What is the primary function of Claude Managed Agents?	A suite of APIs for building and deploying agents at scale, hosted on Anthropic's infrastructure.
What is the core "agent loop" that Claude executes?	Reason, call a tool, read the result, and repeat until the job is done.
What capabilities does the isolated container provide for agent execution?	Full file system access, bash execution, and web search.
What component defines and checks what "done" looks like for an agent?	Outcomes, which are implemented using rubrics and separate graders.
What is the purpose of the Memory store in an agent workflow?	To allow the agent to read context before starting and write state after finishing, enabling continuity.
What does Multi-agent coordination allow an agent system to do?	A coordinator agent delegates tasks to multiple specialist agents, who run in separate contexts but share a file system.
What do MCP servers facilitate in an agent workflow?	Connections to external services like Slack and Asana.
What is the key benefit of using Managed Agents over running the loop yourself?	Anthropic hosts the agent loop, eliminating the need for the user to run the complex infrastructure themselves.
Name three key building blocks provided by Managed Agents.	Agents (definitions), Sessions (individual runs), and Environments (sandboxes).
How are multiple tasks handled in Managed Agents?	By running multiple sessions, which are isolated containers that can operate in parallel.