Description
vLLM is an inference and serving engine for large language models (LLMs). From 0.6.1 to before 0.20.0, there is a a Token Injection vulnerability in vLLM’s multimodal processing. Unauthenticated, text-only prompts that spell special tokens are interpreted as control. Image and video placeholder sequences supplied without matching data cause vLLM to index into empty grids during input-position computation, raising an unhandled IndexError and terminating the worker or degrading availability. Multimodal paths that rely on image_grid_thw/video_grid_thw are affected. This vulnerability is fixed in 0.20.0.
Affected products
vllm: vllm
References2
- github.com/vllm-project/vllm/security/advisories/GHSA-hpv8-x276… ExploitVendor Advisory
- github.com/vllm-project/vllm/issues/32656 Issue Tracking
Metrics
- CVSS 3.1
-
7.5
HIGH
Score source: NVD
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H - EPSS — exploit probability
- 0% percentile: 36 · as of 14 November 2026
- CWE
- CWE-129
Could this be exploited in your environment?
A patch closes the known hole. A penetration test shows what an attacker can still reach — with proof.