ExploitCVE-2024-6868
Improper Input Validation (CWE-20)
Summary
A vulnerability in mudler/LocalAI version 2.17.1 allows for arbitrary file write due to improper handling of automatic archive extraction. When model configurations specify additional files as archives (e.g., .tar), these archives are automatically extracted after downloading. This behavior can be exploited to perform a 'tarslip' attack, allowing files to be written to arbitrary locations on the server, bypassing checks that normally restrict files to the models directory.
Impact
This vulnerability can lead to remote code execution (RCE) by overwriting backend assets used by the server. The potential impacts are severe, as it allows attackers to write files to arbitrary locations on the server, potentially compromising the entire system. This could result in unauthorized access, data breaches, system manipulation, and complete compromise of the LocalAI application and possibly the underlying server.
Exploitation
One proof-of-concept exploit is available on huntr.com. There is no evidence of proof of exploitation at the moment.
Patch
A patch is available. The vulnerability has been addressed in a commit on GitHub, specifically at https://github.com/mudler/localai/commit/a181dd0ebc5d3092fc50f61674d552604fe8ef9c. The patch was added on 2024-11-13.
Mitigation
1. Immediate patching: Update mudler/LocalAI to a version newer than 2.17.1 that includes the security fix. 2. Access control: Restrict network access to the LocalAI server to trusted sources only. 3. Input validation: Implement strict validation for any user-supplied input, especially for model configurations and archive files. 4. Monitoring: Implement robust logging and monitoring to detect any suspicious file system activities or unexpected file creations. 5. Principle of least privilege: Ensure the LocalAI application runs with minimal necessary permissions to limit potential damage in case of exploitation. 6. Regular security audits: Conduct frequent security reviews of the LocalAI deployment and its configurations.
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Timeline
NVD published the first details for CVE-2024-6868
Feedly found the first article mentioning CVE-2024-6868. See article
Feedly estimated the CVSS score as HIGH
EPSS Score was set to: 0.04% (Percentile: 9.9%)
A CVSS base score of 9.8 has been assigned.