Modular Architecture Prompts
AI prompts for modular architecture from the LearnWithHasan AI Coding Building Blocks (Architecture).
Analyze & Plan Module Structure
Start here to plan your refactoring From the Modular Architecture AI Coding Building Block.
Analyze my codebase and propose a modular architecture. I want to restructure it so AI coding assistants can work more effectively with isolated, focused modules. Current structure: [paste your folder structure or describe your app] For each proposed module: 1. Name and single responsibility 2. What files/functions belong in it 3. What it exports (public interface) 4. What it imports from other modules 5. Estimated size (should be <500 lines ideally) Goals: - Each module should be small enough for AI to fully understand - Clear boundaries so updates don't cascade - Easy to test and debug in isolation My stack: [your language/framework] I'm learning, so explain each part simply.
Extract a Module
For pulling a feature into its own module From the Modular Architecture AI Coding Building Block.
Extract [feature name] into its own isolated module from my current codebase. Current code location: [file path or paste relevant code] Requirements: 1. Create a clean public interface (what other modules can call) 2. Hide implementation details (private functions/classes) 3. Define clear types/interfaces for inputs and outputs 4. Add a module-level docstring explaining its purpose 5. List all external dependencies this module needs 6. Ensure no circular imports with other modules The module should be completely self-contained. I want to be able to: - Ask AI to modify just this module without loading the rest of my codebase - Run tests on this module in isolation - Swap the implementation without changing calling code My stack: [your language/framework] I'm learning, so explain each part simply.
Design Module Interfaces
For designing clean module boundaries From the Modular Architecture AI Coding Building Block.
Design clean interfaces between my modules. I want to minimize coupling so each module can evolve independently. My modules: [list your modules, e.g., auth, billing, notifications, users] For each module boundary: 1. Define the contract (function signatures, types, what's promised) 2. Use dependency injection where appropriate 3. Create interfaces/protocols rather than concrete implementations 4. Document what can and cannot cross module boundaries 5. Suggest a pattern for inter-module communication (events, direct calls, message queue) Goal: If I ask AI to "change how billing works," it should only need to understand the billing module, not touch auth, users, or notifications. My stack: [your language/framework] I'm learning, so explain each part simply.
Add AI-Friendly Module Documentation
For making modules AI-readable From the Modular Architecture AI Coding Building Block.
Add documentation to my module that helps AI coding assistants understand it quickly. Module path: [folder/file path] Create: 1. A MODULE_README.md with: - One-paragraph purpose statement - Public interface summary (what other modules should call) - Example usage snippets - What this module does NOT handle (boundaries) 2. Inline documentation: - Docstrings for all public functions - Type hints for all parameters and returns - Brief comments for non-obvious logic 3. An ARCHITECTURE.md if the module has multiple files explaining: - How files relate to each other - Data flow through the module - Key design decisions Goal: An AI assistant should be able to read these docs and immediately understand how to work with this module without reading every line of code. I'm learning, so explain each part simply.