Skip to content

Cloud Project Compromise Assessment

ID: S1005

Description: Evaluate a cloud project for malicious or unknown activities and determine if the cloud project is compromised. This is distinct from evaluating whether a particular instance in that cloud project is compromised.

Has an attacker gained access to the cloud project? (Initial Access) F1013

  • Q1039  Are there any detections of exposed service account credentials?
  • Q1040  Are there any interaction with the cloud project resources from an unknown external IP address (including Tor Exit nodes, C2 tagged addresses)?
  • Q1041  Are there any unexpected successful API calls from an unknown external IP address (including Tor Exit nodes, C2 tagged addresses)?

Are there any signs of an attacker trying to maintain access to the cloud project? (Persistence) F1014

  • Q1042  Are there any APIs recently enabled?
  • Q1043  Are there any role bindings to cluster admin for anonymous users?
  • Q1044  Are there any unexpected resources created under the cloud project?
  • Q1045  Are there any unknown accounts created?
  • Q1046  Are there any unknown role bindings created?

Are there any signs of interference with cloud project protective measures? (Defense Evasion) F1015

  • Q1047  Are there any suspicious modifications to the cloud project firewall rules/policies?
  • Q1048  Are there any suspicious modifications to the cloud project logging configuration?

Are there indications of an attacker moving between resources? F1016

This may take the form of an attacker moving between cloud resources within the project or between the cloud project and on-prem systems.

  • Q1049  Is there an unusual increase of traffic between resource in the cloud project?

Are there any signs of data exfiltration from resources inside the cloud project? (Data Exfil) F1017

  • Q1050  Are there any GCS buckets within the cloud project shared externally?
  • Q1051  Are there any signs of data within the cloud project being transferred externally?
  • Q1039  Are there any detections of exposed service account credentials?