MLflow: Schwachstelle ermöglicht Umgehen von Sicherheitsvorkehrungen
WID-SEC-2026-2946 · first 20.08.2026 · updated 20.08.2026 · CVSS 9,3
What it is about
MLflow ist eine Open-Source-Plattform zur Verwaltung des gesamten Machine-Learning-Lebenszyklus, einschließlich Experimenten, Reproduzierbarkeit und Bereitstellung. Ein entfernter, anonymer Angreifer kann eine Schwachstelle in MLflow ausnutzen, um Sicherheitsvorkehrungen zu umgehen und Daten offenzulegen oder zu manipulieren.
Under what conditions?
Over the network and without logging in. This is the most dangerous combination: attackers can scan reachable devices automatically.
What can happen in the worst case
An attacker could use the vulnerability to
- Sicherheitsvorkehrungen zu umgehen
- Daten offenzulegen oder zu manipulieren
Am I affected?
Affected according to the BSI:
- Open Source MLflow <=3.13.0
The installed version is usually shown under “Help → About” or in the device’s admin interface.
What helps?
Update to at least:
- Open Source MLflow <=3.13.0
Is the vulnerability already being exploited?
- CVE-2026-64849: exploited in the wild according to CISA, in the catalogue since 19.08.2026.
Information and sources
- Original BSI advisory (WID)
- MLFlow Releases vom 2026-08-19 github.com
- MLFlow GitHub Security Advisory vom 2026-08-19 github.com
- CISA Known Exploited Vulnerabilities Catalog vom 2026-08-19 www.cisa.gov
CVE: CVE-2026-64849
Source: BSI warning and information service (CERT-Bund); CISA. The assessment of the conditions is my reading of the BSI text; the external pages are in their original language.