I've learnt that if a test only fails sometimes, it can take a long time for somebody to actually investigate the cause,in the meantime it's written off as just another flaky test. If there really is a bug, it will probably surface sooner in production than it gets fixed.
I dislike the intent too. A website should simply not be able to see which apps I've got installed. Imagine Facebook doing stuff like this in order to know what ads they should serve.
You, as a maintainer, are free to ignore any such expectations and do what you want. There are no obligations. You only risk disappointing people (or corporations), and losing Github stars. If that leads to unmaintained libraries, that probably means the open-source model doesn't work for this project. And that's fine.
Is the consensus that software will be both developed and tested by machines? Will there still be a human in the loop? I hope at least some testing or approval will still be done by people, otherwise the software we use everyday will become even worse than it is now. Unless we envision that machines will also be the only end users of the software. At that point there hopefully will be an interface that allows for immediate reporting and fixing of defects.
The parent comment was talking about a deployed application, and making sure users can't modify the source code.
In that case, your image hopefully was built using source code from a VCS and no such changes will have been applied. This is not so different from building an app by pulling the latest version of the code from Git.
And if you did manage to commit changes that completely mess up your environment, you throw that image away, take a vanilla one and load your source code into it. Again, not so different from using Git.
How do you define cryptography? Let's say my files are written in a format that only my software can read. Is it then illegal to distribute said files?