Terminology and aspiration: assurance cases

[consideRatio]

From the Developing Secure Software course I learned about "Assurance cases":

Creating an assurance case helps you determine and justify to people that the software is secure, both to others and yourself.

I perceive that this subproject is already working towards providing security assurance through systematic security work, but when I learned today about "Assurance case" as a word, it quickly helped me think more clearly and realize that we could provide one or more assurance cases as artifacts of our work.

[consideRatio]

Do we already have examples of assurance cases in the Jupyter ecosystem?

[dlqqq]

I'm not sure, but I do think it would be good for us to contribute more here. JupyterLab's modular architecture means that any installed package can modify JupyterLab, for good or for evil. Some perceive this as a security risk, but I think the argument we have as maintainers is that this is necessary to make JupyterLab as good as it is today.

I definitely would be interested in seeing an assurance case that reads something like "Assuming you are only installing packages you trust, JupyterLab is secure because...".