SKILL.md
FastAPI Development
Table of Contents
- [Overview](#overview)
- [When to Use](#when-to-use)
- [Quick Start](#quick-start)
- [Reference Guides](#reference-guides)
- [Best Practices](#best-practices)
Overview
Create fast, modern Python APIs using FastAPI with async/await support, automatic API documentation, type validation using Pydantic, dependency injection, JWT authentication, and SQLAlchemy ORM integration.
When to Use
- Building high-performance Python REST APIs
- Creating async API endpoints
- Implementing automatic OpenAPI/Swagger documentation
- Leveraging Python type hints for validation
- Building microservices with async support
- Integrating Pydantic for data validation
Quick Start
Minimal working example:
# main.py
from fastapi import FastAPI, HTTPException, status
from fastapi.middleware.cors import CORSMiddleware
from contextlib import asynccontextmanager
import logging
# Setup logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
# Create FastAPI instance
app = FastAPI(
title="API Service",
description="A modern FastAPI application",
version="1.0.0",
docs_url="/api/docs",
openapi_url="/api/openapi.json"
)
# Add CORS middleware
app.add_middleware(
CORSMiddleware,
allow_origins=["http://localhost:3000"],
allow_credentials=True,
allow_methods=["*"],
// ... (see reference guides for full implementation)Reference Guides
Detailed implementations in the references/ directory:
Guide
Contents
FastAPI Application Setup
Pydantic Models for Validation
Pydantic Models for Validation
Async Database Models and Queries
Async Database Models and Queries
Security and JWT Authentication
Security and JWT Authentication
Service Layer for Business Logic
Service Layer for Business Logic
API Routes with Async Endpoints
API Routes with Async Endpoints
Best Practices
✅ DO
- Use async/await for I/O operations
- Leverage Pydantic for validation
- Use dependency injection for services
- Implement proper error handling with HTTPException
- Use type hints for automatic OpenAPI documentation
- Create service layers for business logic
- Implement authentication on protected routes
- Use environment variables for configuration
- Return appropriate HTTP status codes
- Document endpoints with docstrings and tags
❌ DON'T
- Use synchronous database operations
- Trust user input without validation
- Store secrets in code
- Ignore type hints
- Return database models in responses
- Implement authentication in route handlers
- Use mutable default arguments
- Forget to validate query parameters
- Expose stack traces in production