What problem does it solve? Hand-authoring .ipynb files means writing fragile raw JSON, and ad-hoc notebooks often lack reproducible structure. This Skill generates clean, well-structured notebooks from templates so experiments and tutorials stay consistent and runnable. ## Core Features & Use Cases - Template-Based Scaffolding: Generate experiment or tutorial notebooks via the new_notebook.py helper script without touching raw notebook JSON. - Structured Patterns: Apply proven layouts for exploratory analysis (hypotheses, baselines, metrics) and teaching walkthroughs (prerequisites, exercises, pitfalls). - Safe Editing Guidance: Follow references on notebook JSON structure and a quality checklist when refactoring existing notebooks. - Use Case: A data scientist needs to compare prompt variants; run the helper script with --kind experiment to get a reproducible notebook skeleton, then fill in small runnable cells. ## Quick Start Ask the assistant to create a new experiment notebook titled "Compare prompt variants" using the jupyter-notebook scaffolding script.