Nginxui Vulnerability Exploited Since March 2026

A critical vulnerability in the open-source nginxui web interface for managing Nginx web servers has been actively exploited by attackers since March 2026, enabling unauthenticated takeover of affected systems. Tracked as CVE-2026-33032 with a CVSS score of 9.8, the flaw allows remote attackers with network access to execute privileged commands without credentials, potentially compromising servers worldwide.

The issue affects nginxui's integration with the Model Context Protocol (MCP), a feature added in late 2025 to facilitate communication between Nginx servers and AI models. MCP exposes two HTTP endpoints: "mcp" for establishing connections, which requires both IP whitelisting and authentication, and "mcp_message" for processing tool invocations, which previously applied only IP whitelisting. With the default whitelist configured to allow all traffic, attackers could bypass authentication entirely and invoke 12 MCP tools. These include seven destructive functions, such as creating, modifying or deleting Nginx configuration files, restarting the service and triggering automatic reloads, as well as five for reconnaissance, like reading existing configurations to map infrastructure.