py-pit — Claude Code plugin: Python API patterns
A skill pack for Python API work: FastAPI, uv, SQLAlchemy, Alembic. The agent picks the right skill on its own when a task matches.
/plugin marketplace add anthropics/claude-plugins-community
/plugin install py-pit@claude-communityWhat py-pit does
py-pit gives Claude a fixed set of conventions for building Python APIs. Sixteen skills ship inside. fastapi-init scaffolds a service, uv handles environments and lockfiles, sqlalchemy-models and alembic-migrations cover models and schema changes, pydantic-schemas deals with request and response shapes, settings-config replaces scattered environment lookups, and click-cli plus click-cli-linter look after command-line tools. Testing, error handling, Docker Compose, request tracing and outbound API clients each get their own skill too.
Nothing needs to be invoked by hand. Skills fire when the agent reads the intent behind your request. Sitting above them is implementation-protocol, which runs on any non-trivial feature and routes the work to the domain skills. That ordering is the point: one skill decides the sequence instead of you.
The stack is narrow on purpose. Django, Flask or poetry projects will find several skills simply do not apply, because everything here assumes FastAPI, uv and SQLAlchemy 2.x. Worth installing when you are starting a service on that stack, or when you want an existing one brought in line with it.
What it helps with
- scaffold a new FastAPI service from scratch
- move a project off pip or poetry onto uv
- design SQLAlchemy 2.x models and their relationships
- generate an Alembic migration and review the diff
- wire up configuration with pydantic-settings
Example prompts for the agent
- Scaffold a new FastAPI service with an orders endpoint
- Migrate this project's dependencies from poetry to uv
- Design SQLAlchemy models for users and orders, then prepare an Alembic migration
How to install py-pit
Plugin marketplace
/plugin marketplace add anthropics/claude-plugins-community/plugin install py-pit@claude-communityStep by step, with commands for each platform: Claude Code plugins: marketplaces, installing, and building your own
Questions about py-pit
How do I install py-pit in Claude Code?
Add the marketplace first: /plugin marketplace add anthropics/claude-plugins-community. Then run /plugin install py-pit@claude-community.
What does py-pit do?
It supplies Python API development skills covering FastAPI, uv, SQLAlchemy, Alembic, Pydantic, Click, testing and Docker. The implementation-protocol skill coordinates the rest.
Do I have to trigger the skills myself?
No. They activate when the agent recognises a matching task, such as a request to create a service or set up migrations.
Source text
SKILL.md · MIT
# py-pit  Opinionated Python API development skills for AI coding agents (Pi, Claude Code). `py-pit` encodes opinionated Python API development patterns as skills for AI coding agents — works with both **Pi** and **Claude Code**. Covers the modern Python API stack: FastAPI services, uv environments, SQLAlchemy models, Alembic migrations, configuration management, and CLI tooling. ## Install ### Pi Install from npm: ```bash pi install npm:@robhowley/py-pit-skills ``` Or directly from GitHub: ```bash pi install https://github.com/robhowley/py-pit-skills ``` ### Claude Code Add the plugin marketplace, then install py-pit: ``` /plugin marketplace add robhowley/py-pit-skills /plugin install py-pit@py-pit-skills ``` ## Skills Skills activate automatically when the agent detects relevant intent — no manual prompting required. | Skill | Activates when LLM detects | |----------------------|--------------------------------------------------------------------------------------------| | `implementation-protocol` | any non-trivial feature implementation — orchestrates domain skills below | | `fastapi-init` | new FastAPI service, scaffold an API, new microservice | | `uv` | dependency management, lockfiles, env setup, migration from pip/poetry | | `click-cli` | designing or generating a new Click CLI | | `click-cli-linter` | auditing or improving an existing Click CLI | | `pydantic-schemas` | request/response schema design, Pydantic v2 models, schema patterns | | `code-quality` | linting setup, Ruff, pre-commit, code health tooling | | `dockerize-service` | local dev Docker Compose setup, containerizing a project | | `settings-config` | environment variable management, pydantic-settings, replacing os.getenv | | `sqlalchemy-models` | ORM model design, SQLAlchemy 2.x patterns, relationships, migration-ready schema | | `alembic-migrations` | adding Alembic, generating migrations, reviewing autogenerate diffs, safe schema evolution | | `fastapi-errors` | FastAPI error handling, exception hierarchy, consistent API error responses | | `pytest-service` | writing tests for a FastAPI service, SQLAlchemy test fixtures, DI overrides, test setup | | `request-correlation` | request tracing, correlation IDs, observability wiring, propagating context across task/job boundaries | | `http-client-integration` | outbound API calls, building a vendor client, introducing retries and timeouts, replacing ad-hoc HTTP usage | | `background-jobs-boundaries` | background task guardrails, misuse of FastAPI BackgroundTasks, when to use a real job system |
© robhowley · MIT · robhowley/py-pit-skills
