Jupyter notebooks: scaffold for reproducible work

Choose experiment or tutorial mode, generate from a template, and validate the .ipynb.

Skills

Use this workflow only for a Jupyter notebook request. Decide whether the notebook is an exploratory experiment or a reader-facing tutorial, then confirm its objective, audience, and output path. Run the bundled `scripts/new_notebook.py` with the experiment or tutorial template to create valid notebook JSON, then add small focused code cells and short explanations. When editing an existing notebook, preserve its order and intent and make targeted improvements. If the required kernel is available, execute top to bottom and inspect results; otherwise state that execution could not be verified. JupyterLab and ipykernel are optional environment changes, so follow the project’s existing uv or environment-management rules.

Permissions

Prerequisites

The user asks to create, structure, or edit an `.ipynb` experiment, analysis, or tutorial; choose experiment or tutorial mode.
Confirm output location, audience, and runnable environment while preserving existing notebook intent.

Permissions

May write a local notebook from a template using the standard-library helper; running it executes cell code.
Installing optional Jupyter or ipykernel packages changes the local environment and should follow an agreed setup.

Limitations

The scaffold helper sets notebook structure and title, not data correctness. If execution is unavailable, report that it was not run and preserve existing cell intent.