I was on a team that tried it. It didn’t work well for us.

Actual issue trackers are far more efficient. Juggling pull requests, code review, and inevitable things like edit conflicts and rebases is a small price to pay for code that people are working on for hours at a time. However, it’s way too much overhead to simply manage issues. It’s not fun to have to type a lot of commands and click around code review to accomplish the same thing I can do in a few seconds with a couple clicks in an issue tracker.

The system also hits a dead end as soon as you want to bring non-coders into the loop. It’s not fun when your managers have to go through a coder to get guidance to interact with the issue tracker.