Autonomous Agent Memory: A Restart Checklist

An autonomous agent's memory design should answer a practical question: if the process stops now, what will the next run need to continue correctly? Start with