feat(mcp-app-sample): Add new simple MCP App sample

[dmandar]

Description:

This PR introduces a new sample application demonstrating how to integrate a secure MCP (Model Context Protocol) App within the A2UI environment. It provides a complete, minimal reference implementation covering both the agent-side logic and the client-side UI components.

Key Features:
Native Module Resolution: Uses an <script type="importmap"> inside the sandboxed iframe to resolve the @modelcontextprotocol/sdk at runtime, avoiding complex bundler workarounds.
Specification Alignment: Implements the recommended "Double-Iframe" security model (Host -> Proxy -> Guest) and uses allow-same-origin as specified by the community guidelines.
Documentation: Includes a detailed README.md covering architecture, setup steps, and security considerations for production deployment.
Files Added/Modified:
samples/agent/adk/mcp-app-sample/: New directory containing the Python agent server (agent.py) and instructions.
samples/client/lit/mcp-app-sample/: New directory containing the Lit-based client UI, custom component wrapper, and optimized Vite configuration.
Verification:
Verified that the client successfully renders the injected HTML content.
Verified that clicking the button inside the sandboxed iframe successfully triggers the expected JSON-RPC tool call to the agent.

Pre-launch Checklist

• I signed the CLA.
• I read the Contributors Guide.
• I read the Style Guide.
• I have added updates to the CHANGELOG.
• I updated/added relevant documentation.
• My code changes (if any) have tests.

If you need help, consider asking for advice on the discussion board.

[gemini-code-assist]

Code Review

This pull request adds a standalone sample for integrating Model Context Protocol (MCP) Apps into the A2UI environment, featuring a FastAPI agent and a Lit-based client. The implementation uses a double-sandboxed iframe architecture for security. Feedback focuses on resolving local environment dependencies in the agent's HTML, fixing a memory leak from an unremoved global message listener, and removing redundant component registration. Additionally, improvements were suggested for server response error handling and aligning Python version requirements between the project configuration and documentation.