A flaw in the odh-dashboard component of Red Hat OpenShift AI allows a malicious actor inside the cluster to bypass authentication by sending a forged X-Forwarded-Access-Token header on backend port 8080. The server fails to validate the origin of this token, enabling the attacker to impersonate any user. This results in unauthorized access to the Kubernetes API, which can lead to arbitrary code execution, privilege escalation, or the disclosure of sensitive information. The weakness is classified as CWE‑346, reflecting an improper origin check of a trustable credential.

The vulnerability impacts the Red Hat OpenShift AI (RHOAI) platform, specifically the odh-dashboard web console. All installations that expose odh-dashboard on the default backend port 8080 without additional origin validation are affected. No specific RHOAI version ranges are listed, so any version running this component is potentially susceptible.

The CVSS score of 8.8 indicates a high severity risk. The EPSS score of <1% shows a low but non-zero probability of exploitation; attackers would need network access within the cluster. The vulnerability is not yet listed in the CISA KEV catalog. Based on the description, the likely attack vector is via internal cluster traffic, exploiting the lack of origin validation on the X-Forwarded-Access-Token header to impersonate users.