An unauthenticated remote code execution vulnerability in GPT Researcher v3.3.7 allows attackers to execute arbitrary system commands via malicious configuration injection.

This vulnerability involves a lack of authentication and role-based access control, allowing an unauthenticated attacker to inject malicious Model Context Protocol (MCP) configurations into the user interface. By tricking a user into interacting with a crafted HTML page, an attacker can trigger the execution of arbitrary commands on the underlying host system.

The CVSS score of 9.8 reflects the critical nature of this flaw, as it allows for full system compromise without requiring prior authentication. Successful exploitation enables attackers to gain persistent access to the host, exfiltrate sensitive data, or deploy malware, which could lead to severe reputational damage and operational disruption.

Immediate Action: Restrict access to the GPT Researcher interface behind a secure authentication gateway and verify if the vendor has released a security patch to address the underlying command injection flaw.

Proactive Monitoring: Monitor server logs for unexpected process execution, unusual outbound network traffic, and unauthorized attempts to modify configuration files.

Compensating Controls: Deploy a Web Application Firewall (WAF) to block suspicious input patterns and ensure the application is running in an isolated container environment with minimal privileges.

Public Exploit Available: No

Given the critical severity of this remote code execution flaw, organizations should immediately restrict public access to the GPT Researcher interface. Until a verified patch is applied, ensure the application is only accessible via secure, authenticated internal networks to mitigate the risk of unauthorized command execution.