Exposes the full pypreset CLI through MCP, letting Claude scaffold Python projects from YAML presets or augment existing ones with CI workflows, tests, Docker configs, and documentation. It handles project creation with Poetry, uv, or setuptools, generates GitHub Actions workflows for testing and linting, and can add components like Codecov integration, MkDocs or Sphinx docs, tox configs, and devcontainer setups. The augment command is smart enough to detect your existing tooling from pyproject.toml and fill in the gaps. It also wraps act for local workflow verification and manages PyPI metadata. Reach for this when you're bootstrapping Python projects or want Claude to add missing infrastructure without manually writing boilerplate configs.
A meta-tool for scaffolding Python projects with configurable YAML presets.
Supports Poetry, uv, and setuptools, generates CI workflows, testing scaffolds, type checking configs, and more.
mcp-name: io.github.KaiErikNiermann/pypreset
src/ layout and flat layout.dockerignore, and VS Code devcontainer configs (Docker or Podman).python-version for pyenv and uv, with python-version-file in CI workflowsgh CLIact (auto-detect, auto-install, dry-run and full-run modes)pyproject.toml metadata~/.config/pypreset/config.yamlpip install pypreset
# With MCP server support
pip install pypreset[mcp]
# Create a CLI tool project with Poetry
pypreset create my-cli --preset cli-tool
# Create a data science project with uv
pypreset create my-analysis --preset data-science --package-manager uv
# Create an empty package with src layout (default)
pypreset create my-package --preset empty-package
# Create a Discord bot
pypreset create my-bot --preset discord-bot
# Create a project with Docker support
pypreset create my-service --preset cli-tool --docker --devcontainer
# Create with .python-version for pyenv/uv
pypreset create my-lib --pyenv --python-version 3.13
# Create with Podman, Codecov, docs, and tox
pypreset create my-project --preset empty-package \
--container-runtime podman --docker \
--coverage-tool codecov --coverage-threshold 80 \
--docs mkdocs --docs-gh-pages \
--tox
create -- Scaffold a new projectpypreset create <name> [OPTIONS]
| Option | Description |
|---|---|
--preset, -p | Preset to use (default: empty-package) |
--output, -o | Output directory (default: .) |
--config, -c | Custom preset YAML file |
--package-manager | poetry or uv |
--layout | src or flat |
--type-checker | mypy, pyright, ty, or none |
--typing | none, basic, or strict |
--python-version | e.g., 3.12 |
--testing / --no-testing | Enable/disable testing scaffold |
--formatting / --no-formatting | Enable/disable formatting config |
--radon / --no-radon | Enable radon complexity checking |
--pre-commit / --no-pre-commit | Generate pre-commit hooks config |
--bump-my-version / --no-bump-my-version | Include bump-my-version config |
--extra-package, -e | Additional packages (repeatable) |
--extra-dev-package, -d | Additional dev packages (repeatable) |
--docker / --no-docker | Generate Dockerfile and .dockerignore |
--devcontainer / --no-devcontainer | Generate .devcontainer/ configuration |
--container-runtime | docker or podman |
--coverage-tool | codecov or none |
--coverage-threshold | Minimum coverage % (e.g., 80) |
--docs | sphinx, mkdocs, or none |
--docs-gh-pages / --no-docs-gh-pages | Generate GitHub Pages deploy workflow |
--tox / --no-tox | Generate tox.ini with tox-uv backend |
--pyenv / --no-pyenv | Generate .python-version and use python-version-file in CI |
--git / --no-git | Initialize git repository |
--install / --no-install | Run dependency install after creation |
--dry-run | Preview what would be created without generating anything |
augment -- Add components to an existing projectAnalyzes pyproject.toml to auto-detect your tooling, then generates the selected components. Runs in interactive mode by default (prompts for values it can't detect); use --auto to skip prompts.
pypreset augment [path] [OPTIONS]
Available components:
| Flag | Component | What it generates |
|---|---|---|
--test-workflow / --no-test-workflow | Test CI | GitHub Actions workflow that runs pytest across a Python version matrix |
--lint-workflow / --no-lint-workflow | Lint CI | GitHub Actions workflow for ruff, type checking, and complexity analysis |
--dependabot / --no-dependabot | Dependabot | .github/dependabot.yml for automated dependency updates |
--tests / --no-tests | Tests directory | tests/ with template test files and conftest.py |
--gitignore / --no-gitignore | Gitignore | Python-specific .gitignore |
--pypi-publish / --no-pypi-publish | PyPI publish | GitHub Actions workflow for OIDC-based publishing to PyPI on release |
--dockerfile / --no-dockerfile | Docker | Multi-stage Dockerfile and .dockerignore (Poetry, uv, or setuptools aware) |
--devcontainer / --no-devcontainer | Devcontainer | .devcontainer/devcontainer.json with VS Code extensions |
--codecov / --no-codecov | Codecov | codecov.yml configuration |
--docs | Documentation | Sphinx or MkDocs scaffolding (--docs sphinx or --docs mkdocs) |
--tox / --no-tox | tox | tox.ini with tox-uv backend for multi-environment testing |
--readme / --no-readme | README | README.md generated from the shared template (badges, install, features) |
--pyenv / --no-pyenv | pyenv | .python-version file for pyenv and uv version pinning |
# Interactive mode (prompts for missing values)
pypreset augment ./my-project
# Auto-detect everything, no prompts
pypreset augment --auto
# Generate only specific components
pypreset augment --test-workflow --lint-workflow --gitignore
# Add Docker and devcontainer
pypreset augment --dockerfile --devcontainer
# Add PyPI publish workflow
pypreset augment --pypi-publish
# Add documentation scaffolding
pypreset augment --docs mkdocs
# Generate a README from your project metadata
pypreset augment --readme
# Overwrite existing files
pypreset augment --force
workflow -- Local workflow verificationVerify GitHub Actions workflows locally using act. The proxy auto-detects whether act is installed, can install it on supported systems, and surfaces all act output directly.
