AWS cuts agent setup to 3

Amazon Web Services updated Amazon Bedrock AgentCore on April 22, adding a managed harness that lets developers launch an AI agent in three API calls. (aws.amazon.com 1) (aws.amazon.com 2) AgentCore is AWS’s service for running AI agents, which are software systems that call models, tools and memory in sequence to complete tasks. The new harness moves that wiring into configuration, so teams declare a model, instructions and tools instead of writing orchestration code first. (docs.aws.amazon.com) (aws.amazon.com) AWS paired the harness with an AgentCore command-line interface, or CLI, for local setup, testing and deployment. It also added a persistent filesystem so agents can pause work, keep files and resume later without losing state. (aws.amazon.com 1) (aws.amazon.com 2) The company also introduced AgentCore skills for coding assistants including Claude Code, Kiro and Codex. AWS said those packaged workflows are meant to give developers reusable patterns for common coding tasks inside agents. (aws.amazon.com 1) (aws.amazon.com 2) The release is aimed at a bottleneck in enterprise AI projects: teams can prototype a chatbot quickly, but production agents need identity controls, tool permissions, memory and runtime management. AWS said AgentCore’s harness bundles compute, tooling, memory, identity and security into one managed layer. (aws.amazon.com) (docs.aws.amazon.com) AWS said the managed harness preview is live in four regions: US West (Oregon), US East (N. Virginia), Europe (Frankfurt) and Asia Pacific (Sydney). The AgentCore CLI is available in 14 AgentCore regions, while availability for other features varies by service. (forbes.com) (aws.amazon.com) The timing lines up with a broader AWS push to make Bedrock the default control plane for enterprise AI tools. On April 21, GitLab said joint customers can route GitLab Duo Agent Platform inference through Amazon Bedrock using existing AWS accounts, Identity and Access Management policies and spending commitments. (about.gitlab.com) (finance.yahoo.com) Amazon also said on April 21 that it would invest an additional $5 billion in Anthropic, with plans for up to $20 billion more, while Anthropic committed $100 billion to AWS technologies over 10 years. Amazon said the deal includes up to 5 gigawatts of Trainium capacity for Anthropic’s models. (aboutamazon.com) The through line is that AWS is trying to sell more than models. It wants Bedrock and AgentCore to be the place where companies build, secure and operate agents without stitching together separate infrastructure first. (aws.amazon.com) (docs.aws.amazon.com)