Build Production-Ready Python Backends with FastAPI

This workflow builds a production-ready Python backend with FastAPI across eight phases: project setup, database setup, API routes, authentication, error handling, testing, documentation, and deployment. Each phase names specific sub-skills to invoke and gives copy-paste prompts for invoking them, and the stack is fixed to FastAPI, Python 3.11+, SQLAlchemy 2.0, Pydantic v2, PostgreSQL, Alembic migrations, JWT/OAuth2 auth, and pytest.

Use this workflow when building new REST APIs with FastAPI, creating async Python backends, implementing database integration with SQLAlchemy, setting up API authentication, or developing microservices. It is not a substitute for environment-specific validation, testing, or expert review, and work should pause for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.

Phase 1 (Project Setup) sets up the Python environment (uv/poetry), creates the project structure, configures the FastAPI app, sets up logging, and configures environment variables, invoking app-builder, python-development-python-scaffold, fastapi-templates, and uv-package-manager. Phase 2 (Database Setup) designs the schema, sets up SQLAlchemy models, creates the database connection, configures Alembic migrations, and sets up session management, invoking prisma-expert (as an alternative ORM), database-design, postgresql, and pydantic-models-py. Phase 3 (API Routes) designs endpoints, creates routers, implements CRUD operations, adds request validation, and configures response models, invoking fastapi-router-py, api-design-principles, and api-patterns. Phase 4 (Authentication) chooses a strategy (JWT, OAuth2), implements registration and login, creates auth middleware, and adds password hashing, invoking auth-implementation-patterns and api-security-best-practices. Phase 5 (Error Handling) creates custom exceptions, exception handlers, error responses, request logging, and error tracking, invoking fastapi-pro and error-handling-patterns. Phase 6 (Testing) sets up pytest, fixtures, unit and integration tests, and a test database, invoking python-testing-patterns and api-testing-observability-api-mock. Phase 7 (Documentation) configures the OpenAPI schema, endpoint docs, usage examples, API versioning, and generates docs, invoking api-documenter and openapi-spec-generation. Phase 8 (Deployment) creates a Dockerfile, sets up docker-compose, configures production settings and a reverse proxy, and deploys to cloud, invoking deployment-engineer and docker-expert.

Each phase's skill is invoked with a copy-paste prompt, for example:

Use @fastapi-templates to scaffold a new FastAPI project

The technology stack is fixed: FastAPI on Python 3.11+, SQLAlchemy 2.0 for the ORM, Pydantic v2 for validation, PostgreSQL as the database, Alembic for migrations, JWT/OAuth2 for auth, and pytest for testing. Six quality gates close the workflow: all tests passing with over 80% coverage, mypy type checking passing, ruff/black linting clean, complete API documentation, a passed security scan, and met performance benchmarks. It's related to four neighboring bundles: development, database, security-audit, and api-development.

Backend engineers building a FastAPI service end-to-end - from project scaffolding through database integration, routing, auth, error handling, tests, docs, and deployment - who want a consistent phase-by-phase structure with named sub-skills and ready prompts rather than improvising the build order each time.