Multi-Agent Design Doc Generator

Draft a multi-agent system design doc from this rough idea: {{idea}}. Success should look like: {{success}}. Constraints: {{constraints}}.

Use these sections: 1 Goal and success metric; 2 Single-agent baseline and expected failure modes; 3 Why multi-agent (justify each agent with a genuine reason); 4 Agents table (model tier, tools, permissions, sees raw untrusted data?); 5 Orchestration pattern(s) and flow; 6 Hand-off contracts and shared state (with reducers); 7 Reliability (budgets, timeouts, retries, termination criteria, partial-result policy); 8 Cost and latency estimate with napkin math; 9 Evaluation plan (offline set, baseline comparison, component checks, online); 10 Risks and rollout (security, human-in-the-loop, kill switch). Where information is missing, write an explicit OPEN QUESTION instead of inventing details.