Home icon

Securing Amazon Bedrock AgentCore Runtime with AWS WAF

Machine Learning Blog



This article presents two architecture patterns for securing Amazon Bedrock AgentCore Runtime with AWS WAF through an Application Load Balancer and VPC Interface Endpoint.

  • Pattern 1 uses a Lambda proxy between ALB and VPC Endpoint for request transformation and custom authentication logic
  • Pattern 2 targets VPC Endpoint ENI IPs directly from ALB for minimal latency and simplicity
  • Both patterns solve the health check authentication challenge by using unauthenticated health checks while passing authenticated production traffic
  • Resource policies prevent direct access to AgentCore, ensuring all traffic flows through AWS WAF
  • Defense-in-depth security includes AWS WAF, security groups, authentication, and resource policies
  • Pattern 1 adds 50-200ms latency but enables request transformation; Pattern 2 has no Lambda overhead

Both patterns provide tested implementations for securing AI agent deployments with multiple security layers while supporting SigV4 and OAuth authentication methods.



Go to article

The AWS News Feed is currently looking for gold sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.

Related articles

Jun 29
2026
AWS WAF adds support for Amazon Bedrock AgentCore Gateway
Aug 13
2025
Securely launch and scale your agents and tools on Amazon Bedrock AgentCore Runtime
Apr 30
2026
Configuring Amazon Bedrock AgentCore Gateway for secure access to private resources
Oct 14
2025
Securing AI agents with Amazon Bedrock AgentCore Identity

The AWS News Feed is currently looking for silver sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.