Executive Summary

Between October 14 and 26, 2026, an unknown threat actor compromised two IAM user accounts belonging to Qubole, an enterprise cloud data-processing platform, and exploited cross-account IAM trust relationships to silently assume roles inside dozens of downstream customer AWS environments. Once inside each tenant, the attacker enumerated S3 buckets, harvested secrets from AWS Secrets Manager, SSM Parameter Store, and Lambda environment variables, then validated credentials with the open-source TruffleHog tool before bulk-downloading accessible objects—following an identical sequence across every victim account. Sentra confirmed dozens of corporate victims and millions of files stolen from at least one organization, while the full scope of compromised credentials—including Microsoft Entra ID and Okta secrets—remains under investigation by Wiz and Mandiant.

This incident underscores a rapidly maturing attacker playbook centered on supply-chain identity abuse: rather than attacking hardened perimeters directly, adversaries compromise a trusted third-party SaaS or data-platform account and ride pre-approved IAM trust policies into multiple high-value cloud tenants simultaneously, achieving massive scale with minimal noise and no malware required.

Why This Matters Now

Cross-account IAM trust abuse is surging as enterprises accelerate multi-cloud SaaS integrations, leaving dormant but permissive trust policies as silent attack surfaces. Regulatory bodies including CISA and ENISA have highlighted third-party cloud identity risk in 2026 guidance, and several parallel campaigns targeting similar integration patterns have been disclosed this year—making unused cross-account roles an urgent remediation priority for any organization using managed data or analytics platforms.

Attack Path Analysis

MITRE ATT&CK® Techniques

Potential Compliance Exposure

Sector Implications

Sources

Frequently Asked Questions

The attackers compromised two Qubole IAM user accounts and then leveraged pre-existing cross-account IAM trust policies that customer organizations had configured to grant Qubole access. By calling the AWS STS AssumeRole API with Qubole's credentials, they silently inherited permissions inside each downstream tenant without needing to compromise those accounts directly.

Cloud Native Security Fabric Mitigations and ControlsCNSF

Based on the attack progression modeled above, these are the defensive controls that would constrain each stage.

This incident exploited trusted cross-account IAM relationships to silently traverse dozens of AWS environments, making it a direct use case for Aviatrix Zero Trust CNSF — where identity-aware segmentation, east-west enforcement, and controlled egress policies would likely have reduced the attacker's reachability and constrained the blast radius across victim tenants.

Initial Compromise

Control: Cloud Native Security Fabric (CNSF)

Mitigation: The CNSF fabric's real-time behavioral inspection could have flagged abnormal API call patterns from the compromised Qubole IAM credentials, potentially reducing the window in which the attacker operated undetected and limiting their ability to initiate AssumeRole calls from unexpected network origins.

Privilege Escalation

Control: Zero Trust Segmentation

Mitigation: Zero Trust Segmentation enforcing least-privilege, identity-scoped policies could have meaningfully limited which external identities were permitted to assume cross-account roles, reducing the number of environments the attacker could access from a single compromised source.

Lateral Movement

Control: East-West Traffic Security

Mitigation: East-west traffic controls monitoring inter-account and service-to-service API call patterns would likely have identified the anomalous volume of cross-account enumeration activity, potentially constraining how far the same compromised identity could propagate across tenant boundaries.

Command & Control

Control: Multicloud Visibility & Control

Mitigation: Centralized multicloud visibility across the AWS control plane could have surfaced the high-velocity, repetitive API enumeration pattern — including the automated credential harvesting rate of roughly 100 credentials per seven minutes — potentially enabling faster detection and throttling of attacker activity.

Exfiltration

Control: Egress Security & Policy Enforcement

Mitigation: Egress security controls enforcing per-role outbound data transfer limits and FQDN-based policy enforcement would likely have constrained the volume of S3 data leaving each environment, reducing the scale of exfiltration even if the attacker's initial access could not be stopped.

Impact (Mitigations)

Had earlier CNSF controls constrained lateral movement and egress activity, the attacker's reach across customer tenants would likely have been reduced — meaning fewer environments compromised, fewer secrets harvested, and a significantly smaller volume of S3 content successfully exfiltrated during the twelve-day campaign window.

Impact at a Glance

Affected Business Functions

  • Cloud Data Analytics and Big Data Processing
  • AWS S3 Object Storage and Data Lake Operations
  • Secrets and Credentials Management (AWS Secrets Manager, SSM Parameter Store)
  • Third-Party SaaS Integration and Cloud Identity Federation
Operational Disruption

Estimated downtime: N/A

Financial Impact

Estimated loss: N/A

Data Exposure

Millions of files exfiltrated from S3 buckets across dozens of victim organizations, including AWS access keys, Microsoft Entra ID client secrets, Okta client credentials, Lambda environment variables, and potentially sensitive business data, PII, and proprietary datasets stored in cloud data lakes. The attacker harvested approximately 100 passwords and secrets per seven minutes across affected environments. The breadth of exposure spans multiple downstream enterprise customers of Qubole, amplifying the supply-chain blast radius significantly beyond any single organization.

Recommended Actions

  • • Apply Zero Trust Segmentation with least-privilege identity-based policies to all cross-account IAM roles, enforcing that unused or dormant roles are automatically scoped, flagged, and require periodic revalidation before AssumeRole operations are permitted from third-party provider accounts.
  • • Deploy Multicloud Visibility & Control with centralized traffic observability across AWS control plane API activity to baseline normal Qubole-initiated API patterns and generate high-fidelity alerts when anomalous cross-account enumeration or high-velocity secrets retrieval is detected.
  • • Enforce Egress Security & Policy Enforcement with per-role S3 transfer rate limits and FQDN-based outbound filtering to detect and block bulk data exfiltration events, triggering automated session revocation when GetObject call volumes exceed defined thresholds within a rolling time window.
  • • Implement East-West Traffic Security controls to monitor and restrict inter-account API pivoting, ensuring that a single compromised third-party identity cannot laterally traverse multiple customer tenants without triggering workload-to-workload flow anomaly detection and policy-based isolation.
  • • Leverage CNSF fabric-level inline enforcement and distributed policy controls to require real-time contextual verification — including source network validation and behavioral scoring — before any cross-account IAM role assumption is honored, reducing the attack surface of supply-chain identity trust relationships.

Secure the Paths Between Cloud Workloads

A cloud-native security fabric that enforces Zero Trust across workload communication—reducing attack paths, compliance risk, and operational complexity.

Cta pattren Image