A remote code injection vulnerability in FoundationAgents MetaGPT allows unauthenticated attackers to execute arbitrary code, posing a significant risk to system integrity.

The vulnerability resides in the generate_thoughts function within the metagpt/strategy/tot.py file. This flaw allows an unauthenticated, remote attacker to perform code injection by manipulating the application logic.

Successful exploitation of this vulnerability permits unauthorized code execution, which could lead to full system compromise, data theft, or the installation of persistent malicious backdoors. Given the CVSS score of 7.3, this represents a high-severity risk that could result in significant operational disruption and loss of confidentiality.

Immediate Action: As no official patch is currently available, users should restrict network access to affected instances and consider disabling the Tree-of-Thought Solver functionality if it is not business-critical.

Proactive Monitoring: Monitor application logs for suspicious entries originating from the generate_thoughts function and investigate any unexpected process execution behavior on the host server.

Compensating Controls: Deploy a Web Application Firewall (WAF) with rules configured to detect and block malicious injection patterns targeting Python-based application endpoints.

Public Exploit Available: Yes, a published proof-of-concept exists as detailed in the GitHub issue report linked in the vulnerability documentation.

The presence of a public proof-of-concept combined with the ability for unauthenticated remote exploitation makes this a high-priority concern. Organizations using MetaGPT should audit their deployments immediately and apply the recommended network restrictions until an official security update is released by the project maintainers.