Amazon Bedrock AgentCore Identity adds a managed consent portal

What changed

Source summary

AWS added a managed consent portal to Bedrock AgentCore Identity, giving each AgentCore Gateway a hosted web client so end users can authorise and check the status of AI agent connections to third-party services such as GitHub, Salesforce and Slack, without developers having to build custom OAuth callback infrastructure.

Why it matters to your work

Practical interpretation

Mainly relevant if your organisation builds or manages AI agents on AWS Bedrock. This removes a chunk of custom engineering work needed to let staff safely approve which external tools, such as GitHub, Salesforce or Slack, an AI agent can access on their behalf.

When this could help

Worth adopting if you are building agents on Bedrock AgentCore that connect to OAuth-protected third-party services and currently maintain custom callback infrastructure.

The managed consent portal removes the need to build, host, and maintain custom OAuth callback infrastructure for three-legged authorization flows. Each AgentCore Gateway gets its own hosted portal where admins share a URL with their team, users grant permissions before sessions start, and IDE-based agent clients that cannot display consent URLs use it as a separate authorization surface.

How to get started

1. Configure your AgentCore Gateway with the third-party credential providers. Share the portal URL with your team so users can grant agent permissions before sessions start. IDE-based clients use the portal as a separate authorization surface.

   Provider guidance ↗

Build it into your routine

SCALA suggested workflow

1. Simplify agent OAuth setup

   When adding a new third-party tool connection to your Bedrock agent, use the managed consent portal instead of building custom OAuth callback infrastructure.

What to check

• Pricing, supported region list, and limits on gateways/providers/users are not stated in the announcement. Source ↗