Yes, in fact it even seems to work with FRLG.
As far as I can tell I get exactly the same ARM dump.
As for the Ageto Pikachu, I wasn't able to test how it is written to the GBA but unlike what I had thought the two specimen I know of (PID 1568718567, spread 28/24/7/24/14/31 and PID 3836930036, spread 22/25/22/18/18/14) do not use any simple variation of known GBA/GC PID methods. I tried looking for more legitimate samples but they are hard to find and hard to make without a GC and writeable memory card.
I tried patching the disc to use the Pikachu function at 8012d5dc instead of the Celebi function at 8012d4f0 when writing to the Colosseum save and the result is a perfectly fine Ageto Pikachu, except with a standard GC PID.
Since there's is no example of a legal Ageto Celebi with a non-standard GC PID (among more than 150, some of which must have been written to the GBA) I am really wondering how the bonus disc writes to GBA saves. I have serious doubts Ageto Pokemon are actually made on the GBA, but if that's not the case then why does Celebi seem to get a GC PID but Pikachu doesn't?
Was it ever confirmed whether or not the Celebi sent directly to the GBA can be shiny?
likely not shiny ever, as either GC/GBA receiving saves probably have the celebi generated on the gamecube. Why have two separate versions of code that generate the same thing?
Could this workwith Pokemon Channel to see if you can recive a shiny Jirachi?