CVE-2026-48746
Published: 22 June 2026
Summary
CVE-2026-48746 is a critical-severity HTTP Request/Response Smuggling (CWE-444) vulnerability in Vllm Vllm. Its CVSS base score is 9.1 (Critical).
Operationally, exploitation aligns with the MITRE ATT&CK technique Exploit Public-Facing Application (T1190); ranked in the top 46.0% of CVEs by exploit likelihood; it is not currently listed in the CISA KEV catalog.
OWASP Top 10 for Web (2025)
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2026-38401
Vulnerability details
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…
more
to use the API without providing the configured VLLM_API_KEY or --api-key. This vulnerability is fixed in 0.22.0.
- CWE(s)
Related Threats
MITRE ATT&CK Enterprise TechniquesAI
Why these techniques?
Auth bypass on public-facing LLM inference API directly enables exploitation of Internet-facing application for unauthorized access.
Affected Assets
Mitigating Controls
Likely Mitigating Controls AI
Per-CVE control mapping for this CVE has not run yet; the list below is derived from the weakness types (CWEs) cited in the NVD entry.
Establishes and maintains trust boundaries with external organizations before allowing their systems to interact with organization resources.
Prevents information from crossing trust boundaries without explicit approved authorizations.
Defining interfaces, controls, and trust responsibilities in agreements helps prevent violations of trust boundaries during data exchanges.
Authorizing and reviewing connections helps maintain proper trust boundaries between internal components.
Controlling media movement outside controlled areas maintains separation between internal and external trust boundaries.
Review of inter-system matching programs identifies and corrects trust-boundary violations before data crosses organizational or policy domains.
Defines explicit trust boundaries for PII use via documented purposes and prevents processing outside those boundaries.
Explicitly binding attributes to information crossing trust boundaries prevents loss of security context that leads to trust-boundary violations.