mlflow anfällig für Path Traversal

The core of this vulnerability lies in MLflow's improper handling of URL parameters. Attackers can leverage the semicolon (';') character within URLs to inject path traversal sequences. This allows them to bypass intended access controls and potentially read sensitive files, execute arbitrary code (depending on file contents), or even gain control of the underlying server. The impact is amplified if MLflow is deployed in environments with sensitive data or integrated with other critical systems, as a successful attack could lead to broader data breaches and system compromise. This attack pattern shares similarities with other URL parameter manipulation vulnerabilities, highlighting the importance of robust input validation.

Organizations deploying MLflow for machine learning model tracking and management are at risk, particularly those using older versions (≤2.9.2). Shared hosting environments where MLflow is deployed alongside other applications are also at increased risk, as a successful exploit could potentially compromise the entire host.

• python / mlflow:

import re

def check_mlflow_params(url):
    match = re.search(r'params=(.*?)(&|$)', url)
    if match:
        params = match.group(1)
        if ';' in params:
            print(f"Potential path traversal detected in params: {params}")

• generic web:

curl -I 'http://your-mlflow-server/some/endpoint?params=;../sensitive/file'

Inspect the response headers and body for any unexpected file disclosures.

1. 2024-04-16Disclosure

   disclosure

1. Reserviert2024-02-16
2. Veröffentlicht2024-04-16
3. Geändert2025-02-04
4. EPSS aktualisiert2026-04-02

The primary mitigation for CVE-2024-1593 is to upgrade MLflow to a version that includes the fix. Consult the MLflow release notes for the specific patched version. If upgrading immediately is not feasible, consider implementing temporary workarounds such as strict URL input validation and sanitization to prevent the injection of malicious path traversal sequences. Web application firewalls (WAFs) configured to block suspicious URL patterns can also provide an additional layer of defense. Thoroughly review MLflow's configuration to ensure that file access permissions are appropriately restricted.