NeuVector CVE-2025-54469

Exploitation requires the attacker to control the values of the CLUSTER_RPC_PORT or CLUSTER_LAN_PORT environment variables of the NeuVector enforcer container - i.e., the ability to define or modify the enforcer's container/pod environment (Kubernetes deployment or orchestration-level access). The injected command only runs when the enforcer container stops and the monitor process invokes popen() to check whether the consul subprocess ports are still active; it does not fire during normal steady-state operation. Limiting factors: this is not reachable over the network without first altering the container environment, it depends on the stop/teardown code path being triggered, and only unpatched versions (<5.3.5 in the 5.3.x line, <=5.4.6 in the 5.4.x line) are affected. Patched builds validate the variables as numeric ports and terminate on failure.

The input assigns CVSS 9.9 (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H), but this score conflicts with the described mechanism and should be treated with caution. The vulnerability (CWE-78 command injection) is only reachable because the enforcer's monitor process passes the CLUSTER_RPC_PORT and CLUSTER_LAN_PORT environment variables into a popen() shell command without sanitization. Setting or tampering with those environment variables is not a network operation - it requires the ability to define the enforcer container's environment, which in practice means control over the Kubernetes pod/deployment spec or equivalent orchestration privilege. The AV:N/PR:L vector therefore appears overstated: realistic exploitation is local/configuration-bound and requires meaningful prior privilege, so the 9.9 'critical' rating likely overstates real-world risk (CVSS indicates network vector but the description implies control of the container environment - verify with vendor). No EPSS score, no CISA KEV listing, and no public POC were provided, so there is no evidence of active or opportunistic exploitation. Net assessment: a genuine command-injection defect worth patching, but functionally a privilege-dependent / defense-in-depth issue rather than a remotely-exploitable crisis; prioritize patching but below internet-facing unauthenticated RCEs.

An attacker who can influence the enforcer container's environment (e.g., a tenant or compromised account able to modify the NeuVector enforcer pod spec/deployment) sets CLUSTER_RPC_PORT or CLUSTER_LAN_PORT to a value containing shell metacharacters and injected commands. When the enforcer container stops, the monitor process invokes popen() to check the consul subprocess ports, executing the attacker's injected command inside the container. No public exploit code was identified in the provided data.