ai_security▌

← feed

NOTABLE

vLLM publishes six GHSAs: multimodal-preprocessing DoS batch, engine crash, and a heap-address leak

Six vLLM advisories with CVEs (CVE-2026-126394 through 126399) were published in one NVD batch: request-controlled mm_processor_kwargs.max_patches forces quadratic CPU work on Cohere2 Vision preprocessing (126397, high), num_frames and min_pixels values bypass media-processing budgets to OOM or exhaust the API server (126394, 126396, 126399), embeddings-only /v1/completions requests trip an assertion that kills the engine on XD-RoPE models with --enable-prompt-embeds (126398, high, fixed in 0.30.0), and a lowercase-only sanitize_message() regex leaks live server heap addresses through uppercase PIL image addresses in crafted PNGs, an ASLR-bypass building block (126395). Five are fixed in 0.31.0.

released 2026-11-13 · ingested 2026-11-14T00:45Z

Sources

Primary source · corroboration · corroboration

Cite this

APA

AI Security Index. (2026). vLLM publishes six GHSAs: multimodal-preprocessing DoS batch, engine crash, and a heap-address leak. AI Security Index. https://github.com/advisories/GHSA-8hqr-48h6-r93j

BibTeX

@misc{aisec-2026-11-13-vllm-six-ghsa-multimodal-dos-heap-leak,
  title  = {vLLM publishes six GHSAs: multimodal-preprocessing DoS batch, engine crash, and a heap-address leak},
  year   = {2026},
  howpublished = {\url{https://github.com/advisories/GHSA-8hqr-48h6-r93j}},
  note   = {AI Security Index item 2026-11-13-vllm-six-ghsa-multimodal-dos-heap-leak}
}