mcp-todo-skill

You can manage todo lists for AI agents by deploying a stateless MCP server that exposes five core tools: add, list, complete, delete, and update. It uses a PostgreSQL-backed store with per-request database transactions to ensure data persistence.

User_id isolation is an explicit security requirement enforced in every tool call to separate task data. It ensures that AI agents manage todo lists strictly for the requesting user, preventing unauthorized access across different chat-based workflows.

To set up the PostgreSQL-backed MCP server, you must configure your DATABASE_URL environment variable. After configuring the database connection, you can run the FastMCP Python server locally and register the extension for immediate use in production.

Yes, the server is designed for scalable task orchestration in chat-based workflows. Its stateless architecture allows AI assistants to handle concurrent user requests efficiently without maintaining session data between interactions.

A stateless server does not retain session context between AI agent requests, requiring a valid user_id and a persistent PostgreSQL database connection for every operation. All task state is managed externally through the database rather than in memory.

Yes, the MCP server implements five tools with strict typing and robust error handling. This ensures reliable execution of add, list, complete, delete, and update operations while maintaining data integrity during per-request database transactions.