Логотип exploitDog
Консоль
Логотип exploitDog

exploitDog

redhat логотип

CVE-2026-12491

Опубликовано: 10 июн. 2026
Источник: redhat
CVSS3: 4.8
EPSS Низкий

Описание

A flaw was found in vLLM, an open-source library for large language model inference. This vulnerability arises from improper handling of image metadata, specifically EXIF orientation and PNG transparency (tRNS) data, during image processing. When images are converted to RGB, transparency information may be implicitly discarded or remapped, leading to unexpected rendering of transparent pixels and distortion of input content. This can result in the model misinterpreting image content, potentially affecting the integrity of processed data.

Отчет

This Moderate flaw in vLLM, as used in Red Hat AI Inference Server, Red Hat OpenShift AI, and Red Hat Enterprise Linux AI, stems from incorrect image metadata handling during processing. Specifically, EXIF orientation and PNG transparency data are not normalized, causing large language models to misinterpret image content. This can lead to a loss of data integrity in processed inputs, as the model's understanding of an image may differ from its intended representation.

Затронутые пакеты

ПлатформаПакетСостояниеРекомендацияРелиз
Red Hat AI Inference Serverrhaiis/vllm-cpu-rhel9Fix deferred
Red Hat AI Inference Serverrhaiis/vllm-cuda-rhel9Fix deferred
Red Hat AI Inference Serverrhaiis/vllm-neuron-rhel9Fix deferred
Red Hat AI Inference Serverrhaiis/vllm-rocm-rhel9Fix deferred
Red Hat AI Inference Serverrhaiis/vllm-spyre-rhel9Fix deferred
Red Hat AI Inference Serverrhaiis/vllm-tpu-rhel9Fix deferred
Red Hat AI Inference Serverrhaii/vllm-cpu-rhel9Fix deferred
Red Hat AI Inference Serverrhaii/vllm-cuda-rhel9Fix deferred
Red Hat AI Inference Serverrhaii/vllm-gaudi-rhel9Fix deferred
Red Hat AI Inference Serverrhaii/vllm-neuron-rhel9Fix deferred

Показывать по

Дополнительная информация

Статус:

Moderate
Дефект:
CWE-115
https://bugzilla.redhat.com/show_bug.cgi?id=2489786vllm: vllm: image EXIF Rotation & PNG tRNS Transparency Not Normalized, Causing Mismatch Between Model Input and Expectations

EPSS

Процентиль: 15%
0.00239
Низкий

4.8 Medium

CVSS3

Связанные уязвимости

CVSS3: 4.8
nvd
2 месяца назад

A flaw was found in vLLM, an open-source library for large language model inference. This vulnerability arises from improper handling of image metadata, specifically EXIF orientation and PNG transparency (tRNS) data, during image processing. When images are converted to RGB, transparency information may be implicitly discarded or remapped, leading to unexpected rendering of transparent pixels and distortion of input content. This can result in the model misinterpreting image content, potentially affecting the integrity of processed data.

CVSS3: 4.8
debian
2 месяца назад

A flaw was found in vLLM, an open-source library for large language mo ...

CVSS3: 4.8
github
2 месяца назад

vLLM: image EXIF Rotation & PNG tRNS Transparency Not Normalized, Causing Mismatch Between Model Input and Expectations

EPSS

Процентиль: 15%
0.00239
Низкий

4.8 Medium

CVSS3