subagent-driven-development

Development automation divides tasks into isolated subagents for execution, ensuring code meets specifications and quality standards. It uses a two-phase review to verify adherence without context leakage, minimizing manual intervention for high-quality iterations.

A two-phase review verifies that code generated by isolated subagents meets specifications and adheres to quality standards. It checks specification compliance and code quality sequentially, ensuring the automated development output aligns with the original plan.

Prevent context leakage by distributing tasks to isolated subagents, each operating with its own independent context. This ensures specification compliance and code quality without cross-task interference during the automated development workflow.

Use subagent-driven development automation when executing plans for tasks with clear, independent specifications. It is ideal for scenarios requiring strict specification compliance, code quality, and quick iterations with minimal manual intervention.

Subagent-driven development is designed for tasks with clear, independent specifications. It may not be suitable for tasks with overlapping dependencies or unclear requirements, as isolated subagents require distinct contexts to ensure specification compliance.

Start subagent-driven development automation by running the skill command, which creates subagents for each task. The subagents execute the development plan, undergo a two-phase review, and ensure specification compliance automatically.