alphaparkinc/genpark-natural-language-sql-schema-optimizer-skill ? reverse-engineered prompt

Build me a Python tool that lets me type a question in plain English and get back optimized SQL, plus a simple explanation of what it’s doing. I want it to feel like an AI helper for database work, where it can turn a natural language request into a good SQL query, tune the query plan using AST style logic, and suggest indexing or other performance improvements when that makes sense.

Please make it easy to run from the command line with a small example script, and also expose it as an MCP server so I can plug it into an AI agent workflow. Keep it lightweight and simple to use, with a clean request and response format. If you need to check current MCP or Python docs online, go ahead and look them up. I’d also like a tiny client example that shows how to send a prompt and get the optimized SQL back.