A Server-Side Request Forgery (SSRF) vulnerability exists in the MediaConnector class within the vLLM project's multimodal feature set. The load_from_url and load_from_url_async methods fetch and process media from user-provided URLs without adequate restrictions on the target hosts. This allows an attacker to coerce the vLLM server into making arbitrary requests to internal network resources.

Metrics

No CVSS v4.0

Attack Vector Network

Attack Complexity High

Privileges Required Low

Scope Unchanged

Confidentiality Impact High

Integrity Impact Low

Availability Impact High

User Interaction None

No CVSS v3.0

No CVSS v2

This CVE is not in the KEV list.

The EPSS score is 0.00039.

Exploitation none

Automatable no

Technical Impact partial

Affected Vendors & Products

Vendors Products
Redhat
  • Ai Inference Server
  • Enterprise Linux Ai

No data.

No data.

No data.

Advisories
Source ID Title
Github GHSA Github GHSA GHSA-3f6c-7fw2-ppm4 vLLM is vulnerable to Server-Side Request Forgery (SSRF) through `MediaConnector` class
Fixes

Solution

No solution given by the vendor.

Workaround

Red Hat has investigated whether a possible mitigation exists for this issue, and has not been able to identify a practical example. Please update the affected package as soon as possible.

References
Link Providers
https://access.redhat.com/security/cve/CVE-2025-6242 cve-icon cve-icon
https://bugzilla.redhat.com/show_bug.cgi?id=2373716 cve-icon cve-icon
https://nvd.nist.gov/vuln/detail/CVE-2025-6242 cve-icon
https://www.cve.org/CVERecord?id=CVE-2025-6242 cve-icon
History

Wed, 08 Oct 2025 00:15:00 +0000

Type Values Removed Values Added
References
Metrics threat_severity

None

 threat_severity

Important

Tue, 07 Oct 2025 20:15:00 +0000

Type Values Removed Values Added
Metrics ssvc

{'options': {'Automatable': 'no', 'Exploitation': 'none', 'Technical Impact': 'partial'}, 'version': '2.0.3'}

Tue, 07 Oct 2025 20:00:00 +0000

Type Values Removed Values Added
Description A Server-Side Request Forgery (SSRF) vulnerability exists in the MediaConnector class within the vLLM project's multimodal feature set. The load_from_url and load_from_url_async methods fetch and process media from user-provided URLs without adequate restrictions on the target hosts. This allows an attacker to coerce the vLLM server into making arbitrary requests to internal network resources.
Title Vllm: server side request forgery (ssrf) in mediaconnector
First Time appeared Redhat
Redhat ai Inference Server
Redhat enterprise Linux Ai
Weaknesses CWE-918
CPEs cpe:/a:redhat:ai_inference_server:3
cpe:/a:redhat:enterprise_linux_ai:1
Vendors & Products Redhat
Redhat ai Inference Server
Redhat enterprise Linux Ai
References
Metrics cvssV3_1

{'score': 7.1, 'vector': 'CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:H/I:L/A:H'}
cve-icon MITRE

Status: PUBLISHED

Assigner: redhat

Published:

Updated: 2025-10-07T19:57:03.472Z

Reserved: 2025-06-18T15:26:11.100Z

Link: CVE-2025-6242

cve-icon Vulnrichment

Updated: 2025-10-07T19:55:32.062Z

cve-icon NVD

Status : Awaiting Analysis

Published: 2025-10-07T20:15:36.187

Modified: 2025-10-08T19:38:09.863

Link: CVE-2025-6242

cve-icon Redhat

Severity : Important

Publid Date: 2025-10-07T00:00:00Z

Links: CVE-2025-6242 - Bugzilla

cve-icon OpenCVE Enrichment

No data.