CVE-2026-33866

[CVE catalog]

Severity

Medium · CVSS 6.5CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N

Published

Apr 7, 2026, 12:57:44 PM UTC

Source

[NVD] · [CVE record]

EPSS

0.36% · Oct 25, 2026

A flaw was found in MLflow. A user without proper permissions to an experiment can exploit an authorization bypass in the AJAX endpoint, which is used for downloading saved model artifacts. This missing access-control validation allows the unauthorized user to directly query the endpoint and retrieve model artifacts they are not authorized to access, leading to information disclosure. MLflow: MLflow: Information disclosure via authorization bypass in AJAX endpoint This Moderate-severity information disclosure flaw in MLflow, as deployed in Red Hat OpenShift AI, allows authenticated users to bypass authorization checks. An attacker with existing access to the platform, but without explicit permissions to a specific experiment, could directly query the AJAX endpoint to retrieve model artifacts, leading to unauthorized data access. The CVSS score(s) listed for this vulnerability do not reflect the associated product's status, and are included for informational purposes to better understand the severity of this vulnerability.