Honesty and alignments: Honesty is less of a Good-Evil axis trait and more of a Lawful-Chaotic one. I'm fairly sure the LE villain who never lies and never breaks a promise but still commits atrocities is quite common.
Rolepgeek: The income is fixed now, I think. I'm pretty sure the mistake was subtracting instead of adding 3E, because there's nothing else.
Quartz_Mace: That's probably true, but I didn't really weight any particular days in the last week or so, so it's probably an artefact of the time before the silence proposal passed.
Worshippers worshipping multiple gods: There's no penalty in the end, but it probably wouldn't start out with perfect efficiency. The justification for this is that not every worshipper of X god would also worship Y god whole-heartedly even if commanded or suggested to do so by X god. Thus, it's not a simple addition of incomes. Worship quality can still be full for both gods eventually, which is where the no-penalty part probably came from.
Fiat preventing mostly uninteresting grand outcomes: That's definitely considered to keep things interesting.
Stars being cheap: That was for constellations/signs. Mundane stars that actually support life (=plot-relevant) are far more expensive. I don't really ever neglect to consider plot-relevance, so it would have been expensive whether it was brought up or not.
KI'TORK and not KL'TORK: Why did it have to use an upper case I!?
Keshan dust ring and low-level Void cantrips: They still work. The resulting influence, however, gets cleaned up more quickly.
Aligning city to Good and limiting its power: That will probably work.
Oracle: The Oracle is not malevolent and won't attack anyone. It's a bit hard to attack when you're immobile, mostly treated as a bakery curiosity, and don't really have anything with which to attack anyway.
Inspiration for Melyssa: That looks interesting (and a bit far from the shipmap depiction).
ADDIT:
Were the Oracle so helpful, the player base would probably end up becoming very unstable.