Would i need to modify the code of every services to man in the middle myself surly since i control the private key i can just do it after it leaves the service then if i pipe that back to my “universal” incoming it would distribute it to the relevent services?

If you do a perfect MitM (which is quite hard and requires constant modifications as software you’re MitMing updates) you wouldn’t need to change anything about the applications you’re running. I don’t think that’s very viable, though.

I dont understand why the last approach breaks mastadon-lemmy interaction? It seems like the best solution if it wasnt for that.

The way Webfinger works is that a remote server fetches details about the correct destination for messages once and then uses that to post a bunch of updates. If you tell every Lemmy server to post to user@lemmy.server, the Lemmy servers wouldn’t be able to post to user@mastodon.server.