My gut feeling was that a standard accessibility protocol doesn't exist yet, but that a Wayland extension is probably the solution.

Looks like I was right: https://blogs.gnome.org/a11y/2024/06/18/update-on-newton-the-wayland-native-accessibility-project/

However that's a recent proof-of-concept, so nothing generally available at this stage. Details would have to be ironed out, changes upstreamed, the protocol extension accepted, and every graphics toolkit would have to be updated to support the protocol. Seems like a very worthy goal to pursue though and it's great to see funds allocated for a11y.

I don't know how that would work with XWayland apps, but given how clunky that gets with stuff like fractional scaling I wouldn't get my hopes up. The sooner we can native Wayland all the things, the better. X11 must die, regardless of the feeling of the graybeards on the matter.

Thanks OP; now I need this 😜.

For the last year or so, I've been using Vimium and mouseless. For the latter, I'd probably use warpd or wl-kbptr instead if I could get them to work on Fedora Silverblue.