Ask HN: Best practice for diagraming the output/side effects of complex scripts?

I'm a SWE, and part of my job right now is migrating a legacy data pipeline to the cloud. Part of that is understanding layers of complex bash scripts that have tons of side effects and do lots of things at once (make 2 copies of this, awk that, over write this, copy it again, rename, append to this one, change the name, append again, copy again, etc).

I'm a very visual person, and my dream goal would be to have some robust diagramming system that allows me to read the script, and, line-by-line, physically diagram it out in such a way that it's very easy for anyone who reads the diagram to understand what's going on.

Some goals I have for this system:

- never having to erase what I've already drawn as I iterate through each line

- the end result makes it easy to spot unnecessary logic

- borrows a lot of symbolism from existing diagrams (state machines, UML/Sequence/etc)

Does this exist? Do I need to invent this? Does this seem useful?