Building with Claude Managed Agents
You already built the agent harness once, by hand. Anthropic just shipped one. Building with Claude Managed Agents is about handing work to an agent that Anthropic runs for you, and getting back a finished task instead of a transcript.
Across thirteen parts it goes from your first managed agent to a deep research agent that composes every primitive at once: reading the event stream, custom tools, permissions, MCP, skills, memory, spec-driven hand-off, parallel agent teams, multi-user isolation, and deployment.
Start at Part 1, or jump to any part below.
The parts
- You Already Built the Agent Harness. Anthropic Just Shipped One.
- Your First Claude Managed Agent: From Zero to a Finished Task in Four Calls
- "Idle" Does Not Mean Done: How to Actually Read a Managed Agents Event Stream
- The Agent Cannot Call Your Database. Custom Tools Are How You Hand It Back.
- Two Words Stand Between Your Agent and a Deleted Ledger
- Your Agent Lives in a Box. MCP Is How It Reaches Your Billing System.
- The Skill You Write in Claude Code Is the Same Skill Your Agent Ships With
- Your Agent Solves the Same Problem Every Month. Memory Stores End That.
- Stop Conversing With Your Agent. Hand It a Spec and Walk Away.
- One Agent Works in Sequence. A Team Works in Parallel.
- One Agent, Many Users, Zero Secret Store
- Your Agent Has No Deploy Button. It Is Already Deployed.
- How to Build a Deep Research Agent That Composes Every Primitive at Once