CVE-2026-15581 describes an authentication bypass vulnerability in the TrustyAI Service (TAS) deployment that permits any pod on the cluster network to directly access the TAS backend API without proper authentication. Exploitation of this flaw can lead to unauthorized reading, tampering, or deletion of monitoring data and configurations, as well as injection of arbitrary data into the service. The vulnerability is associated with CWE-306 (Missing Authentication for Critical Function). Red Hat's advisories for Red Hat OpenShift AI mention this CVE among others and indicate updated images are available for certain versions, but no explicit fix for this CVE is currently provided. Users are directed to follow upcoming documentation for upgrade and mitigation steps.

Successful exploitation allows an attacker with access to any pod in the cluster network to bypass authentication controls on the TAS backend API. This can result in full compromise of monitoring data integrity and availability, including reading sensitive data, modifying or deleting configurations, and injecting malicious data. Such actions could disrupt tenant operations relying on TAS, potentially causing denial of service or data corruption.

Red Hat has published advisories acknowledging the vulnerability and has released updated images for Red Hat OpenShift AI versions 2.25.10 and 3.4.3. However, these advisories indicate that no direct fixes for this CVE are included in these releases yet. Users should monitor Red Hat's official documentation and errata for updated instructions on upgrading their clusters and applying forthcoming patches. Until a fix is available, restricting pod network access to the TAS backend API and implementing strict network segmentation may reduce exposure.