Design a meta-prompt architecture for building a smart prompt database. This database will store and categorize prompts, optimized for diverse AI applications.  The system should facilitate efficient prompt retrieval, modification, and version control. The database should include fields for prompt type (e.g., chain-of-thought, few-shot, role-play), target AI model (e.g., GPT-3.5-turbo, GPT-4), intended outcome (e.g., creative writing, code generation, data analysis), keywords for searching, prompt rating (effectiveness score), and associated metadata (e.g., date created, last modified, author).  The system should also include mechanisms for automatically suggesting related prompts based on user input, identifying and flagging low-performing prompts, and incorporating user feedback to improve prompt effectiveness.  The output should be a detailed design document outlining the database schema, data structures, search algorithms, and user interface elements.  Prioritize scalability and maintainability in your design. Include examples of how different prompt types would be categorized and stored. The success of this meta-prompt will be measured by the creation of a comprehensive design document that is clear, detailed, and easily implementable, leading to a functioning, robust, and scalable prompt database.