/notebook
<!-- ~ Copyright (c) 2024- Datalayer, Inc. ~ ~ BSD 3-Clause License -->
How it fires
How this command gets triggered: by you, by Claude, or both.
- Fires itselfClaude auto-loads it when your prompt matches the work.
- You can call itInvoke it directly when you want it.
- Slash command
/notebook
Context preview
What this command does when you run it.
<!-- ~ Copyright (c) 2024- Datalayer, Inc. ~ ~ BSD 3-Clause License -->
Command definition
notebook.md<!-- ~ Copyright (c) 2024- Datalayer, Inc. ~ ~ BSD 3-Clause License -->
--- description: Open a Datalayer notebook and work in it argument-hint: [notebook name or path] ---
Connect to a Datalayer notebook and make it the one we work in.
Steps:
1. Call `list_notebooks` to see what this account can reach. 2. If `$ARGUMENTS` is empty, show the list and ask which notebook to open. Otherwise pick the notebook whose name or path best matches `$ARGUMENTS`; if several match, list the candidates and ask rather than guessing. 3. Call `use_notebook` with a short, stable alias and the notebook's path. 4. Call `read_notebook` with the brief format and summarise what the notebook contains — how many cells, what it appears to be about, and where its last execution stopped.
Do not create a notebook unless you were asked to. If nothing matches, say so and list what is available.
An MCP server developed for AI to connect and manage Jupyter Notebooks in real-time — and scale your Code Sandbox from local to the cloud (Datalayer, Kaggle, Google Colab, Modal, Daytona, E2B, CoreWeave, Cloudflare...)
Repo: datalayer/jupyter-mcp-server

