scaffolding-ml-research-notebook

What problem does it solve?

This Skill helps you start a new machine learning or data-science research project with a reproducible, opinionated directory structure instead of a messy ad hoc notebook folder. It prevents common early mistakes like unpinned dependencies, missing seed handling, accidental data commits, and notebook-output drift.

Core Features & Use Cases

• Creates a greenfield project scaffold with pinned Python, a locked environment, a src package, tests, data/raw and data/processed folders, and a starter notebook.
• Adds a shared seed helper so notebooks and training scripts begin with deterministic behavior from the first cell or first line.
• Installs research-friendly hygiene such as ML-specific gitignore rules, pre-commit checks, and a claudedocs workspace for analysis outputs.
• Use it when you are initializing a new churn model, exploratory analysis repo, coursework research notebook, or any fresh ML project that needs durable structure from day one.
• Refuses to overwrite mature projects, because migration is a separate problem from scaffolding.

Quick Start

Ask the assistant to scaffold a new greenfield ML research project with pinned dependencies, a src layout, data folders, tests, a starter notebook, and reproducibility tooling.