| name | subagent-driven-development |
| description | Use when executing implementation plans with independent tasks in the current session |
Subagent-Driven Development
Execute plan by dispatching fresh subagent per task, with two-stage review after each: spec compliance review first, then code quality review.
Core principle: Fresh subagent per task + two-stage review (spec then quality) = high quality, fast iteration
Continuous execution: Do not pause to check in with your human partner between tasks. Execute all tasks without stopping.
When to Use
Use when you have an implementation plan with mostly independent tasks in the same session.
vs. Executing Plans (parallel session): Same session, fresh subagent per task, two-stage review, faster iteration.
The Process
For each task:
- Dispatch implementer subagent with full task text and context
- Implementer can ask questions before starting
- Implementer implements, tests, commits, self-reviews
- Dispatch spec compliance reviewer subagent
- If spec issues found → implementer fixes → re-review
- Dispatch code quality reviewer subagent
- If quality issues found → implementer fixes → re-review
- Mark task complete
After all tasks: Dispatch final code reviewer, then use finishing-a-development-branch.
Handling Implementer Status
- DONE: Proceed to spec compliance review
- DONE_WITH_CONCERNS: Read concerns, address before review
- NEEDS_CONTEXT: Provide missing context, re-dispatch
- BLOCKED: Assess blocker, re-dispatch with better model or break task
Red Flags
Never: Skip reviews, proceed with unfixed issues, dispatch multiple implementers in parallel, start code quality before spec compliance is done.
