CVE-2026-34755

MEDIUM6.5EPSS 0.05%

vLLM: Denial of Service via Unbounded Frame Count in video/jpeg Base64 Processing

發布日:2026/4/3修改日:2026/5/20

描述

vLLM is an inference and serving engine for large language models (LLMs). From 0.7.0 to before 0.19.0, the VideoMediaIO.load_base64() method at vllm/multimodal/media/video.py splits video/jpeg data URLs by comma to extract individual JPEG frames, but does not enforce a frame count limit. The num_frames parameter (default: 32), which is enforced by the load_bytes() code path, is completely bypassed in the video/jpeg base64 path. An attacker can send a single API request containing thousands of comma-separated base64-encoded JPEG frames, causing the server to decode all frames into memory and crash with OOM. This vulnerability is fixed in 0.19.0.

受影響套件(2)

CVSS 分數

來源版本嚴重程度向量
osvCVSS 3.1MEDIUM6.5CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H

參考連結(5)