A multi-agent exploration of a monorepo crashes about 70% of the way through. Restarting it repeats every agent's work from the beginning. What is the best way to design for recovery?

Answer: A — Have each agent export structured state to a known location, and have the coordinator load a manifest on resume

Structured state exports plus a coordinator manifest let the run resume where it stopped, with each agent's state injected into its prompt. (B) depends on conversational recall, which is unreliable. (C) reduces neither the cost of a crash nor the lack of saved state. (D) depends on memory that didn't survive the crash.