AllegroAI ClearML Path Traversal Vulnerability Leading to Arbitrary File Writes and Potential Remote Code Execution

A path traversal vulnerability has been identified in AllegroAI ClearML version 2.0.1. This issue arises from improper handling of symbolic and hard links in the 'safe_extract' function, allowing arbitrary file writes outside the intended directory. If critical files are overwritten, this could lead to remote code execution.

Impact

Exploitation of this vulnerability allows for arbitrary file writes. Overwriting critical files such as '/etc/crontab' or '~/.bashrc' could result in arbitrary command execution.

Reproduction

The vulnerability can be reproduced by creating a TAR file that includes symbolic links pointing to files outside the TAR archive. When this TAR file is extracted using the 'safe_extract' function, the links can traverse the file system, leading to unauthorized file writes.

Remediation

Users are advised to update to AllegroAI ClearML version 2.0.2, where this vulnerability has been fixed.