Authorization Bypass in khoj AI Chat API

Executive Summary

This vulnerability is a flaw in the khoj-ai khoj software up to version 2.0.0-beta.28, specifically in the Conversation Sharing Handler component located in the file src/khoj/routers/api_chat.py. The issue arises from manipulation of the argument conversation.agent, which leads to incorrect authorization. This means that an attacker could potentially exploit this flaw remotely to bypass authorization controls.

The exploit for this vulnerability has been published and is available for use, although a fix has been proposed and is awaiting acceptance.

Useful 0 Not Useful 0

Detection Guidance

This vulnerability involves incorrect authorization related to conversation agents in the Khoj AI application, specifically when shared conversations are forked and private agent references are retained without proper access checks.

Detection would involve monitoring or inspecting conversation sharing and forking activities to identify if private or protected agent references are being improperly retained or accessed by unauthorized users.

Since the issue is related to application logic rather than network traffic patterns, detection commands would likely involve reviewing application logs or database entries related to conversation agents and their authorization states.

No specific detection commands or network signatures are provided in the available resources.

Useful 0 Not Useful 0

Impact Analysis

This vulnerability can impact you by allowing a remote attacker to manipulate authorization checks within the khoj-ai khoj application. This could lead to unauthorized access or actions within the system, potentially exposing sensitive information or allowing unintended operations.

Useful 0 Not Useful 0

Compliance Impact

The vulnerability in khoj-ai khoj allows unauthorized access to private conversation agents by retaining private agent references when conversations are shared or forked. This incorrect authorization could lead to exposure of sensitive user data or configurations across user boundaries.

Such unauthorized access and potential data exposure may impact compliance with data protection regulations like GDPR and HIPAA, which require strict controls on personal and sensitive data access and sharing.

However, the provided information does not explicitly state the direct compliance impact or any regulatory assessments.

Useful 0 Not Useful 0

Mitigation Strategies

Immediate mitigation involves preventing unauthorized access to private conversation agents when sharing or forking conversations.

A fix has been developed in pull request #1328 which re-authorizes conversation agents before chat execution, prompt construction, and chat-model selection, ensuring private agents are not improperly shared or inherited.

Until the fix is accepted and deployed, administrators should restrict or monitor sharing and forking of conversations that involve private agents to avoid unauthorized access.

Review user permissions and access controls around conversation agents and consider disabling or limiting conversation sharing features temporarily if possible.

Useful 0 Not Useful 0