CVE-2026-7215

7.3

egtai · gmx-vmd-mcp

A command injection vulnerability in the gmx-vmd-mcp MCP tool allows unauthenticated remote attackers to execute arbitrary system commands via crafted file paths in the launch_vmd_gui_tool function.

Executive summary

An unauthenticated command injection vulnerability in egtai gmx-vmd-mcp version 0.1.0 poses a critical risk of full system compromise through the exploitation of the VMD launch handler.

Vulnerability

The application is susceptible to OS command injection (CWE-78) because the launch_vmd_gui_tool function in mcp_server.py insecurely interpolates user-provided file paths into a shell command string that is subsequently executed via os.system(). An unauthenticated attacker can supply paths containing shell metacharacters to break out of the intended command structure and execute arbitrary commands with the privileges of the service account.

Business impact

Successful exploitation of this vulnerability allows an attacker to achieve remote code execution on the host system. This could lead to full unauthorized access to sensitive project data, modification of system integrity, and potential disruption of critical workflows. Although the CVSS score is 7.3, the ability for an unauthenticated attacker to execute arbitrary commands makes this a high-priority security concern.

Remediation

Immediate Action: As no official patch is currently available, users should immediately restrict access to the gmx-vmd-mcp service to trusted internal networks only. If possible, disable the impacted tool until the vendor provides a remediation that replaces the shell-based execution with safer subprocess calls.

Proactive Monitoring: Review application logs for unusual JSON-RPC requests to the launch_vmd_gui_tool method, specifically looking for arguments containing shell metacharacters such as semicolons, ampersands, or backticks.

Compensating Controls: Implement input validation at the gateway level to reject any file path arguments that contain shell metacharacters, effectively neutralizing the injection vector.

Exploitation status

Public Exploit Available: Yes — a published proof-of-concept exists, attributed to the researcher's write-up at the provided GitHub issue tracker.

Analyst recommendation

The presence of a public proof-of-concept for this command injection flaw, combined with the lack of an official patch, creates a significant window of exposure. Organizations utilizing gmx-vmd-mcp must treat this as a high-urgency issue and implement the recommended network restrictions and input validation controls immediately to prevent unauthorized system access.

Sources

Originally found and disclosed by SmallW (VulDB User), with VulDB CNA Team (coordinator), per the CVE Program record.