code-quality

Code Quality Skill (Clean Code for Python & ML/MLOps)

A Claude Code skill that makes Claude apply Clean Code and code-quality best practices when writing, refactoring, or reviewing Python — with a Machine Learning / MLOps slant.

What's inside

This repo is a Claude Code plugin marketplace containing one plugin (code-quality):

.claude-plugin/
├── marketplace.json                # declares this marketplace + its plugins
└── plugin.json                     # the code-quality plugin manifest
skills/
└── code-quality/
    ├── SKILL.md                    # skill entry point
    └── references/                 # deep-dive reference files

File	Purpose
skills/code-quality/SKILL.md	Entry point: when to use it, how to apply it, golden-rule cheat sheet, and a full code-review checklist. Claude reads this first.
skills/code-quality/references/solid.md	The five SOLID principles with Python before/after — including the LSP and ISP examples the source slides leave out.
skills/code-quality/references/clean-code.md	16 Clean Code rules with smell → fix examples (single-purpose functions, naming, few args, no flags, no side effects, polymorphism over conditionals, DRY/KISS/YAGNI, guard clauses, magic numbers, error handling, Python gotchas…).
skills/code-quality/references/design-patterns.md	Factory, Strategy, Adapter, Dependency Injection, Template Method, Builder, Singleton — and when not to use a pattern.
skills/code-quality/references/ml-code-quality.md	ML/MLOps-specific quality: separation of concerns, no data leakage, reproducibility, configuration management, logging & experiment tracking, data validation, efficiency, testing ML code, notebooks vs. production.
skills/code-quality/references/tooling.md	ruff, black, mypy, pytest, pre-commit, bandit — with a ready-to-paste pyproject.toml, .pre-commit-config.yaml, and CI workflow.

The skill uses progressive disclosure: SKILL.md is concise; the detailed references/ files are pulled in only when depth is needed.

How to install it (Claude Code plugin)

In the Claude Code panel (terminal or the VS Code extension), run these two commands:

/plugin marketplace add JaviCeronn/CleanCode_skill
/plugin install code-quality@cleancode-mlops

The first command registers this GitHub repo as a marketplace; the second installs the code-quality plugin from it. No manual file copying. You can also browse interactively by running /plugin and using the menu.

If you added the marketplace before the plugin metadata existed, refresh it first with /plugin marketplace update cleancode-mlops (or remove and re-add it).

Once installed, Claude auto-loads the skill when a task matches its description (writing/refactoring/reviewing Python or ML code; questions about SOLID, clean code, patterns, or quality tooling). You can also trigger it explicitly:

/code-quality review this module

As a plain reference

The Markdown files stand alone — read them directly as a Clean Code / code-quality handbook, or paste a single reference file into any chat as context for a review.

Scope

• Language: English.
• Focus: Python, with Machine Learning / MLOps examples (scikit-learn, pandas, numpy, PyTorch). The principles are general; the examples are ML-flavoured.

Sources & further reading

• Robert C. Martin — Clean Code and Agile Software Development: Principles, Patterns, and Practices (SOLID).
• Gamma, Helm, Johnson, Vlissides — Design Patterns (GoF).
• refactoring.guru — patterns & refactoring catalogue.
• PEP 8 · PEP 20 — Zen of Python.
• scikit-learn Pipeline / ColumnTransformer; Ruff, mypy, pytest, pre-commit docs.