HIGHVulnerability

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.

Properties

severity
HIGH
score
7.8
cve_id
CVE-2026-90553
signal_observed_at
2026-09-21T23:07:07+00:00
vector
CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H
published_at
2026-09-12T13:16:53.887
last_modified
2026-09-16T17:31:40.200

Related Entities (3)

AFFECTS_PRODUCT (1)

[Product]

HAS_WEAKNESS (1)

[Weakness]Improper Control of Generation of Code ('Code Injection')

DESCRIBED_BY (1)

[Source]NVD

Explore deeper with Ninja Signal's threat intelligence graph

CVE-2026-90553 — Ninja Signal Threat Intelligence | Ninja Signal