Disclosed weakness · no CVE recorded
GPUStack could disclose model and user activity through cluster dashboards
GPUStack · GPUStack
First documented disclosure · Review checkpoint
Summary
A cluster member could see model names and usage statistics beyond the activity they were allowed to view.
Severity and exploitation evidence
Medium · 4.3/10 · CVSS 3.1. Severity source.
Exploitation not established: The original source establishes the weakness but does not report malicious exploitation.
Technical severity is not a measure of observed attacks or of risk to your deployment. Attempts, demonstrations, and successful compromise retain the qualifications in the source.
How the weakness works
Access to a cluster was treated as permission to view usage data intended for narrower roles.
Conditions required
An authenticated account can access at least one cluster.
Potential impact
Model names, account activity and token-usage statistics could be disclosed. This does not demonstrate disclosure of prompt content.
Affected versions and configurations
>= 2.2.0, <= 2.3.0
Fix and mitigation status
2.3.1 fixes this issue.
Practical next steps
Upgrade to 2.3.1 or a later release containing this fix. Review dashboard permissions and restrict cluster usage views until patched.
Operating lesson
Test aggregate statistics against the same access policy as individual records.
Evidence and limits
The maintainer assigns the score. No CVE was listed in the original advisory at review.
Sources
- GPUStack security advisory GHSA-h4h6-j8jq-7mxm
Primary source ·
- GPUStack 2.3.1 release notes
Primary source ·