cf. Latin capra, Spanish cabra. In historical linguistics, c's and g's are closely related (the latter voiced, the former unvoiced); ditto f's and b's. I wouldn't be surprised if both cabbyl and gafr both relate to, or to a coeval cognate of the Latin capra.
For completeness' sake, the Irish and Scottish Gaelic (the closest languages to Manx) cognates are both capall (though the ScG has undergone a bit of sense narrowing, so it just refers to colts now - the normal word being each, which is also valid in Irish). The Welsh, ceffyl as qroqqa mentions, is also cognate, obviously, but I'm not aware of a cognate in the other two Brythonic languages, Breton and Cornish; the usual words are marc'h and margh respectively (and Welsh has march). If anyone knows of cognates, I'd be interested.
As for "gaffran", yarb, I don't know it and neither does my Welsh dictionary. The plural of gafr is geifr.
The normal Welsh for "goat" is gafr. The interrelationships between the Celtic and Italic words here are not obvious: Don Ringe says the Celtic original of ceffyl "horse" seems to have gone from Latin to Celtic because it doesn't match up the way it would if it was a cognate, but this is uncertain. It is also unclear whether Italic and Celtic formed a sub-branch of IE and therefore particularly shared inherited vocabulary.
Manx showing a Latin gleaning here. I noticed recently that the welsh for "goat" was gaffran, which is presumably linked to the Latin root. I wonder how big the Latin impact on these Celtic languages was, and what the equivalent term was before the Romans came along.