This includes the functions that I wrote for extracting elements at index paths and by patterns. It's useful for developing mainly and not used to identify the address field. There we use the heuristic with the URL detection as discussed.
Firefox doesn't work w.r.t. the project extraction, but the window title extraction works. It appears opaque to the AX APIs for some reason that I don't understand.
Note: