subagent-driven-development

Automated task execution dispatches a fresh subagent for each independent item in a development plan, maintaining isolated context. It follows a two-stage review process for spec compliance and code quality to ensure your plan meets requirements.

A two-stage code review workflow evaluates subagent output first for spec compliance and then for code quality. This sequential process ensures implemented tasks match specifications before validating the overall code structure and maintainability.

To execute a feature plan, pass your implementation plan file to the automated workflow. The system dispatches isolated subagents for each task and runs the two-stage review process automatically to verify spec compliance and code quality.

Subagent-driven development requires specific prompt templates for implementing subagents and handling reviews. You need the implementer, spec-reviewer, and code-quality-reviewer prompt files to define execution and two-stage review workflows.

Use isolated subagents for development workflow automation when implementing well-defined development plans where tasks can be executed independently. This approach maintains strict context boundaries between distinct features or components without state overlap.