A critical NoSQL injection vulnerability in FastGPT allows authenticated attackers to perform account takeovers by bypassing password verification requirements.

This is a NoSQL injection vulnerability (CWE-943) located in the password change endpoint. An authenticated attacker can inject MongoDB query operators to bypass the requirement for the current password, leading to unauthorized account modification and potential full account takeover.

The ability for a low-privileged user to change the password of their own account, or potentially other accounts if combined with identifier manipulation, poses a severe threat to system integrity and data confidentiality. Given the CVSS score of 8.8, this vulnerability represents a high risk of unauthorized access and persistence within the AI Agent platform, which could lead to significant operational disruption or the compromise of sensitive AI-driven workflows.

Immediate Action: Update the FastGPT deployment to version 4.14.9.5 or later, where the input sanitization logic for the password change process has been corrected.

Proactive Monitoring: Review application access logs for unusual patterns in the password change endpoint, specifically looking for requests containing non-alphanumeric characters or MongoDB query operator syntax.

Compensating Controls: Implement a Web Application Firewall (WAF) rule to inspect and block incoming requests containing common NoSQL injection sequences, such as $gt, $ne, or $where, directed at the authentication or profile management endpoints.

Public Exploit Available: No

This vulnerability presents a high risk to the platform by enabling unauthorized account takeover. Administrators should prioritize the deployment of version 4.14.9.5 immediately. Failure to patch allows even low-privileged authenticated users to escalate their access and gain full control over accounts, necessitating prompt attention to ensure the security of the FastGPT environment.