We're not parsing GitHub comments ATM. That's a good idea though, but it would be difficult to track content. Maybe it's just simpler to track issues in one place and improvements in another?