Changes saved to files in /code and /data  while in a Cloud Workstation session will be automatically synced back to the capsule you launched the session from.
​
This will not be the case for dependencies (libraries). You should note down any dependencies installed during a Cloud Workstation session and re-add them to your environment once you're back in the non-interactive capsule interface.

This article covers instructions for recording this via python and R commands.

From a command line, use either pip freeze or conda list to record all dependencies, which you can then compare to the ones installed in your environment.
​
You can save these results to concrete files for later use as such:
​conda list --export > /code/environment.yml
or
​pip freeze >> /code/requirements.txt 

From an R prompt, use either sessionInfo()  or devtools::session_info()  (Use the devtools  command if you have devtools installed -- R environments typically come with this installed.). 

You can save this to a text file with something like:
​writeLines(capture.output(sessionInfo()), "/code/meta_session_info.txt")