jupyter-notebook-writing — Claude Code skill: Markdown-first notebooks
jupyter-notebook-writing — a skill for editing Jupyter notebooks in Markdown instead of JSON. Runs on Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, and OpenCode; edit .md files first, then convert to .ipynb using jupyter-switch.
no license 3,900 installs updated 2026-08-28
npx skills add zc277584121/marketing-skills --skill jupyter-notebook-writing -a claude-codeWhat jupyter-notebook-writing does
Jupyter notebooks stored as .ipynb are JSON with metadata, execution counts, and cell outputs — difficult for AI to edit directly. This skill flips the workflow: write examples in clean Markdown where Python code blocks become code cells and everything else becomes Markdown cells. After editing the .md file, run uvx jupyter-switch example.md to generate the notebook; Markdown stays as the source of truth for future edits.
The skill walks through running notebooks: detect available Python environments (conda, mamba, venv, registered Jupyter kernels), present them to the user, register a new kernel if needed, then run jupyter execute with the right kernel name. Before execution, comment out cells with pip install and API key placeholders — these must be installed and set outside the notebook to avoid test failures.
The structure template covers Prerequisites, Setup, data preparation, collection creation, queries, and cleanup. Focused on Milvus application examples: RAG, semantic search, hybrid search.
What it helps with
- edit notebooks without managing JSON and metadata
- keep Markdown documentation and
.ipynbin sync - pick the correct Python environment to run a notebook
- run a notebook in CI without conflicts from setup cells
- quickly create a Milvus application example
Example prompts for the agent
- Create a Markdown file with a Milvus RAG example and convert it to a notebook
- Edit an existing
.ipynb, convert to Markdown, revise the text, convert back - Run a notebook in the right environment by selecting a Jupyter kernel from a list
How to install jupyter-notebook-writing
1Skills CLI (Vercel)
npx skills add zc277584121/marketing-skills --skill jupyter-notebook-writing -a claude-code2Manually: put the files in place
~/.claude/skills/jupyter-notebook-writing/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/jupyter-notebook-writing/
1Skills CLI (Vercel)
npx skills add zc277584121/marketing-skills --skill jupyter-notebook-writing -a codex2Manually: put the files in place
.agents/skills/jupyter-notebook-writing/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add zc277584121/marketing-skills --skill jupyter-notebook-writing -a cursor2Manually: put the files in place
.cursor/skills/jupyter-notebook-writing/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install zc277584121/marketing-skills skills/jupyter-notebook-writing2Manually: put the files in place
.github/skills/jupyter-notebook-writing/SKILL.md1Skills CLI (Vercel)
npx skills add zc277584121/marketing-skills --skill jupyter-notebook-writing -a gemini-cli2Platform CLI
gemini skills install https://github.com/zc277584121/marketing-skills --path skills/jupyter-notebook-writingManually: put the files in place
.devin/skills/jupyter-notebook-writing/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add zc277584121/marketing-skills --skill jupyter-notebook-writing -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about jupyter-notebook-writing
How do I install jupyter-notebook-writing?
Choose your platform. For Claude Code: npx skills add zc277584121/marketing-skills --skill jupyter-notebook-writing -a claude-code, then open ~/.claude/skills/jupyter-notebook-writing/SKILL.md. For GitHub Copilot: gh skill install zc277584121/marketing-skills jupyter-notebook-writing. For other platforms, use npx skills add with -a codex, -a cursor, -a gemini-cli, or -a opencode.
Why edit in Markdown instead of .ipynb directly?
.ipynb files are JSON with metadata, execution counters, and cell outputs—messy for AI editing. Markdown is clean: Python code blocks automatically become code cells, and everything else becomes Markdown cells. The Markdown file serves as the source of truth for all future edits.
What are the requirements to use this skill?
Python >= 3.10 and the uvx command (part of the uv tool). The skill uses jupyter-switch to convert between .md and .ipynb, and jupyter execute to run notebooks. Install dependencies (pymilvus and others) in your chosen Python environment before running, not inside the notebook itself.