HighVulnerability
CVE-2026-4944: vllm-project/vllm remote code execution via trust_remote_code
- Source
- NVD(opens in a new tab)
- Identifier
- CVE-2026-4944
- Published
- Record updated
- Known exploitation
- Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
- EPSS
- 1.0%
Summary
vllm-project/vllm version 0.14.1 hardcodes `trust_remote_code=True` in two model implementation files, `vllm/model_executor/models/nemotron_vl.py` and `vllm/model_executor/models/kimi_k25.py`. This overrides a user's explicit `--trust-remote-code=False` setting, allowing remote code execution through malicious HuggingFace model repositories. The issue is an incomplete fix for CVE-2025-66448 and CVE-2026-22807 and is particularly impactful for deployments loading NemotronVL or KimiK25 models.
Mitigation
The source does not state a fix yet. Check the original advisory for updates.
Topics
Related items
- MediumCVE-2026-100653: vLLM unpinned Hugging Face artifact loads for FunAudioChat and Tarsier2Same vendor · NVD/CVE Database
- HighGHSA-3hmm-rh5q-gwwr: LMDeploy vulnerable to arbitrary code execution via eval() of untrusted quant_dtype in model config loadingSame vendor · Hugging Face Security Advisories
- MediumCVE-2026-80047: Hugging Face Transformers writes remote Python files to disk before trust checkSame vendor · NVD/CVE Database
- HighCVE-2026-58474: whichllm code injection in run and snippet commands via GGUF filenamesSame vendor · NVD/CVE Database
- HighCVE-2026-79784: Vocos class instantiation from configuration via from_pretrainedSame vendor · NVD/CVE Database