Which versions of MLflow are affected by EUVD-2026-74038?

MLflow is a widely-used machine learning workflow platform relied on by data science and ML engineering teams to manage model development and deployment.

How to fix or mitigate EUVD-2026-74038?

Within 24 hours: Inventory all MLflow deployments and model registries across your organization and notify all data science, ML engineering, and research teams of this vulnerability. Restrict model loading to only trusted internal registries with strict access controls and disable external registry access. Within 7 days: Implement sandboxed, air-gapped environments for any necessary model evaluation and testing, deploy cryptographic signature verification for all loaded model artifacts, and audit recently accessed or loaded models for signs of compromise or suspicious changes. Within 30 days: Monitor Databricks and MLflow project channels for patch availability (currently unavailable), evaluate alternative ML platforms with stronger supply-chain security controls, and establish a migration plan if a patch is not released within 60 days. Coordinate with your network security team to segment and restrict traffic to model registries from critical infrastructure.

MLflow EUVD-2026-74038

Q: What is the CVSS score of EUVD-2026-74038?

EUVD-2026-74038 has a CVSS 4.0 base score of 8.6 (High). CVSS vector: CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X. EPSS exploitation probability: 0.3%.

| CVE-2026-79721 HIGH

Inclusion of Functionality from Untrusted Control Sphere (CWE-829)

2026-09-08 6f8de1f0-f67e-45a6-b68f-98777fdb759c

GHSA-jw72-756q-9hfr

RCE

8.6

CVSS 4.0 · Vendor: 6f8de1f0-f67e-45a6-b68f-98777fdb759c

Severity by source

Vendor (6f8de1f0-f67e-45a6-b68f-98777fdb759c) PRIMARY

8.6 HIGH

CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

vuln.today AI

8.0 HIGH

Network-accessible registry fetch (AV:N), no complexity (AC:L), low-privileged load operation (PR:L), victim must actively load the artifact (UI:R), no scope change, full system impact.

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

4.0 AV:N/AC:L/AT:N/PR:L/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N

Primary rating from Vendor (6f8de1f0-f67e-45a6-b68f-98777fdb759c).

CVSS VectorVendor: 6f8de1f0-f67e-45a6-b68f-98777fdb759c

Attack Vector

Network

Attack Complexity

Low

Privileges Required

Low

User Interaction

Scope

Lifecycle Timeline

Analysis Generated

Sep 08, 2026 - 21:11 vuln.today

CVE Published

Sep 08, 2026 - 19:19 cve.org

HIGH 8.6

DescriptionCVE.org

Code execution can occur in versions of the MLflow platform running version 0.0.1 or newer, enabling a maliciously crafted model artifact to execute arbitrary code on an end user's system when loaded by the project.

AnalysisAI

Arbitrary code execution in MLflow (all versions ≥ 0.0.1) is achievable through a maliciously crafted model artifact that exploits the platform's model-loading pipeline, triggering code execution on the end user's system when the artifact is deserialized. An attacker with the ability to publish or distribute a poisoned model to a shared registry can fully compromise any data scientist or ML engineer who loads it, gaining complete control over their workstation or compute environment. …

Unlock full vulnerability intelligence

Risk assessment & exploitation conditions
Attack chain visualization
Remediation with exact patch versions
Threat intelligence from 22 sources
Personal watchlist & email alerts

Continue with Google Continue with GitHub

Free forever · No credit card required

Attack ChainAIDerived

Hypothetical attack flow derived from CVE metadata

Access

technique details hidden

Delivery

technique details hidden

Exploit

technique details hidden

Execution

technique details hidden

Persist

technique details hidden

Impact

technique details hidden

Access

technique details hidden

Delivery

technique details hidden

Exploit

technique details hidden

Execution

technique details hidden

Persist

technique details hidden

Impact

technique details hidden

Vulnerability AssessmentAI

Exploitation	The victim must actively call an MLflow model-loading API (e.g., mlflow.load_model(), mlflow.pyfunc.load_model(), or a flavor-specific loader) against an attacker-controlled artifact. … Additional conditions and limiting factors are described in the full assessment.
Risk Assessment	The CVSS 4.0 base score of 8.6 reflects high CIA impact on the vulnerable system with no subsequent-system impact recorded (SC:N/SI:N/SA:N). … Full risk analysis with EPSS, KEV, and SSVC signal comparison available after sign-in.
Exploit Scenario	Full exploit scenario with step-by-step reproduction available after sign-in.
Remediation	No vendor-released patch version is identified in the available data - the affected version range spans all releases, and the HiddenLayer advisory (https://www.hiddenlayer.com/sai-security-advisory/2026-09-mlflow) should be consulted for any coordinated fix timeline. … Detailed patch versions, workarounds, and compensating controls in full report.

Recommended ActionAI

Within 24 hours: Inventory all MLflow deployments and model registries across your organization and notify all data science, ML engineering, and research teams of this vulnerability. …

Threat intelligence, references, and detailed analysis are available after sign-in.