Beyond valid credentials: How exposed AWS keys are tested for Amazon Bedrock access

Datadog Security Labs reports active credential-harvesting activity targeting AWS with tooling that validates stolen keys and specifically checks for Amazon Bedrock/LLM access; the analysis details a KMON_NOC platform frontend that flags Bedrock-capable credentials, two malicious Python scripts (SHA-256s provided) that enumerate and invoke Bedrock models (using ListFoundationModels, ListInferenceProfiles, and Converse/InvokeModel), observed telemetry across multiple organizations and Datadog customers, and a list of IP and file indicators to aid detection and response.