Executive brief
egtai gmx-vmd-mcp is a tool used to manage molecular dynamics simulations and visualizations. A security flaw allows remote attackers to execute arbitrary commands on the host system by providing specially crafted file paths for simulation data. This could lead to a total compromise of the server, including unauthorized data access, modification of research results, or service disruption.
Technical details
A command injection vulnerability exists in the launch_vmd_gui_tool function within mcp_server.py of egtai gmx-vmd-mcp (up to version 0.1.0). The application accepts user-controlled 'structure_file' and 'trajectory_file' arguments, which are interpolated into a shell command string and executed using os.system(). Because the application fails to sanitize shell metacharacters (such as semicolons or ampersands) in these paths, a remote attacker can break out of the intended command to execute arbitrary code with the privileges of the service account. A public proof-of-concept (PoC) has been released, and the vendor has not yet provided a patch. Security engineers should migrate to subprocess.run with shell=False to mitigate this risk.
Affected products
- egtai gmx-vmd-mcp 0.1.0
Timeline
- 2026-04-10: disclosed: Vulnerability reported to vendor via GitHub issue
- 2026-04-27: advisory: Initial disclosure by VulDB
- 2026-04-28: other: NVD publication date