Description

FakeFish handles incoming credentials by passing them down to scripts. This works for real hardware because in the end it's up to the BMC to validate them. However, KubeVirt relies on a KUBECONFIG file mounted to the container and completely ignores the credentials. This allows any user of the cluster to control VMs of the user that created fakefish, power them on and off, and mount arbitrary CD images to them.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "9.3",
  "pubDate": "2026-08-17T15:16:57.610Z",
  "pubdate": "2026-08-17T15:16:57.610Z",
  "executiveSummary": "An authorization bypass vulnerability exists in FakeFish regarding its credential handling mechanism when deployed within KubeVirt environments. The vulnerability type is improper access control and authentication delegation. The impact allows unauthorized multi-tenant cluster users to gain full control over virtual machines belonging to the user who instantiated FakeFish. Affected systems include FakeFish deployments utilizing KubeVirt where a KUBECONFIG file is mounted to the container. The risk implications are severe, exposing sensitive tenant workloads to cross-tenant administrative control without requiring prior credentials or elevated privileges within the target context. Attacker capabilities include powering virtual machines on and off, as well as mounting arbitrary CD images to victim VMs. Exploitation requirements involve an attacker having access to the Kubernetes cluster where FakeFish is deployed, allowing them to leverage the statically mounted KUBECONFIG file which bypasses credential validation entirely.",
  "technicalDetails": "The vulnerability resides in the architectural assumption made by FakeFish during the credential delegation process. The vulnerable component is the credential handling logic within FakeFish that normally passes incoming credentials down to underlying scripts, relying entirely on the target Baseboard Management Controller (BMC) to perform strict validation for real hardware deployments. However, when deployed in KubeVirt environments, the application model changes fundamentally. Instead of utilizing dynamic credentials provided per request, KubeVirt relies on a statically mounted KUBECONFIG file present within the container filesystem. This architectural mismatch causes FakeFish to completely ignore incoming user-supplied credentials. Consequently, any request processed by FakeFish authenticates against the Kubernetes API using the broad permissions granted by the mounted KUBECONFIG file rather than enforcing tenant-isolation boundaries based on the submitting user's identity. The step-by-step attack flow begins when an unprivileged or standard user within the Kubernetes cluster interacts with the FakeFish instance. Because the application fails to validate or map the incoming credentials to specific tenant contexts, the interaction is processed using the authority of the container's service account or mounted cluster configuration. The attacker can then issue arbitrary control commands, bypassing any intended authorization checks. The post-exploitation impact includes complete administrative control over the targeted virtual machines, enabling malicious actors to manipulate power states and attach arbitrary CD images containing potentially malicious payloads or extracting sensitive data from the guest operating systems."
}