OpenShift AI Privilege Escalation Flaw Exposes Clusters
🔒 Red Hat has disclosed a severe privilege escalation vulnerability in OpenShift AI (CVE-2025-10725) that can allow an authenticated, low-privileged user to escalate to full cluster administrator and fully compromise a deployment. The issue carries a CVSS score of 9.9 but is rated Important by Red Hat because exploitation requires an authenticated account. Affected releases include OpenShift AI 2.19, 2.21 and RHOAI. Administrators are advised to avoid broad ClusterRoleBindings such as binding kueue-batch-user-role to system:authenticated, and to grant job creation permissions only on a granular, need-to-know basis while applying vendor guidance.
