SUSE Rancher Fleet EUVD-2026-115251

Requires an authenticated Rancher account holding only the default 'user' global role with the intentional, default self-service ability to create Fleet workspaces (verb 'create' on management.cattle.io fleetworkspaces) - no non-default feature or configuration needs to be enabled, since Fleet is enabled by default in Rancher. The attacker must be able to define a Fleet Cluster object in a namespace they control that references a kubeconfig secret carrying management-cluster administrative credentials. Exploitation is LIMITED or prevented where administrators had already removed Fleet workspace self-service via custom global roles, restricted write access to fleet.cattle.io clusters/bundles in non-system namespaces, or deploy admission policy (e.g. Kubewarden) rejecting Cluster objects that reference kubeconfig secrets in other namespaces; these reduce but, per the vendor, do not fully eliminate exposure on unpatched versions. Note the vendor states upgrading is the only complete fix.

This is a genuine high-priority issue, not an inflated CVSS score. The multi-source signals align toward urgency: the vuln.today Decision label is act_now, the CVSS base score is 9.1 with a scope change (S:C), and the impact is full cluster-admin on the Rancher management cluster, which the advisory confirms exposes all management-cluster secrets, downstream-cluster credentials, API tokens, and RBAC. The most important caveat is a conflict in the privileges metric: the supplied CVSS vector states PR:H, but the SUSE/Rancher advisory explicitly describes an authenticated user holding only the default 'user' global role with no role bindings on the local cluster - i.e. a low-privileged, self-service-capable account. That makes the real precondition closer to PR:L than PR:H, which arguably understates practical exposure rather than overstating it. No EPSS score and no KEV status were provided in the input; absence of those is not evidence of safety, but it means there is currently no confirmation of active exploitation or a quantified exploitation probability. No working public exploit is identified in the input (the only reference is the GHSA advisory itself, not exploit code). Overall: real priority driven by default-enabled, low-privilege-reachable total compromise of the management plane and the vendor's act_now signal, tempered by the fact that exploitation status and EPSS are unknown from the provided data.

An authenticated Rancher user with only the default 'user' global role creates a Fleet workspace (self-service, default behavior) and defines a Fleet Cluster object in a namespace they control that references a kubeconfig secret carrying administrative credentials for the management (local) cluster. The Fleet import controller, running on the management cluster with cluster-admin, then manages the management cluster on the user's behalf and deploys the user's Fleet content there with elevated privileges, yielding full cluster-admin. No public working exploit is identified in the available sources; the trigger mechanism is described by the vendor advisory rather than demonstrated by released exploit code.