A critical path traversal vulnerability in the dexhunter kaggle-mcp tool allows unauthenticated remote attackers to manipulate filesystem paths and create arbitrary directories.

The application fails to validate the competition_id parameter in the prepare_kaggle_dataset function, allowing path traversal sequences to escape the intended data directory. This flaw is reachable by unauthenticated remote attackers and triggers directory creation before any authentication or data validation occurs.

The CVSS score of 7.3 reflects a high-severity threat to system integrity and service reliability. Successful exploitation allows an attacker to create directories in unauthorized locations, which can lead to filesystem clutter, interference with application state, or potentially pave the way for further malicious activity if those locations are later accessed by the system.

Immediate Action: As no official patch is currently available, implement input validation to restrict the competition_id parameter to expected alphanumeric formats and reject any input containing path traversal sequences like double dots.

Proactive Monitoring: Review system logs for unusual directory creation events or suspicious API calls to the prepare_kaggle_dataset tool that contain path-like characters.

Compensating Controls: Deploy a Web Application Firewall or similar filtering layer to inspect incoming requests and block those containing path traversal patterns targeting the Kaggle-MCP service.

Public Exploit Available: Yes, a published proof-of-concept exists as detailed in the referenced GitHub issue report.

Given the availability of a public proof-of-concept and the lack of a vendor-provided fix, this vulnerability poses a significant risk to affected environments. Administrators should restrict access to the vulnerable tool and apply strict input validation rules immediately to prevent unauthorized filesystem operations. Monitor the repository for future security updates and apply them as soon as they become available.