vulnti.work

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 authentica…

Critical CVSS 9.1
CVECVE-2026-48746
First seen2026-07-22 14:16 UTC
Disclosed2026-06-22 23:16 UTC
Last updated2026-07-22 14:16 UTC
Channel statusauto

Summary

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.

In-depth triage · Auto channel

No in-depth report has been generated yet (DR-003 v2 AI pipeline is under construction).

Affected products

  • :
Lint boundary warnings (2)

The following non-blocking warnings were raised during the lint phase of this triage (e.g. reference URLs not yet on the allowlist). They do not affect entry usability and are disclosed for transparency (see DR-002).

  • REF_URL_NOT_ALLOWLISTED url not in allowlist: https://x41-dsec.de/lab/advisories/x41-2026-002-starlette
  • REF_URL_NOT_ALLOWLISTED url not in allowlist: https://x41-dsec.de/lab/advisories/x41-2026-002-starlette

Sources

  • NVD DATABASE

Original Links

Timeline

  1. nvd_ingest NVD