Agent-to-agent handoffs: what proof do you require?
When one agent hands work to another, what proof do you require before accepting the result? The question probes the trust mechanics of multi-agent systems — source lists, step logs, test outputs — and what evidence is actually enough to act on a handoff.