| name | PolicyPulse Development-2 |
| description | AI-powered synthetic population simulator with clean architecture, Python type safety, and professional dashboard design patterns |
PolicyPulse Development
This skill guides AI agents in developing PolicyPulse, a professional analytical tool for stress-testing policies against synthetic populations using hybrid AI (LLM + Neural Network).
When to use
Activate this skill when:
- Building or modifying PolicyPulse application code
- Creating modular Python components following clean architecture
- Designing dashboard UI components for Streamlit
- Implementing type-safe Python code with proper annotations
- Writing tests for PolicyPulse functionality
Core Principles
1. Document-Driven Development
- Source of Truth: PRD.md, DESIGN_DOC.md, and TECH_STACK.md are authoritative
- No Feature Invention: Only implement features explicitly defined in documents
- Incremental Progress: Small, safe steps that keep the project runnable
- No Summary Documents: Don't create documentation unless explicitly requested
2. Clean Architecture (Modular Monolith)
Follow the defined module structure:
app.py # Entry point, session management
ui_sections.py # UI component rendering
simulation.py # Simulation orchestration
llm_client.py # Gemini API wrapper
nn_model.py # Neural network training/inference
population.py # Synthetic citizen generation
stats.py # Aggregation and statistics
data_models.py # Dataclass definitions
utils.py # Helper functions
config.py # Environment loading
ml_data.py # Training dataset management
Module Dependency Rules:
data_models.py has no dependencies (stdlib only)
