Kedro hat eine Pfadüberschreitung beim Laden versionierter Datensätze über eine nicht bereinigte Versionszeichenkette.

The core of the vulnerability lies in the getversioned_path() method, which directly interpolates user-supplied version strings into filesystem paths without proper sanitization. This means that malicious input, such as ../, can be injected into the path, allowing an attacker to traverse outside the intended versioned dataset directory. Successful exploitation could lead to unauthorized access to sensitive files and directories on the system, including configuration files, credentials, or other critical data. The provided example demonstrates accessing a 'secrets' file via the CLI, highlighting the potential for data exfiltration. This vulnerability poses a significant risk to data pipelines that rely on Kedro for managing datasets and workflows.

Data science teams and organizations using Kedro for data pipeline orchestration are at risk, particularly those relying on older versions (0.0.0 - 1.2.9). Environments where Kedro pipelines process or store sensitive data, such as financial or healthcare information, face a higher risk of data compromise. Shared hosting environments where multiple Kedro pipelines are deployed on the same server could also be vulnerable.

• python / kedro:

import os
import kedro

def check_versioned_path(version):
    try:
        path = kedro.io.core._get_versioned_path('dataset', version=version)
        # Check if the path is within the expected directory
        if '..' in version:
            print(f"Potential path traversal detected with version: {version}")
    except Exception as e:
        print(f"Error checking path: {e}")

# Example usage with a malicious version string
check_versioned_path('dataset:../../../secrets')

• generic web: Check Kedro pipeline configuration files for version strings that include traversal sequences (../). Examine access logs for requests containing suspicious path parameters.

1. 2026-04-06Disclosure

   disclosure

1. Reserviert2026-04-01
2. Veröffentlicht2026-04-06
3. Geändert2026-04-07
4. EPSS aktualisiert2026-04-14

The primary mitigation is to upgrade Kedro to version 1.3.0 or later, which includes a fix for this vulnerability. If upgrading is not immediately feasible, consider implementing input validation on version strings used within Kedro pipelines. Restrict access to the Kedro CLI and related configuration files to prevent unauthorized modification of version strings. Employ a Web Application Firewall (WAF) or proxy to filter out malicious path traversal attempts. Monitor Kedro logs for suspicious activity, particularly attempts to access files outside the expected versioned directories. After upgrading, confirm the fix by attempting to access a file outside the intended directory using a crafted version string; the access should be denied.