CVE-2026-71211

Received Received - Intake

SSRF via Unvalidated API Base in MLflow AI Gateway

Publication date: 2026-08-05

Last updated on: 2026-08-10

Assigner: 309f9ea4-e3e9-4c6c-b79d-e8eb01244f2c

Description

MLflow's AI Gateway accepts an auth_config.api_base value when creating a gateway secret (mlflow/server/handlers.py, _create_gateway_secret) with no validation of scheme, host, or IP range; the value is stored verbatim. The gateway proxy endpoint (mlflow/server/gateway_api.py, raw_proxy) subsequently issues an HTTP request to that stored api_base plus a caller-supplied path and returns the full response body.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published

2026-08-05

Last Modified

2026-08-10

Generated

2026-08-17

AI Q&A

2026-08-05

EPSS Evaluated

2026-08-16

NVD

CVE-2026-71211

EUVD

EUVD-2026-53208

Affected Vendors & Products

Vendor	Product	Version / Range
mlflow	mlflow	*
mlflow	mlflow	3.14.0

Helpful Resources

Exploitability

CWE

KEV

CWE ID	Description
CWE-918	The web server receives a URL or similar request from an upstream component and retrieves the contents of this URL, but it does not sufficiently ensure that the request is being sent to the expected destination.

Attack-Flow Graph

Executive Summary

This vulnerability in MLflow's AI Gateway allows authenticated users to create a gateway secret with an unvalidated api_base URL. The system stores this URL as-is and uses it to proxy HTTP requests. Since there's no validation of the scheme, host, or IP range, users can point the proxy to internal addresses. Additionally, the permission check is missing for the CreateGatewaySecret action, allowing even read-only accounts to exploit this. This could expose cloud-instance IAM credentials via metadata services.

Detection Guidance

Check for unauthorized gateway secrets pointing to internal addresses by reviewing MLflow server logs and API calls to CreateGatewaySecret. Inspect proxy endpoint traffic for requests to unexpected hosts.

Impact Analysis

An attacker with basic authentication could create a malicious gateway secret pointing to an internal service. By accessing the proxy endpoint, they could retrieve sensitive data or credentials from internal systems, potentially leading to unauthorized access, data breaches, or privilege escalation within the cloud environment.

Compliance Impact

This vulnerability could lead to unauthorized access to sensitive data, violating GDPR's data protection principles and HIPAA's security requirements. It may result in data breaches, unauthorized disclosure of personal or health information, and non-compliance with regulatory obligations for data security and access controls.

Mitigation Strategies

Update MLflow to the latest version that includes validation for api_base. Restrict CreateGatewaySecret permissions to require elevated scopes. Add network-level controls to block unauthorized outbound requests from the MLflow server.

Hi! I’m here to help you understand CVE-2026-71211. Ask me anything about the vulnerability, its impact, or mitigation strategies.

0/70