the-controller-subagent-driven-development

To automate iterative development, this Skill dispatches independent subagents for each task within an implementation plan, managing execution and iterative fixes within a single session. It enforces a two-stage review process to ensure high quality.

The two-stage review process evaluates each task's implementation first for spec compliance, then for code quality. This mechanism ensures generated code meets both the original plan requirements and high software engineering standards.

You execute a feature implementation plan by pointing the Skill to your plan file. It automatically decomposes the plan into individual tasks, dispatching a dedicated subagent to handle each specific user story.

Yes, you can use subagent-driven development for TDD and code review. It automates code generation for each user story and rigorously reviews the output for spec adherence and quality before iterating.

You need to provide an implementation plan file, such as a markdown document located at docs/plans/feature-plan.md. The Skill reads this file to break down complex plans into independent subagent tasks.

Using a subagent approach eliminates manual context switching by automatically generating code for each user story within a single session. It handles question answering and iterative fixes, enabling rapid development without losing context.