The vulnerability exists in the odh-model-controller's loadSecret function, where it fails to validate the namespace parameter supplied by an authenticated user with custom resource creation permissions. This improper validation allows the attacker to read Secrets from namespaces they should not have access to, resulting in information disclosure of sensitive credentials. The CVSS v3.1 score is 6.5 (medium), reflecting network attack vector, low attack complexity, low privileges required, no user interaction, unchanged scope, high confidentiality impact, and no integrity or availability impact. The flaw is tracked as CWE-441 (Improper Restriction of a Resource to a Legitimate User).

An attacker with authenticated access and permission to create custom resources can read sensitive API keys and cloud credentials from other namespaces, leading to information disclosure. This could compromise confidentiality of critical secrets used by the system.

Red Hat has published security advisories RHSA-2026:53261 and RHSA-2026:53262 addressing this vulnerability as part of updates to Red Hat OpenShift AI. Updated images and instructions for upgrading to fixed versions (e.g., RHOAI 2.25.10 and 3.4.3) are available from Red Hat. Users should apply these official updates promptly to remediate the issue. Patch status is confirmed as fixed in these updated releases. No additional mitigations are indicated by the vendor advisory.