I wrote a program at work that gets deployed to hundreds of thousands of systems and is very hard to fully test or instrument. This program recently had a bug that was hard to track down. Using the command line, I connected to one of these boxes over ssh and ran a series of commands to detect the bug and dump details of what happened. Then, I took all those commands and turned them into a onliner that I could pass in over ssh, so I could get everything I needed for an individual maxhine. I then used xargs to run that command in parallel over every single one of the systems my code was running on and in the end, I was left with a nice directory of files whose name was the IP of an affected system, each filled with useful information. I started by manually running command over ssh, but the composable nature of the shell allowed me to transition that into a script in a matter of minutes.

I provided a more residential example of why I exclusively use the terminal for file management in a different top level comment.