The Agent Skills Directory

[SAFE]: The skill uses well-known, industry-standard Jupyter tooling (jupyterlab, nbconvert, ipykernel) for its primary operations, adhering to standard development practices.
[COMMAND_EXECUTION]: Executes a bundled local Python script scripts/new_notebook.py to scaffold notebooks from pre-defined templates.
[COMMAND_EXECUTION]: Utilizes shell commands for notebook execution (jupyter nbconvert --execute) and environment inspection (jupyter kernelspec list), which are standard for notebook management.
[EXTERNAL_DOWNLOADS]: Recommends the installation of standard Python packages (pandas, matplotlib, jupyterlab) from the official Python Package Index (PyPI) or Homebrew.
[DYNAMIC_EXECUTION]: Uses short python3 -c snippets in documentation to demonstrate programmatic JSON transformations on notebook files, facilitating safe manual editing.
[INDIRECT_PROMPT_INJECTION]: The skill processes and executes existing notebook files, which constitutes an ingestion point for potentially untrusted data.
Ingestion points: references/manual-editing.md (instructions for reading and parsing existing .ipynb files).
Boundary markers: None (the agent processes notebook cells as data without explicit ignore-instruction delimiters).
Capability inventory: Full code execution via jupyter nbconvert, and file-system write access for saving modifications.
Sanitization: No content sanitization or validation of cell source code is performed prior to execution.

jupyter-notebook