AWS Bedrock AgentCore Execution Role Used Outside Its Runtime

Description

Identifies an Amazon Bedrock AgentCore execution role (an AssumedRole identity whose role name begins with "AgentCore-" or contains "BedrockAgentCore") making an AWS API call to a service it has not previously called. AgentCore runtimes normally interact only with Bedrock inference, AgentCore data-plane, and observability services (CloudWatch Logs, X-Ray, CloudWatch metrics), so an execution role suddenly calling STS, EC2, IAM, Secrets Manager, or other services is a strong indicator that the role's temporary credentials were exfiltrated from the agent's microVM (for example, via the Code Interpreter instance-metadata-service credential theft) and are being used outside the runtime for reconnaissance, privilege escalation, or lateral movement. Because the stolen credentials are recorded in CloudTrail under the execution role's own identity, the anomalous service usage, not the identity, is the detectable signal.

Detection logic

Its licence does not clear it for publishing here

Sunturai publishes a detection's own text where the licence it arrived under has been reviewed and permits it, and Elastic License 2.0 has not. The query as its source wrote it, its canonical form and the hash that pins this revision are in the workspace record.

Detection requirements

Platform: Containers ESXi IaaS Identity Provider Linux macOS Network Devices Office Suite SaaS Windows
The rule states no platform. This is derived from the 2 ATT&CK techniques it maps to.
Log source category: event_index

Known benign triggers

An AgentCore agent that legitimately integrates with additional AWS services will produce a first-time service call for its execution role. Verify the role in "aws.cloudtrail.user_identity.session_context.session_issuer.arn", the action in "event.action" and "event.provider", and the origin in "source.ip" and "source.as.organization.name", and confirm the activity matches the agent's intended design. Known agent integrations can be excluded after validation.

References

Links the source published beside this detection. Reference is the link the source wrote, unmodified. Class is decided by this catalogue from the host and path, not stated by the source. Cited by counts published detections citing the same resource, including this one, and the sources beneath it count the repositories this catalogue ingests, which is not the same as independent research.
Reference	Class	Cited by
sonraisecurity.com/blog/sandboxed-to-compromised-new-research-exposes-credential-exfiltration-paths-in-aws-code-interpreters/	Unclassified	2
unit42.paloaltonetworks.com/bypass-of-aws-sandbox-network-isolation-mode/	Threat report	3

Detection rules

AWS Bedrock AgentCore Execution Role Used Outside Its Runtime

Description

Detection logic

Detection requirements

Known benign triggers

MITRE ATT&CK mappings

References

From the source

Licence

Authorship

Detection rules