Koko wrote:/i/ was dropped to /ɪ/, and in turn this was dropped to /e/.
Not very naturalistic. There's a very strong tendency for languages to keep the "corners" of the vowel triangle occupied, since this maximises the perceptual distance between phonemes. If /i/ dropped to /e/, I would expect some other phoneme to move into the space vacated by /i/. (Consider English: When /iː/ diphthongised to /əi/, /eː/ was raised to take its place, thus triggering the Great Vowel Shift.)
Koko wrote:/j/ remained, but only because this is a sound much closer to an aproximant /e/, which I have no idea how to represent nor make.
You can't have an approximant at that vowel position. The closest you can get is the semivowel [e̯].
Koko wrote:As for the affricate /ps/, this is an evolved form of /pfʃ/. The /f/ merged with the /p/ while the esh was alveolised to /s/. And yes, /pfʃ/ is considered one consonant in the proto language; they tended to slur their consonants together: many weird clusters formed, but were later abolished (of course, not consciously.)
"Considered one consonant" by whom, for what purposes? Linguists have their own criteria for these situations, and without knowing at least something about the phonotactics of your language, I can't begin to say whether they are close to being satisfied.
[Insert rant about how syllable structure/phonotactics is a basic part of any phonological description, and without it what you have is really just a phoneme inventory.]