CVE-2026-5483: Red Hat OpenShift AI odh-dashboard: Kubernetes service account token exposure
According to the record, the odh-dashboard component of Red Hat OpenShift AI can disclose Kubernetes Service Account tokens through a NodeJS endpoint. NVD rates the issue 9.9 (critical) and Red Hat rates it 8.5, and no confirmed exploitation appears in the sources consulted. The affected ranges given are from 2.16 before 2.16.4, from 2.25 before 2.25.4, and 3.2 and 3.3.
Overview
Red Hat has published CVE-2026-5483 for the odh-dashboard component of Red Hat OpenShift AI (RHOAI). The record states that a NodeJS endpoint in this component can disclose Kubernetes Service Account tokens. The weakness is classified as CWE-201, insertion of sensitive information into sent data.
Impact
- Exposed service account tokens could allow an unauthorized party to reach Kubernetes resources with the privileges of that account.
- NVD scores the issue 9.9 under CVSS, which is critical. Red Hat scores it 8.5.
- The record does not say which Kubernetes resources are reachable. This depends on the privileges granted to the exposed account, so it remains unknown from the available information.
Affected products and versions
- Red Hat OpenShift AI from 2.16 before 2.16.4.
- Red Hat OpenShift AI from 2.25 before 2.25.4.
- Red Hat OpenShift AI 3.2 and 3.3, with no narrower range given in the record.
- The record lists the product lines 2.16, 2.25, 3.2 and 3.3 without further version detail for each. The record contains no fixed-version field.
Exploitation status
The record does not list CVE-2026-5483 in the CISA Known Exploited Vulnerabilities catalog. The SSVC data in the record shows exploitation as none and automatable as no. No confirmed exploitation appears in the sources consulted.
Remediation and mitigation
- Review the Red Hat errata linked for this CVE (RHSA-2026:7397, RHSA-2026:7398, RHSA-2026:7403 and RHSA-2026:7404) and apply the one that matches your OpenShift AI release. The record does not map each erratum to a specific product line.
- Treat the version boundaries as a guide: the stated ranges suggest that 2.16.4 and 2.25.4 are the first versions outside them, which is an inference to confirm with the vendor.
- If exposure is suspected, rotate the affected service account tokens and review which permissions those accounts hold in the cluster.
- Limit access to the dashboard to trusted networks and users while patching is in progress.
Detection guidance
- Review dashboard and API server audit logs for unexpected use of service account identities, especially from sources outside normal administrative paths.
- Monitor for service accounts performing actions they do not normally perform.
- Confirm that service account permissions follow least privilege, since this limits what a leaked token can do.
Limitations of the available information
The record gives no fixed-version field, no detail on the specific endpoint, and no information about the permissions of the tokens involved. The NVD and Red Hat scores differ, and the record does not explain the difference. The sources consulted do not confirm exploitation.
Frequently asked questions
Is CVE-2026-5483 being actively exploited?
The record does not list it in the CISA KEV catalog, and no confirmed exploitation appears in the sources consulted.
Which Red Hat OpenShift AI versions are affected?
According to the record, the affected ranges are from 2.16 before 2.16.4, from 2.25 before 2.25.4, and 3.2 and 3.3.
What should administrators do first?
Review the Red Hat errata linked for this CVE, apply the one matching your release, and rotate service account tokens if exposure is suspected.
This analysis was drafted with AI assistance from verified source data and automatically checked against it.