subagent-driven-development

Subagent-driven development orchestrates independent subagents to execute project plan tasks with isolated context, preventing cross-task contamination. It applies to complex plans with many parallel tasks across different components, requiring staged reviews and clear handoffs.

Managing parallel development tasks with subagent-driven development reduces context leakage by giving each task a fresh subagent. This approach scales to large plans and parallel tasks, ensuring clear handoffs and structured workflows across components.

You implement two-stage code reviews by running spec compliance checks first, then code quality checks after each subagent completes a task. This structured workflow ensures quality control across parallel tasks in complex project plans.

You need subagent task decomposition when your project plan involves many parallel tasks across different components requiring staged reviews and clear handoffs. It orchestrates complex plans by distributing work to independent, isolated subagents to reduce context leakage.

Yes, subagent-driven development works for large plans with parallel tasks by using structured prompts and workflows that scale. It dispatches a fresh subagent per task with isolated context to prevent cross-task contamination during implementation.

Context leakage happens when multiple tasks share the same context window in a single agent. Subagent-driven development prevents this by dispatching a fresh subagent per task with isolated context, reducing cross-task contamination and speeding iteration.