CVE-2026-90553
vLLM before 0.28.0 contains a remote code execution vulnerability in the LlavaOnevision2 processor loader that ignores the trust_remote_code parameter when loading remote processor classes. Attackers can craft a malicious model with arbitrary code in processing_llava_onevision2.py that executes with vLLM process authority even when trust_remote_code is set to False.
Scoring
- Severity
- HIGH
- CVSS base score
- 8.5
- CVSS vector
- CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N
- EPSS probability
- 0.21%
- CWE
- CWE-94
- Published
- 2026-09-12
- Last modified
- 2026-09-14
Affected products
- vllm-project vLLM
- vllm-project vLLM
Weakness type
Related vulnerabilities
- CVE-2026-60004 — Gitea before 1.27.1 allows remote code execution via the diffpatch API through Git hook installation.
- CVE-2026-76605 — Joomla Extension - fabrikar.com - Remote code execution via image element in Fabrik < 4.7.2
- CVE-2026-76604 — Joomla Extension - fabrikar.com - Unauthenticated remote code execution via PHP form element in Fabrik < 4.7.2
- CVE-2026-92937 — vm2 3.11.6 Remote Code Execution via Promise call/apply
- CVE-2026-85978 — Unauthenticated Remote Code Execution in Akana API Platform
- CVE-2026-62104 — WordPress Migratico Lite plugin <= 2.6.8 - Remote Code Execution (RCE) vulnerability
- CVE-2026-53710 — MCP Context Forge: RestrictedPython sandbox bypass via getattr builtin in python_sandbox_server
- CVE-2026-55565 — Yamcs: Authenticated remote code execution via unescaped StreamSQL `LIKE` pattern compiled by Janino (`LikeExpression`)