quoting con quesa, Layperson, California:

Also I reject your claim that a language having /ɾ/ as it's only liquid is "weeby". Lots of languages lack /l/.

Sure—but these two statements can coexist. The tongue-in-cheek insinuation comes down to a matter of the motivation for only including one liquid/approximant/trill/etc. As I understand it, the Berber languages are generally pretty rich in this area, so the inspiration isn't immediately apparent.

I also think that talking about how the writing system (if one exists) ought to work is a bit premature,

Divergence between a written language and its spoken counterpart can fossilize etymological relationships that would otherwise be less obvious, e.g. any fluent English speaker could intuit a definition for "complication" by back-forming from "relation", "relate", and "complicate", essentially intuiting what "-tion" does. But the homophone "shun" is of no help at all. (And if that doesn't convince you, just think about how many homophones French has...)

In short, I'm arguing against generating morphemes based purely on phonology. Assuming the culture is highly literate, the written form of the language shouldn't match 1:1, and if we were to, for example, just use IPA to create the language and completely forgo any attempt at an official transliteration, the result would be like Tok Pisin, or some other really immature creole—which is very unlikely to have a phonological inventory as rich as the one proposed.