Python Project Skill | Agent Skills

Python Project

Use this workflow

Build and maintain Python 3.12+ projects with modern tooling and best practices. Follow the steps below for project setup, dependency management, quality tools, CLI patterns, and packaging workflows.

Key tools:

uv: Fast Python package installer and environment manager
ruff: Lightning-fast linter and formatter
pytest: Testing framework with coverage support
ty: Static type checker
typer: Modern CLI framework
loguru: Simplified logging

For Python coding conventions, see the python-conventions skill.

Set up a new project

Initialize a project:

uv init my-project
cd my-project

Add runtime dependencies:

uv add loguru typer

Add development dependencies:

uv add --dev ruff pytest pytest-cov ty

Verify the setup:

uv run python -V
# Confirm the expected Python version

Use a src/ layout

Use a src/ layout for better import clarity and testing isolation:

my-project/
├── src/
│   └── my_project/
│       ├── __init__.py
│       ├── cli.py
│       └── core.py
├── tests/
│   ├── __init__.py
│   └── test_core.py
└── README.md

Benefits of a src/ layout:

Prevent accidental imports from the project root
Force installation for testing
Keep source and tests separated

Manage dependencies

Basic operations:

Add runtime dependency: uv add <package>
Add dev dependency: uv add --dev <package>
Add to named group: uv add --group <name> <package>
Run commands: uv run <command>
Sync dependencies: uv sync

Key principles:

Use uv run <command> instead of plain python or tool commands.
Use --dev for tools that are not needed in production (ruff, pytest, ty).
Pin versions in production and use ranges during development.

Read references/uv-scripts.md for inline metadata, --no-project, and --with flags.

Run quality tools

Lint and format with ruff:

uv run ruff check         # Check for issues
uv run ruff check --fix   # Auto-fix issues
uv run ruff format        # Format code

Type check with ty:

uv run ty check          # Type check all code

Test with pytest and coverage:

uv run pytest                                     # Run tests
uv run pytest --cov=src --cov-report=term-missing # With coverage
uv run pytest -v tests/test_specific.py           # Specific test file

Use this pre-commit quality gate:

uv run ruff check --fix
uv run ruff format
uv run ty check
uv run pytest

Read references/quality.md for tool configuration and recommended settings.

Build CLIs and logging

Use typer for CLIs:

import typer

app = typer.Typer()

@app.command()
def greet(name: str, count: int = 1) -> None:
    """Greet someone multiple times."""
    for _ in range(count):
        typer.echo(f"Hello, {name}!")

if __name__ == "__main__":
    app()

Use loguru for logging:

from loguru import logger

logger.info("Application started")
logger.warning("Low disk space: {free} MB", free=512)

try:
    # Some operation that might fail
    connect_to_service()
except Exception as err:
    logger.error("Failed to connect: {error}", error=err)

Read references/cli-logging.md for complete examples and advanced patterns.

Package distributions

Build distribution packages:

uv build                 # Build wheel and sdist
uv build --no-sources    # Build wheel only (for publish checks)

Check outputs in dist/:

*.whl - Wheel package
*.tar.gz - Source distribution

Read references/packaging.md for publish workflows and checks.

Read references when needed

Script execution patterns: references/uv-scripts.md
Tool configuration: references/quality.md
CLI and logging: references/cli-logging.md
Packaging details: references/packaging.md
Coding style: use the python-conventions skill

References

references/uv-scripts.md - Running scripts with uv
references/quality.md - Ruff, pytest, and ty configuration
references/cli-logging.md - Typer and loguru patterns
references/packaging.md - Build and publish workflows

Agent Skills: Python Project

Skill Files