Cross-Site Request Forgery (CSRF) in lunary-ai/lunary

Vulnerability

The vulnerability is a Cross-Site Request Forgery (CSRF) in lunary-ai/lunary version 1.2.34. The root cause is overly permissive CORS (Cross-Origin Resource Sharing) settings in the backend, which allow all origins. This misconfiguration makes unauthenticated endpoints susceptible to CSRF attacks [1][2].

Exploitation

To exploit this, an attacker would craft a malicious web page or link that, when visited by a victim who has a local lunary instance running, triggers a forged request. The attack vector is primarily for instances hosted locally on personal machines that are not publicly accessible. No prior authentication is needed because the vulnerable endpoints are unauthenticated [1][2].

Impact

Successful exploitation enables an attacker to sign up for new accounts and create projects on the victim's local instance, impersonating a legitimate user with local access. This can lead to unauthorized use of the instance's resources and potential data exposure within the created projects [1].

Mitigation

The issue exists in version 1.2.34. As of the publication date, a fix has not been detailed, but the recommended mitigation is to restrict CORS policies to trusted origins only, especially for unauthenticated endpoints, and to ensure instances are not exposed to untrusted networks [1][2].