CVE-2025-10725 (CVSS 9.9): Red Hat OpenShift AI Privilege Escalation Flaw Could Lead to Full Cluster Compromise

CVE-2025-10725 (CVSS 9.9) concerns Red Hat OpenShift AI, a platform for building, deploying and managing ML models across hybrid clouds, with a flaw that could let a low-privileged, authenticated user escalate to a full cluster administrator and compromise the entire environment. The advisory explains that the issue stems from overly permissive OpenShift AI role bindings, enabling privilege escalation that could jeopardise the cluster’s confidentiality, integrity and availability.

Exploitation could allow an attacker to steal sensitive datasets and ML models, disrupt services, and take control of the underlying infrastructure. Red Hat classifies the vulnerability as Important rather than Critical because it requires at least some legitimate access, such as a compromised user account, to begin an attack.

Until patches are applied, administrators are urged to follow the principle of least privilege, removing the ClusterRoleBinding tying the kueue-batch-user-role to the system:authenticated group and granting creation rights on a granular basis to specific users or groups.