CVE-2026-22778

CVE Enrichment

CVSS: 9.8 CRITICAL
EPSS: 16.70%
EPSS Percentile: 96.97%
CISA KEV: Not listed

Description

vLLM is an inference and serving engine for large language models (LLMs). From 0.8.3 to before 0.14.1, when an invalid image is sent to vLLM's multimodal endpoint, PIL throws an error. vLLM returns this error to the client, leaking a heap address. With this leak, we reduce ASLR from 4 billion guesses to ~8 guesses. This vulnerability can be chained a heap overflow with JPEG2000 decoder in OpenCV/FFmpeg to achieve remote code execution. This vulnerability is fixed in 0.14.1.

Technical Details

CVSS Version: 3.1
CNA / NVD Source: security-advisories@github.com
Published: 03-Feb-2026 06:16
Last Modified: 15-Jul-2026 09:18
EPSS Date: 2026-10-12
CWE: CWE-532, CWE-209

CVSS Vector


            CVSS:3.1                            AV:N /
                            AC:L /
                            PR:N /
                            UI:N /
                            S:U /
                            C:H /
                            I:H /
                            A:H

Affected Products / Versions

vllm / vllm
Version / affected range: * · from (inclusive) 0.8.3 · before (exclusive) 0.14.1
cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*

References

Last enrichment sync: 13-Oct-2026 16:57
Sources: NVD · FIRST EPSS · CISA KEV
Latest successful provider update; providers refresh independently.

Reported Context (1)

CVE reference: CVE-2026-22778 Metasploit Update Adds Exploit Modules, Windows AArch64 Payloads and Enhancements

CVE (8)

People (10)

Products (8)

Tools (1)

Note: Related entities, including threat actors, malware, CVEs, MITRE ATT&CK techniques, vendors, products, tools, countries, and industries, are shown when they appear in the same reporting. Their presence does not necessarily mean they were targeted, compromised, vulnerable, responsible for the activity, or directly involved in the incident.