This post is misleading and seems aimed at promoting a tool. As far as I can tell, there’s no actual exploit in the MCP server. The described scenario depends on tricking an authorized user into executing malicious commands. That applies to any MCP server with write access.