These sound changes belong to kibylVi~.
  http://tzirtzi.ipage.com/akana/index.php?title=Kib%C3%BCl%CA%8Ci%E1%B9%85

General organisational principle: the order these sound changes are lifted gives the correct ordering, but it might not be strictly chronological.  In particular I like to write
A > B / conditions
B > C
to denote that B was an intermediate outcome which became C by the next time the distinction mattered; I don't mean them to necessarily be immediately successive.  

In A > B / C styie changes, unbracketed symbols are phones and bracketed are features.  In particular V is the wedge vowel while [V] is a cover symbol for vowels.



I'll use 'laryngeal' as a cover term for PPen /C x h/.  

Long [a:] sticks around and behaves as [aa].  Reconstructed long [i:] and [u:] might do likewise, but not next to vowels at any rate (though they do behave differently next to [a], see below).

[stop][laryngeal] > [aspirated stop]
[laryngeal] > 0 / [fricative] _

Where a nasal and a liquid could equally well be syllabic given the phonotactic restrictions, the nasal is.  

m=, n=, N= > u~, i~, a~
m, n, N > u~, i~, a~ / [obstruent or liquid] _ [V]

Subsequent vowel changes happen indifferently to nasalised and unnasalised vowels, unless otherwise specified.

A persistent rule that comes with the introduction of nasals: nasality on vowels spreads throughout maximal domains of vowels and laryngeals.  

[unaspirated obstruent] > [voiced] / internucleically
[laryngeal] > [voiced] i.e. j\ G h\ (everywhere)

ia, ua > E, O / so long as the [i u] aren't reconstructed as long in PPen.

i, u > e, o / _ G, _ r
e, o > E, O

u > u\ / _ [coronals] (incl. [j\] and [i j u\], but not [r])
u\ > y
yi > yy
iu > ii / in some high-yield items

a > E / _ [coronals] (as before)
a > A / _ [velars], _ r
a > V

r= > Er / _ [coronals] (as before)
r= > Ar

l= > Al / _ [velars], _ r
l= > El

[nasal] > [alpha place] / _ [obstruent, alpha place].  This makes a couple /J/ from nasals before /j\/.
J > n

[laryngeal] > 0 / except intervocalically before nonhigh V, or initially

j\, h\ > j, 0 

f > h
[stop]h > [aspirated stop]
h > 0 / non-initially

[stop][obstruent alpha] > [obstruent alpha, tense].  I denote tenseness by *; the resulting consonants are p* t* k* ts)*, the last from stop+[s].  Phonetically it's faucalised ~ stiff voice, like Korean.  

b, d, g, v, z > B, j, G, B, r\.  This introduces new [j G] which don't have nasal spreading; the laryngeal part of the persistent nasality rule falters.
r\ > r

sr > tr
sl, tl > kl
nl > Nl

[liquid] > 0 / _ ([V]) [liquid], and even _ [V][C][liquid] in several cases

[high V] > [glide] / [V] _ [V]
0 > [glide, alpha place] / [high V, alpha place] _ [nonhigh V]
V > [V alpha] / adjacent to nonhigh V alpha (i.e. wedge > a copy of the adjacent vowel)
[low V] > [low V alpha] / _ [low V alpha]
Groups low vowel + high vowel survive, as diphthongs; whether their second element is phonemically a vowel or a glide is probably academic.

[tense C] > [nontense] / after a tense C anywhere earlier in the word.  Any marginal [ts)] this creates fall back into [s].

Consonants become glottalised in absolute word-final position:
p t s k > ? / _ #
m n N > N? alternating with ~? / _# (phonemically distinct from the other nasalisation, though)
l r > l? r? / _ #

(short nonnasal) V > V_x / _ #.  After voiced Cs this /V_x/ tends not to be realised as any sort of vowel at all, but as a prolongation of the C.  After voiceless Cs it may devoice.

(short undiphthongised) i y u > @ @_c @_c / in unstressed closed syllables, whichever those are (note that V_x still exists).
[@] [@_c] are phonemicised as /E ~ V/ and /O/ -- for the first /E/ is usual, but /V/ is found in a few prominent instances.



Resulting phonemic inventory:

m     n                 N
p*    t*    ts)*        k*
p     t                 k    (?)
p_h   t_h               k_h
            s                 h
B                       G
      l     r
                  j (H)(w)

though a sufficiently inspired analysis could take [h] and [B] allophonic.

i  y     u
E     V  O
      A
+ nasal variants of all vowels
+ long variants of all vowels, which this diachrony mostly treats as sequences
+ lots of closing diphthongs.