概要
Subagent Driven Development provides a structured framework for managing complex coding tasks by delegating specific, bounded work to fresh AI instances. By following a standardized handoff process involving clear scopes, success criteria, and deliverables, this skill prevents the context pollution that often leads to AI loops or hallucinations in long-running sessions. It is particularly useful for parallelizing work, conducting deep research without cluttering the main conversation, and breaking through difficult debugging sessions with a fresh perspective.
