Servers and applications
CVE-2026-90553
LlavaOnevision2 processor loader executes attacker model code with `trust_remote_code=False` (inert `trust_remote_code` kwarg to `transformers.get_class_from_dynamic_module`) — RCE from a malicious model
- Severity
- High 8.5
- Published
- 28 Aug 2026
- Last changed
- 16 Sept 2026
- Added to UpHeron
- 3 Nov 2026
- Fixed in
- 0.28.0
- Exploit chance
- 0.3% in the next 30 daysHigher than 24% of CVEs. EPSS by FIRST, 19 Nov 2026
- Weakness
- Code injection (CWE-94)CWE-829
- Attack
- Needs local access · No login needed · Needs someone to view a page or file
- Attacker can
- Read data · Change data · Take it down
- 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/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X- Also known as
- GHSA-3c86-2m5g-59q7
- Sources
- NVD, Project advisories (GitHub)
Description
LlavaOnevision2's processor loader _load_ov2_processor loads the model's remote processor classes via transformers.dynamic_module_utils.get_class_from_dynamic_module(...), passing trust_remote_code=trust_remote_code. That transformers function has no trust_remote_code parameter — the value is swallowed into kwargs and ignored — so the attacker-supplied processing_llava_onevision2.py / video_processing_llava_onevision2.py module is imported and its top-level code executes even when the user set trust_remote_code=False. Because LlavaOnevision2ForConditionalGeneration is a built-in** vLLM architecture, the config loads without any remote code, so nothing upstream refuses the load — the processor loader is the unguarded path. Loading/serving a crafted LlavaOnevision2 model with the safety opt-out enabled runs arbitrary code with the vLLM process's authority.
vllm/model_executor/models/llava_onevision2.py:116-142:
Affected software
| Software | Affected versions |
|---|---|
vLLMServers and applications · cpe:a:vllm:vllm |
|
vLLMServers and applications · github:vllm-project/vllm |
|
About vLLM
open-source software for large language model inference
- Docker Hub pulls
- 44,195,225Docker Hub, 15 Nov 2026
- GitHub stars
- 96,943GitHub, 15 Nov 2026
- Forks
- 24,836GitHub, 15 Nov 2026
- Latest release
- v0.33.0rc3GitHub, 15 Nov 2026
- Released
- 2 Nov 2026GitHub
- License
- Apache Software License 2.0Wikidata
- First published on GitHub
- 9 Feb 2023GitHub
Website Source code Docker image
Vulnerabilities in vLLM
- In our list
- 169
- Published in the last 12 months
- 150
- Critical, last 12 months
- 6
- High, last 12 months
- 63
- Known exploited
- 0
- CVE-2026-130686vLLM 0.6.0 before 0.33.0 Resource Exhaustion via Multi-Image RequestsHigh 8.7 18 Nov 2026
- CVE-2026-130684vLLM before 0.33.0 Path Traversal via speculators_config verifier.name_or_pathMedium 5.3 18 Nov 2026
- CVE-2026-130685vLLM 0.28.0 before 0.33.0 Unauthenticated DoS via Out-of-Range Prompt Token IDHigh 8.7 18 Nov 2026
- GHSA-xgh8-qwjh-q853/v1/systemone bypasses the --trust-request-chat-template gateMedium 6.5 18 Nov 2026
- CVE-2026-129356vllm-project vllm Common Prefix single_type_kv_cache_manager.py race conditionLow 2.3 15 Nov 2026
- CVE-2026-126394vLLM 0.30.0 before 0.31.0 Resource Exhaustion via OpenPangu Video Backend num_framesMedium 6.9 13 Nov 2026
References
- NVD https://nvd.nist.gov/vuln/detail/CVE-2026-90553
- CVE.org https://www.cve.org/CVERecord?id=CVE-2026-90553
- GitHub Advisory Database https://github.com/advisories/GHSA-3c86-2m5g-59q7
- github.com https://github.com/vllm-project/vllm/security/advisories/GHSA-3c86-2m5g-59q7