CVE-2026-34940

OVERVIEW CVE-2026-34940 is a command injection vulnerability in KubeAI, an AI inference operator for Kubernetes. The vulnerability exists in the ollamaStartupProbeScript() function in versions prior to 0.23.2, where unsanitized model URL components are directly concatenated into shell commands executed as Kubernetes startup probes. An attacker with the ability to create or modify Model custom resources can exploit this flaw to inject and execute arbitrary shell commands within model server pods. SEVERITY This vulnerability carries a CVSS score of 8.8 (HIGH), with a network attack vector, low attack complexity, and requirements for low privileges. The impact is severe across all security dimensions: attackers can achieve high confidentiality impact through unauthorized data access, high integrity impact through command execution and system modification, and high availability impact through service disruption. The attack requires only authenticated access to the Kubernetes cluster to create or update resources, making it accessible to users with limited permissions. EXPLOITATION STATUS There is currently no evidence of active exploitation in the wild, with no confirmed entries on the Known Exploited Vulnerabilities (KEV) catalog and no activity on the Hot List. The EPSS score of 0.000140 indicates low predicted exploitation probability relative to other CVEs. However, given the straightforward nature of command injection and the requirement for only basic Kubernetes cluster access, organizations should prioritize patching to version 0.23.2 or later.