I agree. Further, to the extent that students need to learn "Jira-fu, unit testing, peer reviewing, etc." they can learn this in a controlled environment run by the university.

The problem with trying to have all students (not "some exceptional students", but all students) going out and engaging with the real world of open source is that the degree of difficulty in navigating different open source projects is enormous and may not be obvious to the professors nor students ahead of time.

Some students may wind up working on projects with nice guard-rails, helpful communities, plenty of up-to-date high-quality documents for how to contribute. Others... maybe not so much. It will be very hard to judge whether students have been successful and to compare their performances against each other (a perennial chore of academia, but one that people seem to insist on: grades!).