In multi-node deployments, an attacker who had already obtained code execution on one appliance node could submit a value through an internal cluster interface that was written into monitoring configuration on another node without sufficient validation, potentially allowing OS commands to be executed there. Execution was limited to an unprivileged service account on that node.