Pfadüberschreitung in mudler/localai

The path traversal vulnerability in localai allows an attacker to bypass intended directory restrictions and delete files outside of the expected scope. By crafting a malicious request with a manipulated model parameter, an attacker can traverse the file system and target sensitive data, configuration files, or even critical system files. The impact can range from data exfiltration and denial of service to complete system takeover, depending on the permissions of the user account running the localai process. This vulnerability shares similarities with other path traversal exploits where insufficient input validation leads to unauthorized file access and modification.

Organizations deploying localai in production environments, particularly those using older versions (≤2.16.0), are at risk. Shared hosting environments where multiple users share the same localai instance are also vulnerable, as an attacker could potentially exploit the vulnerability to delete files belonging to other users. Users relying on localai for critical data processing or storage should prioritize patching.

• docker: Inspect running containers for localai versions prior to 2.16.0 using docker ps and docker exec -it <container_id> localai --version. • file system: Monitor file system activity for unexpected file deletions, particularly in directories containing sensitive data. Use tools like inotifywait (Linux) or file integrity monitoring solutions. • network: Monitor network traffic for suspicious requests targeting the model deletion endpoint with unusual or manipulated parameters. Use tools like Wireshark or tcpdump. • logs: Examine localai application logs for error messages or warnings related to file access or deletion attempts.

1. 2024-06-19Disclosure

   disclosure

1. Reserviert2024-05-21
2. Veröffentlicht2024-06-19
3. Geändert2024-08-01
4. EPSS aktualisiert2026-04-02

The primary mitigation for CVE-2024-5182 is to upgrade localai to version 2.16.0 or later, which includes the necessary input validation fixes. If upgrading is not immediately feasible, consider implementing stricter file access controls and directory permissions to limit the potential impact of a successful exploit. Additionally, implement a Web Application Firewall (WAF) with rules to detect and block requests containing suspicious path traversal patterns in the model parameter. Monitor system logs for unusual file deletion activity and implement intrusion detection systems (IDS) to alert on potential exploitation attempts. After upgrade, confirm by attempting to delete a test model with a crafted path traversal payload and verifying that the deletion fails.