I want notebook inputs and outputs to be detected automatically so I don't make a mistake

We could use static analysis and/or a watchdog for outputs. It's possible we could read from the kernel operations if it's Python.

The issue proposes automatically detecting notebook inputs and outputs using static analysis, a watchdog, or Python kernel operations. Start by identifying the project's notebook workflow and how Python kernel operations are exposed; done means inputs and outputs are detected reliably without requiring manual declarations.

Written by the indexing model from the issue text.