dev-subagent-driven-development

You can automate task execution by providing an implementation plan that dispatches a fresh subagent per task. The system automatically executes implementer and reviewer rounds, applying a two-stage review to ensure spec compliance and code quality.

Subagent-driven development is an orchestration method that dispatches a fresh subagent per task to maintain isolated reasoning and reduce context bleed. It applies a two-stage review for spec compliance and code quality before running a final integration check.

Provide an implementation plan with independent tasks to initiate automated subagent orchestration. The system delegates per-task work to fresh subagents and automatically executes implementer and reviewer rounds to ensure quality and spec conformity.

Subagent-driven development is specifically applicable to plans with independent tasks where quality and spec conformity matter. Plans with highly dependent tasks may face limitations, as this workflow relies on isolated reasoning and per-task delegation.

Applying a two-stage review workflow is the best way to ensure spec compliance during automated code generation. The first stage checks spec compliance, the second reviews code quality, and a final integration pass confirms overall consistency.

Using a fresh subagent per task enables isolated reasoning and reduces context bleed across the workflow. This ensures each task is evaluated independently for spec compliance and code quality before executing the final integration check.