Amazon Web Services (AWS) updated the Amazon Bedrock AgentCore platform to add features that automate building AI agent infrastructure, SiliconANGLE reported on April 22 local time.

The key update is a "managed agent harness". Previously, deploying an AI agent in a production environment could take days as developers set up infrastructure such as computing resources, authentication protocols, persistent storage and code execution sandboxes. The managed agent harness replaces that process with writing a configuration file, allowing it to be completed in minutes.

The managed agent harness is based on the open-source Strands Agents framework.

Switching models has also become easier. For example, moving from Claude Opus 4.6 to Google Gemini 3 previously required extensive code rewrites, but now only requires adjusting API parameters.

The newly added AgentCore CLI supports keeping the same workflow from prototype development through production deployment. AWS said it applies an infrastructure-as-code approach to ensure reproducibility of agent configurations and version control.

AWS also unveiled prebuilt skills dedicated to Kiro, Claude Code, Codex and Cursor. The company said the goal is to help AI coding agents reduce mistakes by keeping up-to-date knowledge of AgentCore platform best practices.