Anthropic engineers demonstrated Claude Managed Agents by building an autonomous financial research system. The platform executes long-running tasks in secure sandboxes, coordinates specialized sub-agents, applies rubric-based self-supervision, and improves over time using persistent shared memory.