The vulnerability in Data Science Pipelines involves an improper authorization flaw in the setDefaultServiceAccount function. By exploiting this, a restricted user or tenant can specify a more privileged ServiceAccount during a CreateRun request, bypassing authorization checks. This enables the attacker to run containers with elevated privileges, which may lead to the disclosure of secrets and execution of commands in other users' pods. The CVSS 3.1 vector indicates network attack vector, low attack complexity, low privileges required, no user interaction, unchanged scope, high confidentiality impact, low integrity impact, and no availability impact. The vulnerability is tracked as CVE-2026-18620 with CWE-639 (Authorization Bypass Through User-Controlled Key). Vendor advisories from Red Hat OpenShift AI mention updated images and instructions for upgrading clusters to address this issue.

Exploitation of this vulnerability allows a restricted user or tenant to bypass authorization controls and run containers with elevated privileges. This can lead to the disclosure of sensitive information such as secrets and unauthorized execution of commands within other users' pods. The impact on confidentiality is high, while integrity impact is low and availability is not affected.

Red Hat has published security advisories (RHSA-2026:53261 and RHSA-2026:53262) for Red Hat OpenShift AI that include updated images and instructions for upgrading clusters to remediate this vulnerability. Users should follow the official Red Hat documentation to upgrade to the fixed versions of Red Hat OpenShift AI as soon as possible. No other specific mitigations are indicated. Patch status is confirmed by vendor advisories indicating updated images are available.