subagent-coordination

Subagent Coordination

Inspired by NeoLabHQ/subagent-driven-development.

Patterns for effectively coordinating multi-agent task execution.

Core Principle

Fresh context per task + review between tasks = high quality, fast iteration.

Execution Patterns

Sequential (Tightly Coupled Tasks)

When tasks depend on each other:

1. Dispatch one task at a time
2. Review output after each task
3. Only proceed when current task passes review
4. Fix issues before moving to next task

Parallel (Independent Tasks)

When tasks are independent:

1. Group tasks by domain (different files, subsystems)
2. Dispatch one agent per domain
3. Review all outputs together after completion
4. Check for conflicts between parallel changes

Task Dispatch Best Practices

Good Task Prompts

• Focused — One clear problem domain
• Self-contained — All context needed is included
• Specific about output — What should the agent return?
• Constrained — "Don't change files outside of X"

Bad Task Prompts

• ❌ Too broad: "Fix all the tests"
• ❌ No context: "Fix the race condition"
• ❌ No constraints: Agent might refactor everything
• ❌ Vague output: "Fix it"

Review Checkpoints

After each task or batch:

1. Read summary — Understand what changed
2. Check for conflicts — Did agents edit the same code?
3. Run verification — Tests, lint, build
4. Spot check — Review critical changes manually

In Meeting Context

As PM in a meeting:

• Assign clear tasks to DEV: "Implement X in file Y"
• Ask QE to review after DEV completes
• If FAIL verdict, direct DEV to fix specific issues
• Track progress across turns