Jupyter — интерактивный Python REPL для анализа данных

★ 7.1 · data

jupyter-notebook is a Claude Code skill that provides a stateful Python REPL through a live Jupyter kernel, keeping variables, imports, and objects alive across executions — unlike the stateless `execute_code` tool. It relies on the hamelnb script (`jupyter_live_kernel.py`) and exposes commands for `execute`, `variables` inspection, `contents` viewing, cell `edit` (insert, replace, delete), and `restart-run-all` for clean top-to-bottom verification, with all output returned as structured JSON. The skill is the right choice for iterative data exploration, DataFrame inspection, machine learning experimentation, and incremental algorithm development. It requires uv and JupyterLab to be installed and runs on Linux, macOS, and Windows. Use `--compact` to reduce token usage and `--timeout` to extend limits for heavy computations.