This has a lot of batteries included. I include kernels for Python, Julia, R and Racket. I guess you would not want all of these in a typical project. This repo illustrates what is possible.
I have not thoroughly tested all of these.
The Python kernel supports a lot of packages. It is based off of Anaconda which already provides a lot of features, and I have added quite a few more on top of that.
I have made it possible to export your notebooks to PDF via LaTeX and via webpdf.
The Jupyter lab interface is collaborative by default, and the https://github.com/jupyterlab-contrib/jupyterlab-link-share extension is installed. It is easy to invite people who are in your GitHUB organization and who are members of the repo to do real time collaboration. To invite outside people you need to configure the port it is on to be public.
The https://github.com/jupyterlab-contrib/jupyter-videochat extension is also installed, so you can open a video chat in your jupyter lab notebook.
This is a pretty minimal Julia environment. It includes Plots and ForwardDiff.
I include a very minimal R environment. It is mostly a demonstration that you can run R in a Codespace.
I include a minimal Racket environment as a demonstration of running Racket in a Codespace.