jupyter-live-kernel

A stateful Python REPL keeps variables persistent across code executions by interfacing with a live Jupyter kernel. This allows iterative data exploration and ML experimentation without restarting the session or losing loaded datasets.

Yes, you can execute code iteratively in a Jupyter kernel via CLI commands. The skill interfaces with a running kernel session to run code, inspect variables, and edit notebook cells for immediate feedback during development.

Yes, you need uv to manage servers and a JupyterLab-enabled environment to run the stateful Python REPL. A running kernel session is also required to interface with the live Jupyter kernel via hamelnb.

The best way to inspect variables during ML experimentation without restarting is using CLI commands to interact with a stateful Jupyter kernel. This live kernel session preserves all loaded data and model states across execution steps.

Use a live Jupyter kernel REPL when building complex code step-by-step for data science or API exploration. It provides stateful execution and immediate feedback, unlike standard scripts that restart state and lose variables on every run.

A stateful REPL session loses variables when the Jupyter kernel restarts because all in-memory state is cleared. Variable persistence relies on a continuously running kernel session via hamelnb to maintain the Python environment.