I just proposed this task for the 2022 Community Wishlist, see https://meta.wikimedia.org/wiki/Community_Wishlist_Survey_2022/Watchlists/Cascade-watching

Athena skin is being archived, see T299322.

Athena skin is being archived, see T299322.

Regarding the second question whether to ask for confirmation or not, I
would normally vote no, but given that we're already going to show the
diff, then it might aswell contain a text like: Do you want to send public
thanks to User for the following edit/comment?

Hm, yes a bot may be simpler for this particular use case, but looking
beyond, I think a parser function to get page views would open up
possibilities for other developers and template editors across projects. I
think the movement could benefit a lot from more and better tools for
prioritizing work and motivating editors, and a simple way to get page
views to build templates, modules, tables and other tools from them could
go a long way to facilitate prioritization and motivation. That being said,
if you decide not to implement this, no problem I can do the bot for my use
case. :-) Cheers !

The main use case I have in mind for this feature is to build tables of articles sorted by page views so as to prioritize my work and also help others prioritize theirs. (I've been using https://pageviews.toolforge.org/ so far but having the tables on-wiki would make all the difference for new users).

I believe this functionality doesn't exist in the core watchlist. The closest thing is that little number showing the diff in bytes, that when adds up to zero, usually means vandalism-revert. If you implement the byte-count diff as in the core watchlist, then maybe it wouldn't be too difficult to add a filter to the settings to "Hide groups of changes by two or more users that add up to zero-bytes". Notice that changing a single character for another, or a word for another of the same length results in a zero-byte diff, so the filter should probably match groups of changes by two or more users that add up to zero-bytes to minimize the chances of overlooking vandalism. Then the only vandalism that would be overlooked would be when some vandal deletes some content, and then some other user adds some content of exactly the same number of bytes. This is so unlikely that chances are the only way it'd happen would be by some coordinated attack, which would make little sense since other users not using the global watchlist or this filter would easily spot it.

This would be useful also for non-Wikimedia wikis linking to Wikipedia.

Besides RfA and RfCU (the original use cases mentioned on this task description) and mainspace pages with excerpts, other pages that would benefit from cascade-watching would be the Village Pumps of some Wikipedias, such as the French and Portuguese ones, where transcluding subpages is very common (actually, totally standard). See for example here or here. Also, having this feature available may cause other wikis to restructure their Village Pumps (or other pages) in more efficient ways, since I can easily imagine not doing such restructurings due to the lack of cascade-watching. I'm tagging the Editing Team because this task now seems to me relevant to their more general effort to improve communication within the community.

Somewhat related: T55525

This "cascade-watching" feature would also help with mainspace articles that transclude content from other pages, a practice that is gaining momentum (see https://en.wikipedia.org/wiki/Category:Articles_with_excerpts). One concern about this practice, especially when used on Featured and Good articles, is that watchers of the main article are not notified of changes to transcluded content.

Also perhaps not confusing crawlers and other programs.