Oriel — AI & technology worth noticing, every day.

Extend Claude Code to remote sessions on Amazon Bedrock AgentCore with security controls

Web page · https://aws.amazon.com/blogs/machine-learning/extend-claude-code-to-remote-sessions-on-amazon-bedrock-agentcore-with-security-controls

Run Claude Code as a remote agent on Amazon Bedrock AgentCore. Each coding task runs in its own isolated session in your AWS account, every model call goes through Amazon Bedrock, and the result comes back as a GitHub pull request you review and merge.

First seen by Oriel on .

AI summary

This AWS post presents a sample remote coding agent that runs Claude Code in isolated Amazon Bedrock AgentCore sessions and opens reviewable GitHub pull requests with layered security controls.

  • Each task runs in its own AgentCore microVM inside the customer’s VPC, using Claude Opus for planning and review and Claude Sonnet for implementation.
  • The workflow clones a registered repository, retrieves short-lived GitHub access through AgentCore Identity, runs repository-defined setup and tests, and retries up to three times.
  • A Lambda function screens submitted prompts with Amazon Bedrock Guardrails, while private endpoints, restricted IAM permissions, outbound HTTPS controls, and repository-specific runtimes limit access.
  • The agent records verification commands, exit codes, and attempts in draft or ready pull requests, but never merges changes, leaving shipping decisions to humans.
  • The sample targets trusted repositories and scoped tasks; it is unsuitable for untrusted code and has deployment times of 15–25 minutes plus ongoing idle infrastructure costs.

#amazon bedrock #agentcore #claude code #github automation

Generated by gpt-5.6-sol. Check the original for details.

Aliases

SourceExternal IDURL
AWS Machine Learning Blog7f770c3595174ddce3e090066fd2d16df2549723https://aws.amazon.com/blogs/machine-learning/extend-claude-code-to-remote-sessions-on-amazon-bedrock-agentcore-with-security-controls/ Read discussion