Mass Assignment Vulnerability in mintplex-labs/anything-llm

The vulnerability lies in the /api/invite/:code endpoint, which handles account creation via invitation links. Due to a lack of input validation and property allowlisting, an attacker can intercept and modify the HTTP request, injecting a role property with a value of admin. This effectively grants the attacker administrative privileges within the application. Successful exploitation allows for unauthorized access to sensitive data, modification of system configurations, and potentially complete control over the Anything LLM instance. The impact is particularly severe given the potential for privilege escalation and the ability to bypass standard authentication mechanisms. This vulnerability shares similarities with other mass assignment flaws where insufficient input sanitization leads to unauthorized privilege assignment.

Organizations deploying Anything LLM in production environments, particularly those relying on invitation-based account creation, are at risk. Shared hosting environments where multiple users share the same instance of Anything LLM are also particularly vulnerable, as a compromised account could potentially impact other users.

• nodejs / server:

# Check for suspicious account creation requests in access logs
grep 'POST /api/invite/:code' access.log | grep 'role=admin'

• nodejs / server:

# Monitor for unexpected properties in request bodies using a logging framework
# (Example: Winston, Bunyan)
# Log requests to /api/invite/:code and inspect for the presence of 'role' property

1. 2024-04-16Disclosure

   disclosure

1. Gereserveerd2024-01-10
2. Gepubliceerd2024-04-16
3. Gewijzigd2024-08-01
4. EPSS bijgewerkt2026-04-02

The primary mitigation for CVE-2024-0404 is to upgrade to version 1.0.0 or later of Anything LLM, which includes the necessary input validation and property allowlisting to prevent the mass assignment vulnerability. If upgrading immediately is not feasible, consider implementing a temporary workaround by carefully scrutinizing all incoming requests to the /api/invite/:code endpoint. Implement strict input validation to ensure that the role property is not present in the request body and that any other unexpected properties are rejected. Consider using a Web Application Firewall (WAF) to filter out malicious requests containing suspicious properties. Monitor application logs for any unusual account creation attempts or modifications to user roles.