CVE-2026-60089 Details

A path traversal vulnerability has been identified in the PraisonAI pip package, specifically in the 'praisonaiagents' module, prior to version 1.6.78. The issue arises because the application automatically loads default settings from a project-local configuration file (.praisonai/config.toml) when an Agent is created. However, it fails to validate the 'output_file' path specified in these defaults. This oversight allows a repository-controlled config file to direct output to an absolute path or a relative path that traverses up the directory structure. Consequently, when the 'agent.start()' method is called without an explicit output parameter, PraisonAI writes the agent's response to the specified path, potentially overwriting files outside the project root. This file operation is performed with the same user privileges as the individual running PraisonAI, thereby enabling an untrusted project to manipulate files beyond the project's boundaries.

Users can update to PraisonAI version 1.6.78 or later, where this vulnerability has been addressed.