependent tasks. It applies strictly to same-session implementation of largely independent tasks requiring isolated subagents, automated reviews, and commit-based handoffs.