Fascinating. I add:

war>crīgā(gave german 'krieg' and danish 'krig')
(gave -possibly- 'cri' and 'crier', French "shout", noun and verb)
fight>skermō(gave the word 'skirmish')
die>dhewə (gave Deyja(old norse))
(and this word Deyja -possibly- gave 'déjà', French "already")
kill>colḗjō (nothing to explain here)
(gave -possibly- colère, French "anger")
destroy>dheukō (here neither...)
(gave -possibly- dragon, if you pronounce the "h" like in German)
Are you claiming these etymologies to be true, or is this just conjecture? Several are almost certainly completely false, and the rest seem dubious.

For example on Goidelic, which is indeed a form of Celt, what date would one impose for a P-shift?
I don't get this question. Seeing as Irish is Q-Celtic, the proto-celtic labiovelar became a plain velar sound, rather than a labial sound, as in Gaulish and Brythonic. So there was never a P-shift in Goidelic.

Linear B Greek has the Kw sound intact, in Ancient Greek it's become P.
This, presumably, accounts for the difference between the Ancient Greek hippos and the Latin equus? It is interesting that the labiovelar stop>labial stop shift seemingly occurred in what appeared to be separate IE families. I had always believed this to be unique to P-Celtic, but after a little research it also appears to have happened in Oscan as well as Greek.

As a side note, how exactly do you reconstruct the syntax and morphology of proto-gemanic? (I assume that this is what is used for the Sweboz) Is it just a case of making educated guesses based on the current syntax in Germanic languages? Was there even a fixed word order? (SOV would have been my guess)