CVE-2026-90553 PUBLISHED

vLLM before 0.28.0 Remote Code Execution via LlavaOnevision2 processor

Assigner: VulnCheck
Reserved: 12.09.2026 Published: 12.09.2026 Updated: 12.09.2026

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.

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
CVSS Score: 8.5

Exploitability Metrics		Vulnerable System Impact Metrics		Subsequent System Impact Metrics
Attack Vector	Local	Confidentiality	High	Confidentiality	None
Attack Complexity	Low	Integrity	High	Integrity	None
Attack Requirements	None	Availability	High	Availability	None
Privileges Required	None
User Interaction	Passive

CVSS 4.0

CVSS Vector: CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H
CVSS Score: 7.8

Attack Vector	Local	Scope	Unchanged
Attack Complexity	Low	Confidentiality Impact	High
Privileges Required	None	Integrity Impact	High
User Interaction	Required	Availability Impact	High

CVSS 3.1

Product Status

Vendor	vllm-project
Product	vLLM
Versions	Default: unaffected affected from 0 to 0.28.0 (excl.) Version 0.28.0 is unaffected

CVE-2026-90553 PUBLISHED

vLLM before 0.28.0 Remote Code Execution via LlavaOnevision2 processor

Metrics

Product Status

Credits

References

Problem Types