CVE-2026-105754 vLLM: Scale-out disaggregated multimodal transport trusts caller-supplied features
Attackers can cause denial of service or retrieve cross-request cache state by sending malformed multimodal data to the vLLM inference endpoint.
- Exploited in the wild
- Not known
- Login needed
- Any signed-in user
- Attackable
- Over the internet
- Victim action needed
- Not needed
- Fix
- Available
What is CVE-2026-105754?
CVE-2026-105754 is a medium-severity vulnerability in Vllm Vllm, published on 5 Oct 2026. Any signed-in user can exploit it over the network, with no action from the victim.
Official description
vLLM is an inference and serving engine for large language models. Prior to 0.30.0, the /inference/v1/generate endpoint in the disaggregated scale-out path accepts caller-supplied tensors in the features.kwargs_data field, cache identifiers in the features.mm_hashes field, ranges in the features.mm_placeholders field, and wire-selected multimodal field processors without rebinding them to the active model renderer contract. Forged grid geometry, field types, or non-positive placeholder lengths can terminate the shared EngineCore; when an attacker knows or can induce a victim's content hash, forged cache hashes can poison or retrieve cross-request encoder-cache state; and dropped sparse placeholder masks can alter replayed transport semantics. This issue is fixed in version 0.30.0.
Weakness: CWE-704, CWE-668, CWE-639, CWE-617, CWE-20, CWE-1284
Which versions are affected by CVE-2026-105754?
3 open source packages ship an affected version, including vllm. It affects 1 product.
| Package | Affected | Fixed in |
|---|---|---|
vllm PyPI | All versions | 0.30.0 |
vllm PyPI | All versions | 0.30.0 |
vllm PyPI | All versions | 0.30.0 |
Products include Vllm Vllm.
How do I fix CVE-2026-105754?
Upgrade vllm to 0.30.0 or later.
Is CVE-2026-105754 being exploited?
There is no known exploitation of CVE-2026-105754 in the wild and no public exploit code.
Other vulnerabilities in Vllm Vllm
- CVE-2025-32444 vLLM Vulnerable to Remote Code Execution via Mooncake Integration
- CVE-2026-22778 vLLM leaks a heap address when PIL throws an error
- CVE-2025-47277 vLLM Allows Remote Code Execution via PyNcclPipe Communication Service
- CVE-2024-11041 Remote Code Execution in vllm-project/vllm
- CVE-2026-48746 vLLM: OpenAI auth bypass
Sources
Also known as GHSA-ph72-cqr5-qpp7, MINI-9qvv-g62j-xfp6, MINI-rrrx-2p2h-fqcc, MINI-vjjv-ghp2-q798, PYSEC-2026-4309.