Add MCP Gateway shoe store support chatbot sample

[ramyaravi1]

Summary

Adds a new sample under MCP-Gateway/shoe-store-support-chatbot/ — a Streamlit

• AWS Bedrock chatbot that connects to the NetApp Instaclustr MCP Gateway over
  Streamable HTTP. It demonstrates an AI agent reaching live data through a single,
  secured MCP endpoint across two backends:

• ordersapi (HTTP Server) → get_orders lists a shopper's recent orders.
• supporttickets (Kafka) → submit_request publishes a support request to the
  support-requests topic.

This is the companion code for the MCP Gateway tutorial (connecting an AI agent
to Apache Kafka and HTTP Server backends).

What's included

• src/shoe_bot/ — app code (Streamlit UI, Bedrock Converse agent, MCP manager,
  config loader, async runtime).
• example_mcp_server.py — optional minimal stdio MCP server for local testing.
• config.example.json — sanitized config template (copy to config.json).
• pyproject.toml, uv.lock — uv project + locked dependencies.
• README.md — setup, configuration, and run instructions.
• LICENSE — MIT.

Security

No real endpoints or credentials are committed. Only the sanitized
config.example.json (placeholder URL) is included; the live config.json is
intentionally excluded. AWS credentials are read at runtime from the standard
AWS credential chain.

Testing

• uv sync installs dependencies.
• uv run streamlit run src/shoe_bot/app.py launches the chatbot.
• Verified the two sample prompts trigger get_orders (HTTP) and submit_request
  (Kafka) respectively against a configured MCP Gateway.

Notes for reviewers

• Added a per-sample MIT LICENSE. Happy to remove it and defer to the repo's
  overall licensing if that's the preferred convention here.