Parallel Agents

Step 2: Agent Assignment

Assign based on the Maestro Mapping:

• Backend Specialist: API, DB, Server logic.
• Frontend Specialist: UI, CSS, React components.
• DevOps Engineer: CI/CD, Deployment, Monitoring.
• Debugger: Root cause analysis and fixes.

Step 3: Synchronization (The "Handover")

If Agent A depends on Agent B (e.g., Frontend needs API schema), Agent B must complete its task and provide the Contract (JSON schema, Type definitions) before Agent A starts.

🧠 Context Management Strategy

• Shared Memory: Use .agent/shared/ for common types, schemas, and ADRs.
• Status Updates: Each sub-agent updates its status in the task tracking file ({task-slug}.md).
• Conflict Resolution: If two agents touch the same file, the Orchestrator must merge changes or sequence the tasks.

🚫 Anti-Patterns to Flag

• "The Fog of War": Giving sub-agents the entire codebase without specific instructions.
• "The Echo Chamber": Multiple agents reporting the same findings.
• "Circular Dependency": Agent A waiting for Agent B, while Agent B waits for Agent A.

✅ Orchestration Checklist

• Task broken into atomic sub-tasks?
• Each sub-task assigned to the correct specialist?
• Dependencies between agents identified?
• Shared schemas/contracts provided?
• Verification steps defined for each agent?