MCP Server 数据库工具快速配置指南

You are an expert MCP (Model Context Protocol) server architect. I need to set up a database-connected MCP server for my AI agents.

My setup:

• Database type: [PostgreSQL/MySQL/SQLite/other]
• Use case: [data exploration / production agent tools / code generation]
• MCP clients: [Claude Code / Gemini CLI / Codex / custom agent]

Please help me:

1. Server Configuration: Generate a complete MCP server config (tools.yaml) with:

   • Connection pooling settings optimized for my database
   • Authentication setup (API keys or IAM)
   • Rate limiting and query timeout guards

2. Tool Definitions: Create 5-8 useful database tools including:

   • Schema exploration (list_tables, describe_table)
   • Safe read queries (execute_sql with parameterized queries)
   • Semantic search if applicable
   • Data summarization helpers

3. Security Checklist:

   • Read-only vs read-write access boundaries
   • Query allowlist/blocklist patterns
   • Sensitive column masking rules
   • Audit logging configuration

4. Integration Code: Provide a minimal code snippet to connect this MCP server to my agent framework (ADK/LangChain/LlamaIndex).

5. Testing Plan: A quick validation script to verify all tools work correctly.

Output everything in ready-to-use config files and code snippets.