Join our Newsletter — 33% off our NHI Course

AWS Bedrock Mantle SCP bypass: what it means for IAM teams

 

(@nhi-mgmt-group)
Member Moderator
Joined: 1 year ago
Posts: 20739
Topic starter  

TL;DR: Long-lived Bedrock API keys backed by Service Specific Credentials could bypass some SCP enforcement for Bedrock Mantle until AWS corrected the logic, according to Sonrai Security, showing how newly introduced auth paths can outpace centralized controls. The lesson is that org-level policy assumptions must be tested against every credential type, not just the default path.

Editorial analysis by NHI Mgmt Group, based on content published by Sonrai Security: “Cracks in the Bedrock: Bypassing SCP Enforcement with Long-Lived API Keys”.

Key questions

Q: Where do SCPs fail when a cloud service adds a new credential path?

A: They fail when the policy is written for the service’s standard access flow but the new credential path is evaluated differently.

Q: When should IAM teams re-test organization-level deny policies?

A: Re-test them whenever a cloud provider adds a new service namespace, alternate token mode, or service-specific credential type.

Q: What is the operational risk of long-lived API keys in cloud IAM?

A: Long-lived keys increase the chance that a credential path will outlive the assumptions built into central policy enforcement.

Practitioner guidance

  • Validate every new cloud credential mode against SCP deny logic When a service introduces short-term tokens, long-term keys, or service-specific credentials, test each one against explicit organization-level denies before allowing production use.
  • Separate governance for long-lived and short-lived bearer paths Treat long-lived API keys as a distinct control class, with separate approval, monitoring, and revocation checks rather than inheriting assumptions from default user access.
  • Deny service-specific credential creation where it is not required If a workload does not need service-specific credentials, block iam:CreateServiceSpecificCredential at the organization level so alternate access paths cannot emerge later.

Bottom line: This case shows that a centrally defined deny rule can still miss a newly introduced authorization path if the credential type is evaluated differently.

Explore further

View Full Forum →  |  NHI Foundation Course →  |  Our Services →  |  Read the full analysis →


This topic was modified 8 hours ago by NHI Mgmt Group

   
Quote
(@mr-nhi)
Member Moderator
Joined: 5 months ago
Posts: 21545
 

Policy enforcement is only as strong as the credential paths it actually covers. Sonrai Security’s findings show that centralized deny controls can look complete on paper while still missing a newly introduced authorization path. In cloud IAM, the gap is rarely the policy intent itself, but the assumption that every credential form resolves through the same enforcement logic. Practitioners should treat each new service authentication mode as a separate governance surface, not a minor variant of the old one.

A few things that frame the scale:

A question worth separating out:

Q: How should teams govern certificate-based non-human identities in AWS?

A: Treat each certificate as part of a lifecycle-managed identity, with explicit ownership, narrow trust scope, current revocation data, and role bindings reviewed against actual data access. That approach limits hidden access paths and makes machine credentials governable in the same way teams govern other NHIs.

👉 Read our full editorial: SCP enforcement gaps in AWS Bedrock Mantle expose IAM trust assumptions


This post was modified 8 hours ago by NHI Mgmt Group

   
ReplyQuote
Share:

Free weekly newsletter

Subscribe to the NHI & AI Identity Journal

The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.

Bonus 33% off our NHI Course when you subscribe.