Multi-agent: who reports to whom

Your first line in any agent design is what the agent must not do. Most failures come from unclear scope, sloppy tool contracts and missing failure handling, not from a weak model.

Task: A multi-agent setup where the agents talk in circles.

[Before you start, do three things]
1. Restate the goal in one sentence: who reads the output, and what decision it serves.
2. If I have not given you the following, ask me. Do not assume:
   - can the task split into independent subtasks
   - who makes the final call
   - shared state required
   - who retries on failure
3. Say how you plan to do it (at most 5 steps) and wait for my go-ahead.

[Go through each of these]
- one orchestrator decides what happens next
- sub-agents do not talk directly; they share state
- define the handoff format and acceptance
- global step and time caps to stop loops

[Process]
1. Write the boundary: what it may do, must never do, and when to hand back to a human
2. Define each tool's inputs, outputs, failure semantics, idempotency
3. Write the flow as a state machine with exit conditions
4. Design failure paths: retry, degrade, abort, escalate
5. Give one testable success criterion

[Output]
Fixed shape: boundary, tool contracts, state machine, failure paths, criteria.

[Do not]
- No general-purpose agent that does everything
- Never let fuzzy prose stand in for a tool contract
- The agent must not decide whether approval is needed
- No 'do your best' instead of an abort condition