subagent-driven-development

Automated task execution dispatches independent subagents for each task in an implementation plan, managing execution within the current session while ensuring fresh context to prevent pollution. It requires specific prompt templates for implementer, spec reviewer, and code quality reviewer subagents to facilitate deterministic completion.

The two-stage code review process implements a spec compliance review followed by a code quality review after each task's completion. This subagent-driven approach ensures code meets both specification and quality standards before iteration, significantly speeding up development cycles.

To execute a detailed feature plan located in a markdown file, use the skill to autonomously implement each task, review the code through the two-stage process, and iterate until it meets both specification and quality standards. It dispatches independent subagents for each task.

Yes, specific prompt templates for implementer, spec reviewer, and code quality reviewer subagents are required to facilitate deterministic task completion and quality assurance. These templates drive the independent subagents to execute and review tasks within the session.

Yes, dispatching independent subagents for each task manages execution within the current session, ensuring fresh subagent context and preventing pollution. This approach maintains high-quality code iteration by isolating task execution and review processes.

The best way to ensure spec compliance and code quality is through a two-stage review process using independent subagents. After each task's completion, a spec compliance review is followed by a code quality review, iterating until standards are met.