Detections
Analytics

Chainguard: vllm-omni-cuda-13.0: security update to 0.32.0-r0

medium Tenable Cloud Security Plugin ID 477952

Description

There are packages installed that are affected by a vulnerability referenced in the following CVE:

 - vLLM is an inference and serving engine for large language models. Prior to 0.32.0, the Phi-4 multimodal
 audio encoder in phi4mm_audio.py and phi4mm_utils.py uses eager attention by default, allowing a remote
 client to submit one legal long audio clip that causes MultiHeadedAttention.forward_attention to
 materialize and retain full attention matrices in self.attn. The scaled dot-product attention branch does
 not materialize these matrices. The resulting quadratic GPU-memory use can cause an out-of-memory
 condition and terminate EngineCore for all users. This issue is fixed in version 0.32.0. (CVE-2026-115913)

Solution

Update the vllm-omni-cuda-13.0 library and its related packages to version 0.32.0-r0 or later.

Plugin Details

Severity: Medium

ID: 477952

Version: Revision 1.1

Type: Local

Published: 10/30/2026

Updated: 10/30/2026

Risk Information

VPR

Risk Factor: Low

Score: 3

Percentile: 23.72

CVSS v2

Risk Factor: Medium

Base Score: 6.8

Temporal Score: 5.3

Vector: CVSS2#AV:N/AC:L/Au:S/C:N/I:N/A:C

CVSS Score Source: CVE-2026-115913

CVSS v3

Risk Factor: Medium

Base Score: 6.5

Temporal Score: 5.9

Vector: CVSS:3.0/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H

Temporal Vector: CVSS:3.0/E:P/RL:O/RC:C

Vulnerability Information

Exploit Available: true

Exploit Ease: Exploits are available

Vulnerability Publication Date: 10/26/2026

Reference Information

CVE: CVE-2026-115913