Breaking: Ruby2Shell - Ruby Memory Corruption Flaws Expose Applications to Remote Code Execution
Read the Report
Learn How MCP Is Bypassing a Decade of Cloud Security Best Practices
Read the Report
Are you exposed? Connect a repo in minutes.
Scan your repos free

CVE-2025-46722

vllm: vLLM has a Weakness in MultiModalHasher Image Hashing Implementation

Overview

Published
29 May 2025
Modified
17 Jun 2026
High CVE
1 Registry packages
0 OS packages
0 Linux distributions
Yes Fix available

Description

vLLM is an inference and serving engine for large language models (LLMs). In versions starting from 0.7.0 to before 0.9.0, in the file vllm/multimodal/hasher.py, the MultiModalHasher class has a security and data integrity issue in its image hashing method. Currently, it serializes PIL.Image.Image objects using only obj.tobytes(), which returns only the raw pixel data, without including metadata such as the image’s shape (width, height, mode). As a result, two images of different sizes (e.g., 30x100 and 100x30) with the same pixel byte sequence could generate the same hash value. This may lead to hash collisions, incorrect cache hits, and even data leakage or security risks. This issue has been patched in version 0.9.0.

Source: OX vulnerability database

Severity

7.3 · HIGH

CVSS 3.1 base score, scored by NVD.

Attack vector
Network
Attack complexity
Low
Privileges required
None
User interaction
None
Scope
Unchanged
Confidentiality
Low
Integrity
Low
Availability
Low

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L

Weakness: CWE-1023, CWE-1288

Affected packages · 1

Packages affected by CVE-2025-46722
EcosystemPackageVulnerableFixed in
PyPI vllm >=0.7.0, <0.9.0 0.9.0

Open a package to see all its vulnerabilities.

Operating systems

No Linux distribution reports this ID.

References

OX Radar and Response Center

Find threats in your stack.

OX Radar can check high-priority threats, like critical package CVEs, against your context.

Each incident is marked Detected or Not Detected with a remediation plan and proof all tracked in the Response Center.