Critical failures across all dimensions. The single tool 'execute_skill_action' has no visible input schema, no parameter descriptions, and a generic 1-line description (23 chars). The mcp_server.py file does not implement MCP protocol, it merely prints a JSON status message and exits. No tool registration, no schema validation, no error handling. The client.py and example_usage.py files show a working Python class, but this is NOT an MCP server implementation. The code does not handle MCP requests, parse JSON-RPC, or implement the stateless request/response model required by the spec.

• Implement a proper MCP server with a JSON-RPC 2.0 request/response loop. Use the official MCP Python SDK (mcp package) or implement stateless request handling per the 2026-07-28 spec.
• Define the input schema for execute_skill_action as a JSON Schema object with 'viewer_question_text' (string, required, 10-2000 chars) and 'host_language' (enum: en, es, fr, de, zh, ja, etc., default: en).
• Expand the tool description to 50-150 characters explaining: 'Synthesizes a live-host teleprompter cue and product upsell recommendation in response to a viewer question. Supports 110+ languages. Returns instant response (<100ms) suitable for on-air delivery.'
• Document each parameter: 'viewer_question_text: The viewer's question as plain text (10-2000 chars). Examples: "Does this work overseas?", "What's the warranty?"' and 'host_language: ISO 639-1 language code (en, es, fr, de, zh, ja, etc.). Defaults to en.'
• Document the output schema: { instant_teleprompter_cue: string (the host's response), recommended_product_upsell: string (optional SKU and bundle suggestion), response_latency_ms: integer (milliseconds to generate response) }.