Skip to content
OpenAdvisories
Medium WID-SEC-2026-4195 5 CVEs

vllm: Mehrere Schwachstellen

Published Oct 27, 2026 · Updated Nov 12, 2026 · Revision 2 · de

Summary

Ein Angreifer kann mehrere Schwachstellen in vllm ausnutzen, um einen Denial of Service Angriff durchzuführen und um Informationen offenzulegen.

Product

Open Source vLLM ist eine Open-Source-Bibliothek für schnelle und effiziente Inferenz von Large Language Models (LLMs).

Vulnerabilities 5

  • High 8.7

    vLLM from 0.10.1 before 0.31.0 forwards a request-controlled mm_processor_kwargs.max_patches value to the Hugging Face …

  • High 8.2

    vLLM from 0.28.0 before 0.30.0 contains a reachable assertion vulnerability in GPUModelRunner._init_xdrope_positions th…

  • Medium 6.9

    vLLM 0.30.0 before 0.31.0 contains a resource exhaustion vulnerability in OpenCVDynamicOpenPanguVideoBackend, which use…

  • Medium 6.5

    vLLM 0.19.0 before 0.31.0 Resource Exhaustion via Request-Level num_frames

    vLLM from 0.19.0 before 0.31.0 contains a resource exhaustion vulnerability that allows inference clients to bypass the server frame limit by setting media_io_kwargs.video.num_frames in chat completion requests. Attackers can supply num_frames=-1 or a large v…

  • Medium 5.3

    vLLM 0.14.0rc1 before 0.31.0 Heap Address Disclosure via sanitize_message()

    vLLM from 0.14.0rc1 before 0.31.0 contains an information disclosure vulnerability in sanitize_message() whose lowercase-only regex fails to scrub uppercase PIL image addresses. Unauthenticated attackers can send a crafted PNG to endpoints such as /tokenize, …

Revision history

  1. #2

    Referenz(en) aufgenommen: CVE-2026-126395, EUVD-2026-125170, CVE-2026-126397, EUVD-2026-125168, CVE-2026-126399, EUVD-2026-125166, CVE-2026-126398, EUVD-2026-125167, CVE-2026-126394, EUVD-2026-125171

  2. #1

    Initiale Fassung

References