Skip to content
MLflowGHSA-65h7-c7c4-mghx

MLflow Has a Server-Side Request Forgery (SSRF) Vulnerability

High7.1CVE-2026-2393 · Published May 11, 2026 · updated Jul 13, 2026

GitHub advisory

Affected versions

PackageAffectedFixed in
mlflow
PyPI
< 3.9.03.9.0
Details and references

A Server-Side Request Forgery (SSRF) vulnerability exists in MLflow versions prior to 3.9.0. The `_create_webhook()` function in `mlflow/server/handlers.py` accepts a user-controlled `url` parameter without validation, and the `_send_webhook_request()` function in `mlflow/webhooks/delivery.py` sends HTTP POST requests to this attacker-controlled URL. This allows an authenticated attacker to force the MLflow backend to send HTTP requests to internal services, cloud metadata endpoints, or arbitrary external servers. The lack of input sanitization, URL scheme filtering, or allowlist validation on the webhook URL enables exploitation, potentially leading to cloud credential theft, internal network access, and data exfiltration.

CVSS 3.0
CVSS:3.0/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:L/A:N
Severity from
GitHub (reviewed advisory)
Weakness
CWE-918
Also known as
BIT-mlflow-2026-2393, CVE-2026-2393, PYSEC-2026-2219

More MLflow advisories

All MLflow
DateAdvisory
May 11MLflow allows an unauthenticated remote attacker to read arbitrary files from the server's filesystem
CVE-2026-2614High7.5fixed in 3.10.0
May 15MLflow: unauthenticated access to certain FastAPI routes
CVE-2026-2652High8.6fixed in 3.11.0
May 18MLFlow Creates a Temporary File With Insecure Permissions
CVE-2026-4137High7.0fixed in 3.11.0
May 19MLflow: Improper Origin Validation in MLflow Assistant /ajax-api Endpoints Enables Browser-Mediated Local Command Execution
CVE-2026-2611Critical9.6fixed in 3.10.0
May 21MLflow authenticated users can enumerate any registered model versions due to lack of per-model permissions checks
CVE-2026-2734Medium6.5fixed in 3.10.0
May 26MLflow allows unauthorized access to multipart upload endpoints when the `--serve-artifacts` mode is enabled
CVE-2026-2651Critical9.0fixed in 3.11.0rc1

Critical advisories by email

Wednesdays: the week’s critical and high advisories in the AI and data stack, with the fixed versions. Only in weeks that have some.

Double opt-in. Unsubscribe any time.