Executive brief
LORIS, a web-based system for managing neuroimaging research data, contains a vulnerability in its publication module. An attacker with access to this module can manipulate the system into sending emails that appear to be official notifications but contain links pointing to malicious external websites. This could be used to trick researchers or staff into visiting fraudulent sites, potentially leading to further phishing or social engineering attacks.
Technical details
A vulnerability in the LORIS publication module arises from the application incorrectly trusting a 'baseURL' parameter submitted via POST requests in the FileUpload.php component. Instead of using the internal, trusted configuration value, the 'notify' function uses the attacker-controlled input to construct URLs for email notifications (e.g., submission, edit, and review alerts). An authenticated attacker with publication module access can exploit this to conduct a Server-Side Request Forgery (SSRF) variant or phishing attack by injecting an external domain into the baseURL. This results in the system generating and sending legitimate-looking emails that direct recipients to an attacker-controlled site. The issue is resolved in versions 27.0.3 and 28.0.1 by removing the baseURL parameter from the notification logic.
Affected products
- aces LORIS 20.0.0 to 27.0.2, 28.0.0
Timeline
- 2026-04-08: advisory: GitHub Security Advisory published
- 2026-04-08: patched: Fixes released in versions 27.0.3 and 28.0.1