AgentCore harness is now generally available

AWS announces the general availability of the managed agent harness in Amazon Bedrock AgentCore, enabling teams to go from idea to working agents in minutes. The harness handles orchestration loop execution, tool execution, context window management, state persistence, failure recovery, and session isolation. Instead of coding the loop, customers define an agent in configuration: the model it uses, the tools it calls, the skills it accesses, and the instructions it follows. AgentCore assembles and runs that loop from a single definition, producing a production-grade agent in minutes with a filesystem, shell, memory, AWS-curated skills, and web browsing. The harness is decoupled from the model, allowing customers to choose any model and switch providers mid-session without losing context. It is part of a single platform, reaching tools through the same gateway that enforces security policies and connecting the agent to organizational knowledge and web search. Identity, memory, and observability come from that same platform, ensuring every agent action is governed and traced from the first call. When custom orchestration is needed, a single CLI command exports the harness to Strands-based code on the same compute and primitives, with Claude Agent SDK coming soon as an export target.