REDHAT-BUG-2529287: Medium severity Red Hat OpenShift AI (odh-dashboard) vulnerability
A flaw was found in odh-dashboard in Red Hat OpenShift AI. The backend-for-frontend route GET /api/nim-serving/:nimResource reads Kubernetes Secrets using the dashboard service account and returns the full Secret object, including .data, without an authorization check. Any authenticated dashboard user can retrieve the cluster NVIDIA NGC API key Secret (apiKeySecret) and the NIM image pull secret (nimPullSecret). Create and delete of the same NIM credential are admin-gated; the read path is not. This is missing authorization (CWE-862) and insufficiently protected credentials (CWE-522). It is distinct from CVE-2026-5483 (service-account token leak in the Kubernetes client response wrapper on the same route) and CVE-2026-16456 (odh-model-controller cross-namespace confused deputy).
Reported privately by Dmitry Maranik (Sectum AI) via PSIRTSUPT-22896. Confirmed by source review of opendatahub-io/odh-dashboard at HEAD (backend/src/routes/api/nim-serving/index.ts). Not reproduced against a live RHOAI + NGC deployment. Preconditions: authenticated dashboard access and a NIM Account CR with the referenced Secrets present.
Affected Software
Event History
Frequently Asked Questions
Who can access the exposed credentials?
Any authenticated dashboard user can use the affected read route; administrative privileges are not required for that read path. The create and delete operations for the same NIM credential are admin-gated, but the read operation is not.
What conditions must exist for exploitation?
An attacker needs authenticated access to the dashboard and a NIM Account custom resource whose referenced Secrets are present. The affected route uses the dashboard service account to read those Secrets.
Which secrets may be disclosed?
The route can return the full Kubernetes Secret object, including its .data field. This can expose the NVIDIA NGC API key Secret (apiKeySecret) and the NIM image pull Secret (nimPullSecret).
Has this been confirmed in a live deployment?
The issue was confirmed by source review of the odh-dashboard code at HEAD. It was not reproduced against a live Red Hat OpenShift AI and NGC deployment.