Skip to main content
VulnerabilityPublished

CVE-2026-114053: vLLM · Request-controlled mm_processor_kwargs.code_revision allows remote code execution

Affects vllmSource GCVE Database

0
Authoritative record:CVE-2026-114053· Critical 9.2

Not known to be exploited · CVSS 9.2 Critical : upgrade to 0.31.0 as a priority.

Data as of 27 Oct 2026, 16:29 UTC

CVSS 4.0 base score
9.2Critical
How severe the flaw is on the 0–10 CVSS scale.
Patch
Available – fixed in 0.31.0
CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N

Exploitation evidence

  1. Disclosed: evidence found.

    CVE published 21 Oct 2026

  2. Detectable: no evidence on record.

    No public nuclei template indexed

  3. Weaponised: no evidence on record.

    No public exploit code indexed in Metasploit or Exploit-DB

  4. Exploited in the wild: no evidence on record.

    Not in CISA KEV and no source reports exploitation

  5. Observed by OffSeq Mirage: no evidence on record.

    No attack traffic recorded by OffSeq Mirage sensors

A missing step is not evidence of safety. Sources: cve.org (opens in a new tab), the Metasploit, Exploit-DB and nuclei-templates index as of 26 Oct 2026.

What to do

Upgrade vllm to 0.31.0.

Affected software

ProductAffectedFixed in
PyPIvllmAffected
  • >=0.7.3 <0.31.0
Fixed in
  • 0.31.0

Check your own hosts for these packages with threat-finder (opens in a new tab), our free open-source scanner.

AI-suggested

A fixed version of vLLM is available starting at 0.31.0; operators should upgrade to this version or later. If upgrading is not immediately possible, operators should disable the trust_remote_code option to prevent dynamic code execution from untrusted revisions. Additionally, protect the /tokenize endpoint with authentication or an external proxy to prevent unauthenticated access. The default configuration with trust_remote_code disabled is not vulnerable.

Description

vLLM versions from 0.7.3 up to but not including 0.31.0 have a vulnerability where an attacker can control the mm_processor_kwargs.code_revision parameter in multimodal requests. When the operator enables trust_remote_code and serves a model whose Hugging Face processor uses dynamic auto_map code, this allows the attacker to execute arbitrary Python code by importing a malicious processor module revision. The vulnerability is exploitable via unauthenticated /tokenize endpoints unless protected externally. A fix is available in vLLM version 0.31.0.

References

Community reviews

Loading reviews…