Executive brief
A security flaw in KubeVirt's network configuration tool allows users with virtual machine creation permissions to bypass network isolation. By providing specially crafted network names, an attacker can connect their virtual machines to private networks belonging to other users or the system. This could lead to unauthorized access to sensitive data or the ability to impersonate other services on the network.
Technical details
An improper input validation vulnerability exists in KubeVirt's network annotation generator. When a VirtualMachineInstance is created with a Multus network configuration, the 'networkName' value is written directly to the launcher pod's 'v1.multus-cni.io/default-network' annotation without sanitization. If the 'ExternalNetResourceInjection' feature gate is enabled, the system skips the NetworkAttachmentDefinition (NAD) lookup that would normally validate the name. An attacker with 'kubevirt.io:edit' permissions can inject a JSON-formatted array to specify arbitrary namespaces, NAD names, and static IP/MAC addresses, effectively attaching the pod to any network segment on the node.
Affected products
- KubeVirt KubeVirt v1.8.0+
- Red Hat OpenShift Virtualization 4.21+
Timeline
- 2026-06-26: disclosed: Initial vulnerability report and CVE assignment.
- 2026-06-26: advisory: Red Hat published security advisory.