# Verify all workflows (dry-run, no containers)
pypreset workflow verify
# Verify a specific workflow file
pypreset workflow verify --workflow .github/workflows/ci.yaml
# Verify a specific job
pypreset workflow verify --job lint
# Full run (executes in containers, requires Docker)
pypreset workflow verify --full-run
# Auto-install act if missing
pypreset workflow verify --auto-install
# Pass extra flags to act
pypreset workflow verify --flag="--secret=GITHUB_TOKEN=xxx"
# Check if act is installed
pypreset workflow check-act
# Install act automatically
pypreset workflow install-act
Supported auto-install targets: Arch Linux (pacman), Ubuntu/Debian (apt), Fedora (dnf), macOS/Linux with Homebrew. Other systems get a link to the act installation page.
version -- Release managementpypreset version release --bump patch # 0.1.0 -> 0.1.1
pypreset version release --bump minor # 0.1.0 -> 0.2.0
pypreset version release --bump major # 0.1.0 -> 1.0.0
pypreset version release-version 2.0.0 # Explicit version
pypreset version rerun <ver> # Re-tag and push an existing version
pypreset version rerelease <ver> # Delete and recreate a GitHub release
Requires the gh CLI to be installed and authenticated.
metadata -- PyPI metadata managementpypreset metadata show # Display current metadata
pypreset metadata set --description "My cool package" # Set description
pypreset metadata set --github-owner myuser # Auto-generate URLs
pypreset metadata set --license MIT --keyword python # Set license and keywords
pypreset metadata check # Check publish-readiness
badges -- Generate badge markdownReads pyproject.toml to detect your project name, repository URL, and license, then prints badge markdown you can paste into your README.
pypreset badges # Badges for current directory
pypreset badges ./my-project # Badges for a specific project
pypreset list-presets # List all available presets
pypreset show-preset <name> # Show full preset details
pypreset validate [path] # Validate project structure
pypreset analyze [path] # Detect and display project tooling
pypreset config show # Show current user defaults
pypreset config init # Create default config file
pypreset config set <key> <value> # Set a config value
Built-in presets: empty-package, cli-tool, data-science, discord-bot.
Presets are YAML files that define metadata, dependencies, directory structure, testing, formatting, and more. They support single inheritance via the base: field. Presets can override the README template by setting metadata.readme_template to a custom .j2 filename.
Place custom preset files in ~/.config/pypreset/presets/ or pass a file directly:
pypreset create my-project --config ./my-preset.yaml
User presets take precedence over built-in presets with the same name.
Persistent defaults are stored at ~/.config/pypreset/config.yaml and applied as the lowest-priority layer (presets and CLI flags override them).
pypreset config init # Create with defaults
pypreset config set layout flat # Set default layout
pypreset config set type_checker ty # Set default type checker
pypreset config show # View current config
pypreset is published to the MCP Registry as io.github.KaiErikNiermann/pypreset.
Install via the registry (recommended):
# Claude Code
claude mcp add pypreset -- uvx --from "pypreset[mcp]" pypreset-mcp
# Or add manually to ~/.claude/settings.json
{
"mcpServers": {
"pypreset": {
"command": "uvx",
"args": ["--from", "pypreset[mcp]", "pypreset-mcp"]
}
}
}
Or install locally:
pip install pypreset[mcp]
{
"mcpServers": {
"pypreset": {
"command": "pypreset-mcp",
"args": []
}
}
}
Available tools:
| Tool | Description |
|---|---|
create_project | Create a new project from a preset with optional overrides |
augment_project | Add CI workflows, tests, Docker, docs, and more to an existing project |
validate_project | Check structural correctness of a project directory |
verify_workflow | Verify GitHub Actions workflows locally using act |
list_presets | List all available presets with names and descriptions |
show_preset | Show the full YAML configuration of a specific preset |
get_user_config | Read current user-level defaults |
set_user_config | Update user-level defaults |
set_project_metadata | Set or update PyPI metadata in pyproject.toml |
generate_badges | Generate badge markdown links from project metadata |
Resources: preset://list, config://user, template://list
Prompts: create-project, augment-project
All tasks use the Justfile:
just install # Install dependencies
just test # Run tests
just test-cov # Tests with coverage
just lint # Ruff check
just format # Ruff format
just typecheck # Pyright
just radon # Cyclomatic complexity check
just check # lint + typecheck + radon + test
just all # format + lint-fix + typecheck + radon + test
See CONTRIBUTING.md for development setup and guidelines.
MIT