The trustyai-service-operator's LMEvalJob controller in Red Hat OpenShift AI contains a flaw where it does not properly filter protected environment variables in user-supplied sidecar containers. This allows any authenticated user within the cluster to override admin-controlled security policies, enabling remote code execution within these sidecar containers. The vulnerability is classified as CWE-266 (Incorrect Privilege Assignment). The CVSS 3.1 base score is 8.1 with network attack vector, low attack complexity, low privileges required, no user interaction, unchanged scope, and high confidentiality and integrity impacts. Red Hat's advisory confirms the issue but does not currently specify a patch or remediation level. The vulnerability affects the confidentiality and integrity of the cluster by allowing arbitrary code execution.

An authenticated user within the cluster can exploit this vulnerability to bypass security policies and execute arbitrary untrusted remote code within sidecar containers. This compromises the confidentiality and integrity of the cluster environment. Availability impact is not affected. The vulnerability allows privilege escalation within the cluster context.

As of the latest Red Hat advisory, no official patch or remediation level has been published. Patch status is not yet confirmed—check the Red Hat advisory at https://access.redhat.com/security/cve/CVE-2026-15467 for current remediation guidance. Until a fix is available, restrict authenticated user permissions within the cluster to limit the ability to configure sidecar containers. Monitor Red Hat communications for updates on patches or mitigations.