A Claude Code marketplace providing specialized skills for working with the CPython repository - the reference implementation of the Python programming language.
This marketplace hosts Claude Code plugins with AI-agent skills for CPython development. Currently includes one comprehensive plugin with multiple skills providing deep knowledge about CPython's codebase structure, build system, testing infrastructure, and coding standards.
To use these skills in Claude Code within a CPython repository checkout:
-
Add the marketplace:
/plugin marketplace add gpshead/cpython-skills -
Install the plugin:
/plugin install cpython-skills -
In your CPython repository, ask Claude:
write a CLAUDE.local.md that instructs you to load and use all of the cpython-skills plugin in this repo.
Name: cpython-skills
Description: Skills for working with the CPython repository - helping with building, testing, and contributing to Python's implementation
This plugin provides four specialized skills:
When to use: Working in the CPython repository, understanding codebase structure, managing engineering notebooks
Provides:
- Recommended tools (
rg,gh,jq) - Source code organization (Lib/, Modules/, Objects/, Python/, Include/)
- Test location and naming conventions
- Argument Clinic overview
- Engineering notebook management for PRs and branches as task-specific long term memory
- Scratch space usage
- Developer guide and PEP references
When to use: Configuring, building, rebuilding CPython, running tests, collecting coverage, debugging test failures
Includes:
- Build directory setup
- Platform-specific configuration (Linux, macOS, Windows)
- Incremental builds
- Argument Clinic code generation
- Build verification and troubleshooting
- unittest-based testing (not pytest!)
- Running individual tests and test suites
- Using
--matchfor test filtering - Code coverage collection
- Test package handling
- Interactive debugging with tmux
When to use: Writing code, ensuring style compliance, preparing commits
Covers:
- PEP 7 (C code) and PEP 8 (Python code) compliance
- Trailing whitespace and file ending rules
- Type annotation policy (no annotations in Lib/!)
- Pre-commit hooks and patchcheck
- Documentation formatting
When to use: Working with CPython documentation, adding version markers, creating NEWS entries
Covers:
- Documentation in ReST format in Doc/ tree
- Documentation tooling setup and validation
- Version markers (always use
next) - NEWS file entries for bug fixes and features
These tools improve the agentic development experience: rg, gh, jq
cpython-skills/
├── .claude-plugin/
│ └── marketplace.json # Marketplace manifest
├── plugins/
│ └── cpython-skills/ # CPython skills plugin
│ ├── plugin.json # Plugin manifest
│ └── skills/ # Individual skills
│ ├── cpython-codebase/
│ │ └── SKILL.md
│ ├── cpython-build-and-test/
│ │ └── SKILL.md
│ ├── cpython-code-style/
│ │ └── SKILL.md
│ └── cpython-docs/
│ └── SKILL.md
├── LICENSE
└── README.md # This file
To add a new plugin to this marketplace:
- Create a new directory under
plugins/with your plugin name - Add your plugin files (skills, commands, agents, etc.)
- Create a
plugin.jsonmanifest in your plugin directory - Update
.claude-plugin/marketplace.jsonto include your plugin entry
Example marketplace.json plugin entry:
{
"name": "your-plugin-name",
"source": "./plugins/your-plugin-name",
"description": "Brief description of what your plugin does"
}This marketplace follows the Claude Skills specification, making plugins usable by any AI agent that supports this format.
Each skill is a self-contained directory with a SKILL.md file containing:
- YAML frontmatter with
nameanddescription - Markdown content with detailed instructions
When working with CPython, the AI agent will automatically:
- Load
cpython-codebasewhen starting work in the repo - Apply
cpython-build-and-testknowledge when compiling or running tests - Enforce
cpython-code-stylerules when writing code - Follow
cpython-docsguidance when updating documentation
This marketplace was created from the original CPython skills plugin, which was converted from the cpython/CLAUDE.local.md file. The marketplace-based approach provides:
- Modularity: Each skill focuses on a specific domain
- Discoverability: AI agents can find relevant skills based on task descriptions
- Reusability: Skills can be used independently or together
- Extensibility: Easy to add new plugins to the marketplace
- Interoperability: Works with any agent supporting Claude Skills format
CC0 1.0 Universal