Impact
The vulnerability in the OpenShift AI dashboard allows an authenticated user to create RoleBindings without validating the roleRef field. By specifying an arbitrary role, attackers can grant themselves or others highly privileged roles such as cluster-admin, resulting in unauthorized elevated access within their namespace and potentially persistent control over the cluster. This flaw is a form of privilege escalation that can compromise confidentiality, integrity, and availability of cluster resources.
Affected Systems
The affected product is Red Hat OpenShift AI (RHOAI), as identified by the CNA. The flaw is present in all released versions of the odh-dashboard component bundled with RHOAI; no version exclusions are specified in the advisory.
Risk and Exploitability
The CVSS score of 8.8 categorizes this issue as high severity, indicating serious risk if mitigated. EPSS score of 0.00361 indicates a very low exploitation probability, but that does not diminish the risk posed by unauthorized privileged RoleBinding creation. The flaw is not listed in CISA KEV, yet the high CVSS combined with the privilege escalation capability suggests that attackers can gain persistent control if they possess legitimate dashboard credentials. The likely attack vector requires an authenticated session with the OpenShift AI dashboard, after which the attacker can submit a RoleBinding manifest with an arbitrary roleRef. The success of the exploit depends on the existence of privileged ClusterRoles such as cluster-admin in the cluster.
OpenCVE Enrichment