subagent-driven-development

You can automate implementation plan execution by reading a structured plan file, creating a todo list, and dispatching implementer subagents for each task. This ensures systematic task execution and fast iteration in software development workflows.

A two-stage review process verifies implementation against the original specification first, then assesses the code for quality and project conventions. This mechanism ensures high-quality output by separating functional spec compliance from code quality review.

To ensure spec compliance when delegating tasks to subagents, dispatch an implementer subagent to execute the task, followed by a dedicated spec reviewer subagent. This verifies the implementation matches the original specification before proceeding.

Yes, subagent-driven development is ideal for projects where tasks are mostly independent and quality assurance is crucial. It distributes independent tasks to subagents for implementation and subjects them to a two-stage review.

You need a structured plan file and a set of subagents configured for task execution and review. The skill uses these prerequisites to create a todo list, dispatch implementer subagents, and trigger the two-stage review process.

You should avoid subagent task delegation when tasks have heavy interdependencies or when a structured plan file is unavailable. The approach relies on systematic task delegation and independent execution to ensure high-quality iteration.