Want to talk specifics? Let's do it.
Mutation is absolutley not an instant-kill card. It's only 40% of an instant-kill card under the best of circumstances. It has a greater chance of turning the creature into an Abomination or a mutant than it does to kill them, and those potential consequences are not something you can simply ignore when it suits your argument.
I know for a fact that there's an unwritten rule against instant-kill cards because I've actually had a conversation with Zanzarino about it (which resulted in the current version of Aflatoxin, as mentioned above). Shockwave has been added to the game since then, which means Zanz may have changed his mind to a degree, but I would still avoid any kind of creation that said "Target creature dies" unless it had strict consequences or conditions behind it.
Now, let's see...yup, that's the only specific card you've mentioned.
But let's address your general misunderstanding by looking at some other questions you asked:
About not having graveyards:
Has Zanz ever said that? No.
Has Zanz ever said anything that could be interpret as anti-graveyard comments? No.
Can it be done technically? Yes.
Would it make the game better? Maybe.
So why is there an unwritten rule against it?
Because there is currently, to all appearances, no code that involves tracking the existance of cards that have left play. That means that in order to implement any card that involves such tracking, Zanz would have to put together an entirely new pile of code, which is something that he's just not likely to get excited about. Stuff like that is notoriously buggy and time-consuming, so you have to step into Zanz's shoes and ask yourself: "Is it worth the pain in the ass that this coding would require just to have this card idea in my game?"
That is the fundamental source of a majority of the unwritten rules I listed. If your idea is hard to code, it's less likely to make it into the game, period. The stuff that seems not-so-difficult to code (like targeting quanta pools), I mentioned as particularly flexible rules.
The fact that Elements is missing some feature at the moment doesn't mean that this feature will never happen.
Of course that's the case -- but it doesn't change the fact that hard-to-code cards will always be discarded more often than easy-to-code cards, so I don't particularly care.
Now, there are a couple of rules up there that aren't coding issues, but are still things we all recognize. No humans, for example, and no "Control Magic"-type cards for another. Those are things that either don't fit flavor-wise into the game (no humans) or are such painfully ovbious ideas their very nonexistance is a clear inidcator that Zanz doesn't want them in the game (Control Magic).
So, to go over my original list for your own satisfaction:
No cards that affect your opponents hand directly
This falls into the "too obvious to have not been done already" AND potentially the "pain in the ass to code" categories.
No cards that modify either players deck directly (except to put a card directly into a specific position in the deck, thus far only on top).
PITA to code.
No cards that take control of an opponent's creature.
Obvious.
No cards that have multiple non-"all", non-"random" targets. The targeting mechanism clearly is designed to handle one chosen target at a time.
PITA to code.
No cards that kill a creature without taking it's HP into account on some level.
Obvious. And, of course, I talked to Zanz about it.
No cards that require you to target a quanta pool.
PITA to code.
No cards that affect the "graveyard" (to borrow a M:tG term).
HUGE PITA to code.
And there you go. Disagree if you please.