subagent-driven-development

Autonomous subagent orchestration dispatches isolated implementer agents for each task within a single session. It coordinates per-task roles including spec reviewers and code-quality reviewers to ensure structured progress and strict quality gates.

Automated code review during implementation uses a three-stage workflow. It enforces spec compliance, checks code quality, and performs cross-model validation by dispatching dedicated reviewer subagents after every individual task is completed.

Subagent-driven development maintains context isolation by spawning independent subagents for each task within the same session. This ensures structured progress without cross-contamination of context between concurrent implementation tasks.

Multiple reviewer roles are supported through dedicated spec-reviewer and code-quality reviewer subagents. These roles enforce spec compliance and code quality respectively, providing cross-model validation after each task implementation.

A predefined plan is required to dispatch implementer subagents. The workflow supports a plan-to-delivery flow where tasks from the implementation plan are dispatched to isolated subagents with structured three-stage reviews.

Subagent orchestration should be used for complex implementation plans requiring parallel task handling and strict quality gates. It provides isolated context per task and three-stage reviews, unlike sequential execution which lacks autonomous role coordination.