Malicious GCP Service Account Backdoor via serviceAccountTokenCreator Grant

PremiumReviewedSigma · High · v1
Product
gcp
Service
gcp.audit
Author
HuntRule
Published
2026-09-15
Updated
2026-09-15

ATT&CK techniques

Persistence → Lateral Movement
  1. Recon

  2. Resource Dev

  3. Initial Access

  4. Execution

  5. Cred Access

  6. Discovery

  7. Collection

  8. C2

  9. Exfiltration

  10. Impact

What it detects

This rule detects a google.iam.admin.v1.SetIAMPolicy admin activity event that grants the roles/iam.serviceAccountTokenCreator role, allowing a principal to mint access tokens for a targeted service account. Adversaries use this binding to backdoor a service account and later impersonate it through GenerateAccessToken, keeping durable privileged access. This matters because the grant converts any controlled identity into a lasting path to the service account privileges.

Related detections9 linkedT1098.003 — drag to rearrange
Suspicious GCP Service Account Impersonation via GenerateAccessToken
Suspicious Federated Identity Credential On-Behalf-Of Token Flow via Sign-In Logs
Suspicious Service Principal Creation via Azure Audit Logs
Suspicious Salesforce Bulk API Export via aiohttp User Agent
Malicious IAM Group Creation for SES Abuse by JavaGhost (via cloudtrail)
Malicious Administrator Policy Attachment for Privilege Escalation in AWS
Suspicious Azure Contributor Role Assignment to Identity
Suspicious Entra Sign-In With Axios Automation User-Agent via AiTM Token Replay (via azure)
Malicious Azure AD Token Abuse via Roadtx
Malicious GCP Service Account Backdoor via serviceAccountTokenCreator Grant
Pivot detection · T1098.003 · 9 related

Detection content is published as a reviewed draft. Tune thresholds and exclusions against your own telemetry before enabling this rule for alerting.