add-app-to-server

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to clone the official MCP Apps SDK repository from GitHub (modelcontextprotocol/ext-apps). This is a well-known project and is considered a safe and trusted source for developers working with the Model Context Protocol.
  • [COMMAND_EXECUTION]: Includes standard development commands for installing dependencies (npm install), building the UI (vite build), and running the server (tsx server.ts). These commands are necessary for the skill's primary purpose and do not perform unauthorized or privileged actions.
  • [PROMPT_INJECTION]: The skill sets up event handlers (ontoolinput, ontoolresult) to process data from the agent. This represents an indirect prompt injection surface; however, the skill specifies that the UI is rendered within a sandboxed iframe in the host environment, which is a key security feature of the SDK designed to isolate the UI from the host.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 03:57 AM