by Kastro CVE-2026-48746
vllm-project vllm — vLLM: OpenAI auth bypass
Description · from the CVE record
vLLM is an inference and serving engine for large language models (LLMs). From 0.3.0 until 0.22.0, a vulnerability in ASGI web servers and starlette's trust on those web servers enables an authentication bypass of the OpenAI API AuthenticationMiddleware. It allows to use the API without providing the configured VLLM_API_KEY or --api-key. This vulnerability is fixed in 0.22.0.
Affected
ProductAffectedFixed inSource
vllm>= 0.3.0, < 0.22.0–CNA
Red Hat AI Inference Server 3.2see advisory1782951012redhat-SADP
Red Hat AI Inference Server 3.2see advisory1782951244redhat-SADP
Red Hat AI Inference Server 3.3see advisory1782353093redhat-SADP
Red Hat AI Inference Server 3.3see advisory1782352847redhat-SADP
Red Hat Ansible Automation Platform 2.6see advisory1783920640redhat-SADP
Red Hat Ansible Automation Platform 2.7see advisory1783918403redhat-SADP
Red Hat Migration Toolkit for Applications 8.2see advisory1784109883redhat-SADP
Red Hat OpenShift AI 2.25see advisory1783082430redhat-SADP
Exploit Intelligencesee advisory–redhat-SADP
OpenShift Lightspeedsee advisory–redhat-SADP
OpenShift Lightspeedsee advisory–redhat-SADP
Scores by source
GitHub_M (CNA)9.1CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:H
redhat-SADP (CNA)9.1CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:H
CISA ADPSSVCExploitation: poc · Automatable: yes · Technical impact: partial