binary-husky gpt_academic Server-Side Request Forgery Vulnerability

A Server-Side Request Forgery (SSRF) vulnerability has been identified in binary-husky/gpt_academic, specifically in version git 310122f. The application includes a feature to download papers from arxiv.org, but the URL validation is insufficient. This flaw allows an attacker to manipulate the application into accessing any URL, including internal services, and reading the response. Such exploitation could retrieve data accessible only from the server, like AWS metadata credentials, potentially escalating local exploits to network-based attacks.

Impact

Exploitation of this vulnerability could lead to unauthorized access to server-only data, such as AWS metadata credentials. Additionally, it could allow an attacker to escalate local exploits into network-based attacks.

Reproduction

The vulnerability can be reproduced by sending a WebSocket request to the application with a crafted URL that bypasses the arxiv.org validation. This can be done by appending a random string to a URL that the application would normally accept, tricking it into thinking it's a valid arxiv link. Once the request is processed, the application will access the specified URL and return the response, which can then be captured and reviewed.