subagent-driven-development

To prevent context pollution in complex software development workflows, delegate specific implementation tasks to isolated sub-agents. This approach ensures each task maintains a fresh context, preventing drift and preserving focus throughout the multi-task feature development cycle.

Enforcing specification compliance and code quality is best achieved through a two-stage review process. Separate sub-agents independently verify that the implementation strictly adheres to specifications and meets quality standards before final integration.

You can break down a complex feature by using a subagent-driven development approach to divide the implementation plan into independent tasks. Each task is executed by a fresh, isolated sub-agent to ensure focused and high-quality implementation.

Yes, you can use isolated sub-agents for multi-stage verification cycles in TDD. The sub-agents handle strict adherence to specifications and perform rigorous code quality reviews, satisfying the need for automated verification in test-driven development.

You should not use a subagent-driven approach for simple, single-task modifications that do not require strict specification compliance or multi-stage code quality reviews. It is specifically designed for complex, multi-task feature development requiring rigorous automated verification.