An unauthenticated remote code execution vulnerability in the IBM mcp-context-forge AI gateway poses a critical risk of full system compromise.

The python_sandbox_server component fails to implement necessary guards on attribute access, allowing attackers to bypass validation and invoke sensitive functions like subprocess.Popen. This enables unauthenticated attackers to execute arbitrary system commands with the privileges of the underlying server process.

Successful exploitation grants an attacker complete control over the affected server, leading to potential data exfiltration, lateral movement within the network, and the disruption of critical AI services. Given the CVSS score of 10.0, this vulnerability represents the highest level of risk, as it is both remotely exploitable and does not require authentication.

Immediate Action: Update the mcp-context-forge python_sandbox_server component to version 1.0.2 or later immediately.

Proactive Monitoring: Monitor server logs for unexpected subprocess executions or unusual calls to the execute_code tool.

Compensating Controls: If an immediate update is not feasible, restrict network access to the HTTP/SSE transport interface to trusted internal segments only, or transition to stdio-only deployments to reduce the attack surface.

Public Exploit Available: No (exploit_available: false)

This vulnerability is critical and requires immediate attention due to the potential for unauthenticated remote code execution. Administrators should prioritize upgrading to version 1.0.2 across all affected environments to eliminate the underlying code injection vector. Failure to remediate this issue exposes the infrastructure to total system compromise.