HighVulnerabilityLLM-specific
CVE-2026-68903: LLMKube Model source paths allow reading controller and node files
- Source
- NVD(opens in a new tab)
- Identifier
- CVE-2026-68903
- Published
- Record updated
- Affected
- github.com/defilantech/llmkube <= 0.8.28
- Fixed in
- 0.9.0
- released July 5, 2026, 136 days before disclosure
- Known exploitation
- Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
Summary
LLMKube, a Kubernetes operator for self-hosted LLM inference, prior to 0.9.0 has a flaw in its ModelReconciler, which accepts arbitrary absolute paths and file:// URIs in Model.spec.source without restricting them to approved roots. A namespace user who can create or update Model and InferenceService resources can make copyLocalModel read an attacker-selected file in the controller pod, or mount an attacker-selected node file through a HostPathVolumeSource when modelCache.enabled is set, which is the default. The copied file can expose controller service-account credentials or other node secrets outside the attacker's namespace.
Mitigation
Fixed in 0.9.0.
Related items
- HighCVE-2026-77283: obsidian-sync-mcp HTTP /mcp endpoint DNS rebinding and cross-origin vault accessSimilar attack · NVD/CVE Database
- HighCVE-2026-75612: yt-mcp arbitrary file read through add_attachment file_pathSimilar attack · NVD/CVE Database
- MediumCVE-2026-75518: DOMShell MCP server eval tool bypasses write permission checksSimilar attack · NVD/CVE Database
- HighCVE-2026-62964: Gortex arbitrary file write through MCP file toolsSimilar attack · NVD/CVE Database
- HighCVE-2026-127222: SGLang missing authorization on POST /load_lora_adapter_from_tensorsSimilar attack · NVD/CVE Database