subagent-driven-development

To execute an implementation plan with task isolation, this Skill automatically breaks down the plan into discrete tasks and dispatches a fresh subagent for each task. It then runs a two-stage review process for spec compliance and code quality after every task.

Preventing context pollution during code development is achieved by assigning a new, independent subagent to each decomposed task. This ensures each subagent maintains strict focus on its specific implementation requirements without interference from previous tasks.

The two-stage code review process works by running separate automated checks for spec compliance and code quality after each subagent completes its task. This rigorous review ensures each part of the feature is built correctly and meets quality standards before moving forward.

Yes, you can use subagent-driven development for rapid agile iteration cycles. It is specifically designed to provide structured, high-quality software development with rapid iteration by efficiently tracking progress and reviewing tasks within a single session.

Yes, you need a detailed implementation plan before starting task decomposition. The Skill requires a structured plan file, such as a markdown document located at ./plans/feature-plan.md, to automatically break down the work into discrete, manageable tasks for the subagents.