I had always assumed, with no particular evidence, that the frequent Algerian Arabic word rka ركا ("rot", v.) was of some obscure Arabic origin; it looks like a normal Arabic word, after all, with a triliteral root and a weak 3rd consonant and a regular conjugation (although that non-emphatic r is suspicious, in retrospect.) It even has a corresponding adjective, raki "rotten", and there is a verb ركا in Fusha, though its range of meanings ("dig", "fix", "slander"...) show no obvious similarity to "rot". So I was somewhat surprised to see, looking at Kossmann 1999:176, that it occurs throughout Northern and Southern Berber languages, with k shifting to sh in Zenati ones as expected, and can clearly be reconstructed for proto-Berber. A lot of common Algerian Arabic words of obscure origins that I had thought might be from Berber haven't held up to closer examination, but this one looks pretty solid.
So on that note, consider the irregular imperative of "come" in Algerian Arabic: not the impossible *ji, but ayya أيّا. I understand this word is also present as an irregular imperative of as ("come") in Kabyle, Chenoua, and Tumzabt; so does it come from Arabic or Berber? In Arabic, hayyaa هيّا "hurry!" seems a plausible-looking but not indisputable source for it; dropping the h would be irregular, but there are other examples (نوظ "get up", presumably from نهض). So the question hinges on how widely the word yya is distributed among Berber languages. Is it found in Chaoui, for example? Or Tamasheq, or Tashlhiyt, or even Siwa? I'm hoping some readers will be able to help answer these question... :)