Signal is active with 1 mentions in latest observed window
Recommended action window: Monitor and triage in normal cycle
NVD 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.
Priority
LOW
Exploitation
NONE
PoC
YES
Patch
NONE
Momentum
NONE
If you run products in this scope, you should treat this CVE as relevant to your environment.
CVE-2026-44222 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… https://www.cve.org/CVERecord?id=CVE-2026-44222
Post summary
CVE-2026-44222 exposes a token injection flaw in vLLM’s multimodal support for versions 0.6.1 through 0.19.9.
1 of 1 entries
| Part | Vendor | Product | Version | Target SW | Target HW |
|---|---|---|---|---|---|
| App | vllm | vllm | - | - | - |