subagent-driven-development

Subagent-driven development automates task execution by dispatching independent subagents for each task, reducing context pollution. It follows a two-stage review process to verify both spec compliance and code quality, ensuring rigorous quality assurance throughout the development workflow.

To automate implementation plan execution, the skill dispatches fresh subagents for each task and requires specific prompt templates to guide them. It integrates with your development workflow to sequentially execute tasks and perform a two-stage review for spec compliance and code quality.

Yes, the two-stage review process explicitly checks both spec compliance and code quality separately. By separating these reviews, the workflow ensures that the executed tasks meet the original implementation plan requirements while maintaining high code standards.

The best way to maintain context independence is dispatching fresh subagents for each task, which prevents context pollution. This approach ensures that independent tasks in an implementation plan are executed without interference from previous task contexts.

You should not use this two-stage subagent review process for tasks that are not independent or lack a decomposed implementation plan. The workflow requires specific prompt templates and is designed specifically for independent tasks needing rigorous quality control.