I'd like to decline this task. This is basically the same as T972 and T109566: Getting more reports from people who will never reply to any followup questions in Phabricator as they do not have an account in Phabricator and hence do not get any updates from Phabricator.

From my point of view, I often don't even notice the issues posted in github because phab is my main workflow. I mostly want a task for checking out the github issue and trying to address it. I think I'd be OK with people not coming to phab so long as working with them is tracked in phab.

I feel like the only other alternative is to turn off issues in github. But then we have issues not reported as well as people on github not coming to phab.

Indeed, that could be solved by disabling reporting issues in Github so people have to report issues somewhere else.

But then we have issues not reported as well as people on github not coming to phab.

The argumentation could be expanded to people who don't want to create an account neither on Github nor in Wikimedia Phabricator, etc. :)
It boils down to "How many places would you want to actively monitor?" Or more evil: "In how many places would you like to fragment and duplicate conversations?"

I guess github is important due to its market dominance. It's also particularly relevant for many of our libraries as we list the primary repo (not just a mirror) on github to take advantage of that market dominance.

Could be, however than we are back to T253388#6158680. :)

I figured it would be a bot filing the tasks and that the form of the task would essentially be "Go review this issue on github and copy it here if necessary".

(Removing Phabricator and Wikimedia-GitHub as this would not require changes (in my understanding) in these areas if some bot used APIs of both systems)