CVE-2026-42271

LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. From version 1.74.2 to before version 1.83.7, two endpoints used to preview an MCP server before saving it — POST /mcp-rest/test/connection and POST /mcp-rest/test/tools/list — accepted a full server configuration in the request body, including the command, args, and env fields used by the stdio transport. When called with a stdio configuration, the endpoints attempted to connect, which spawned the supplied command as a subprocess on the proxy host with the privileges of the proxy process. The endpoints were gated only by a valid proxy API key, with no role check. Any authenticated user — including holders of low-privilege internal-user keys — could therefore run arbitrary commands on the host. This issue has been patched in version 1.83.7.

### Impact Two endpoints used to preview an MCP server before saving it — `POST /mcp-rest/test/connection` and `POST /mcp-rest/test/tools/list` — accepted a full server configuration in the request body, including the `command`, `args`, and `env` fields used by the stdio transport. When called with a stdio configuration, the endpoints attempted to connect, which spawned the supplied command as a subprocess on the proxy host with the privileges of the proxy process. The endpoints were gated only by a valid proxy API key, with no role check. Any authenticated user — including holders of low-privilege internal-user keys — could therefore run arbitrary commands on the host. ### Patches Fixed in **`1.83.7`**. Both test endpoints now require the `PROXY_ADMIN` role, bringing them into line with the save endpoint. ### Workarounds If upgrading is not immediately possible, developers should block `POST /mcp-rest/test/connection` and `POST /mcp-rest/test/tools/list` at their reverse proxy or API gateway.

### Impact Two endpoints used to preview an MCP server before saving it — `POST /mcp-rest/test/connection` and `POST /mcp-rest/test/tools/list` — accepted a full server configuration in the request body, including the `command`, `args`, and `env` fields used by the stdio transport. When called with a stdio configuration, the endpoints attempted to connect, which spawned the supplied command as a subprocess on the proxy host with the privileges of the proxy process. The endpoints were gated only by a valid proxy API key, with no role check. Any authenticated user — including holders of low-privilege internal-user keys — could therefore run arbitrary commands on the host. ### Patches Fixed in **`1.83.7`**. Both test endpoints now require the `PROXY_ADMIN` role, bringing them into line with the save endpoint. ### Workarounds If upgrading is not immediately possible, developers should block `POST /mcp-rest/test/connection` and `POST /mcp-rest/test/tools/list` at their reverse proxy or API gateway.

LiteLLM es un servidor proxy (Pasarela de IA) para llamar a las API de LLM en formato OpenAI (o nativo). Desde la versión 1.74.2 hasta antes de la versión 1.83.7, dos puntos finales utilizados para previsualizar un servidor MCP antes de guardarlo - POST /mcp-rest/test/connection y POST /mcp-rest/test/tools/list - aceptaban una configuración completa del servidor en el cuerpo de la solicitud, incluyendo los campos command, args y env utilizados por el transporte stdio. Cuando se invocaban con una configuración stdio, los puntos finales intentaban conectarse, lo que generaba el comando proporcionado como un subproceso en el host proxy con los privilegios del proceso proxy. Los puntos finales estaban protegidos únicamente por una clave API de proxy válida, sin verificación de rol. Cualquier usuario autenticado - incluyendo los titulares de claves de usuario interno de bajo privilegio - podía, por lo tanto, ejecutar comandos arbitrarios en el host. Este problema ha sido parcheado en la versión 1.83.7.