Black Friday60% off BLACKFRIDAY2026
PlatformPlansPartners
Resources
ToolsBlogDocs
Sign Up →
high·Product Based Web Vulnerabilities·Updated Nov 9, 2026

CVE-2026-116690 Scanner

CVE-2026-116690 Scanner - Arbitrary File Read vulnerability in MLflow

Est. Time~10 seconds
Scan TypeSingle Scan
Targetsdomain, subdomain, ipv4
CostFree
3
Times Used
continuous scan runs
4.2k
Continuously Checked
assets under CS
0
Vulnerabilities Found
confirmed findings
References
7.5
CVSShigh
Exploitable remotely over the internet · no authentication required.
Description

MLflow is an open source AI engineering platform for agents, large language models, and machine learning models. From 3.10.0 until 3.17.0, the prompt branch in _create_model_version() accepts schemeless relative sources because path validation runs only when a source has a URL scheme, and mlflow.prompt.is_prompt is interpreted by tag presence rather than a true value. A remote caller can store a server-local path as the model version artifact location and use /api/2.0/mlflow/model-versions/get-artifact to read files accessible to the MLflow server process. Default unauthenticated servers require no privileges, while basic-auth deployments are exposed to authenticated users who can create registered models. This issue is fixed in version 3.17.0.

Attack Vector
Network
Privileges Req.
None
User Interaction
None
Affected
mlflowby mlflow
>= 3.10.0, < 3.17.0
Updated Nov 6, 2026View on NVD →
Detail

MLflow is a popular open-source machine learning lifecycle management platform used by data scientists and organizations to streamline and manage their ML models and workflows. It provides tracking, projects, and model registry capabilities to manage machine learning experiments from experimentation to deployment. Organizations utilize MLflow to facilitate collaboration and reproducibility in ML workflows, allowing teams to track and share results. The platform is utilized by both small teams and large enterprises to integrate ML systems more effectively into their applications. MLflow's extensibility allows it to work with any machine learning library, algorithm, language, and deployment tool. With a broad adoption, MLflow acts as a central hub for ML workflow management across different environments.

The Arbitrary File Read vulnerability detected in MLflow versions 3.10.0-3.16.1 stems from a path traversal flaw in the CreateModelVersion handler. Unauthenticated attackers can exploit this vulnerability to read arbitrary files from the server's filesystem. This represents a significant confidentiality risk as sensitive information from configuration or system files could be exposed. The vulnerability is linked to insufficient validation of file paths, allowing relative paths to bypass security checks. As a critical issue, it requires immediate attention and remediation to prevent exploitation. Attackers leveraging this vulnerability could potentially access system-sensitive data without any authentication barriers.

The Arbitrary File Read vulnerability in MLflow occurs in the model-versions/get-artifact endpoint, where path traversal within CreateModelVersion enables attackers to exploit a schemeless relative path. The vulnerable parameter allows paths like ../../../etc, which are processed without being sanitized. The endpoint subsequently serves these unchecked paths, thereby unintentionally exposing arbitrary files host-side. An attacker can easily craft a request to navigate through directories on a vulnerable server and infer sensitive file content. Without proper path restrictions or validation mechanisms, such endpoints are susceptible to being manipulated to reveal hidden files or system configurations.

If exploited, this vulnerability can lead to severe data exposure, including system, user, or other sensitive files present on the server. Attackers could uncover configuration files that might contain credentials, API keys, or other private data. Such information can facilitate further attacks like privilege escalation or more complex infiltration techniques. Organizations relying on MLflow could face data breaches, unauthorized data access, and potential reputational damage. Furthermore, without mitigation, continuous exploits could lead to sustained data leak risks, affecting user privacy and business compliance.

REFERENCES

Solution Advice
  • Upgrade MLflow to version 3.17.0 or later to patch the vulnerability.
  • Implement stringent path validation mechanisms to validate and sanitize inputs effectively.
  • Monitor server access logs for unusual patterns indicating potential exploitation.
  • Regularly audit and update configuration files to remove sensitive data exposure.
  • Enforce least privilege access controls to minimize potential impact from compromised files.

Get AI-powered remediation steps tailored to your asset.

Try AI Solutions →

Check your infrastructure.
Right now.

12,000+ scanners. Free to start. No credit card required.