I generally avoid house-ruling in most games for the sake of simplicity and convenience, but I have actually house-ruled a couple of cards for Mage Wars. Some of the rules in the OP make sense thematically (such as allowing spiders to climb), but I personally prefer to house-rule only when I feel that a card is either so abysmally bad that it cannot be salvaged otherwise, or when the issue is clearly something which should have been addressed when the card was initially designed (an oversight, basically). Here's my current list, with my reasoning:
[mwcard=FWC04]Goblin Builder[/mwcard] - can repair walls which border his zone / reasoning: the devs more or less admitted this was an oversight, but they are apparently busy with other matters and aren't sure when/if they will issue an official errata for the card. The card could already build walls, so this change is also thematically consistent.
[mwcard=MW1J06]Gate to Hell[/mwcard] - after opening the gate, it receives a channeling of +3, the spawnpoint trait, and may summon an unlimited amount of creatures per turn (if the Mage can pay for them, of course) / reasoning: If I had to pick the single most unplayable card in the game, I might well pick this one. It's not just underpowered: the card itself is schizophrenic. It tries to act as both a massive attack card and a spawnpoint, but those two things don't have an overlapping purpose. The attack is far more effective later in the game (when the opponent has more creatures out on the board), but spawnpoints are better laid immediately so they can start to generate mana and summon things. Not to mention that the card costs 24 mana and two full turns, which will severely hinder tempo control and aggression. These changes allow the Adramelech Warlock to make good use of the card, without making it overpowered. She needs a way to summon her low-cost imps in an efficient manner, and potentially setting many of the creatures in the arena on fire is just great in general for her.
[mwcard=FWJ01]Akiro's Hammer[/mwcard] - possesses the indirect attack trait / reasoning: the card was ludicrously easy to shut down without this change, and quite frankly, it's a freaking trebuchet. There is literally no card in the game that more deserves to have the indirect attack trait. Firing indirectly is precisely what trebuchets do in REAL LIFE.
[mwcard=DNJ08]Seedling Pod[/mwcard] - costs 1 mana / reasoning: the card was overcosted and too vulnerable to ever be useful (on top of giving away the future location of your plants). There was basically no benefit to using the card, which meant there was no benefit in using Samara Tree either. That in turn made Vine Tree the go-to Druid partner. Making the card extremely inexpensive means that it is no longer a mana drain, and is instead a trade-off: delay putting down useful plants for 3 turns in exchange for a net gain of 2 mana (instead of the old "trade-off" of delaying and only gaining mana after 3 turns).
[mwcard=WARLORDABILITYOUTLINE]Warlord Ability Card[/mwcard] - the Veterans ability also grants +1 non-spell ranged. Instead of giving a token to a soldier unit that kills an enemy creature, whenever an enemy creature dies, the Warlord gets to choose two soldier creatures in that space to gain a token each / reasoning: as printed, the veterans ability is almost without a doubt the worst Mage ability in the game (the only other ability even in contention is Wounded Prey, which I have given serious thought to also house-ruling). Without an errata, there is zero reason to play the Orc Warlord over the Dwarf Warlord. The reason for granting +1 ranged is that the warlord shouldn't be penalized for using his ranged units (including the awesome Goblin Alchemist). Giving out more than 1 token at a time without allowing stacking encourages unit spam, which is precisely what the Orc Warlord was supposed to be good at in the first place.
[mwcard=MWSTX1CKJ02]Wizard's Tower[/mwcard] - possesses the epic trait / reasoning: the card is simply too powerful to be able to re-summon. This is also more of an informal house-rule (I build pretty much all of the decks for my group, and simply never include more than 1 copy of the card in any deck). This change doesn't outright diminish the card's power in any way, but forces the Wizard to actually defend it if he wants to keep using it.
Huh... that's actually more than I remembered changing, but whatever. I consider those to be pretty conservative edits (with the possible exception of the Warlord). The Goblin Builder and Wizard's Tower changes basically just fix oversights, and the Akiro's Hammer change just prevents players from cheesing it. The three remaining changes fix (IMHO) broken cards as simply as possible.
Gate to Hell. ...Either the real card is overcosted or more likely (IMO) people just don't know how to use it.
Considering it's been out since the first release and almost everyone seems to think the card is nearly useless or in need of a change, I'm gonna go ahead and say that there simply is no way to use it correctly. It's not even necessarily about cost: the card just tries to do too many unrelated things that don't synergize well. It would actually be far more useful to have one card that only gave all demons +1 melee, one card that only performed the massive attack, and one card that only acted like a spawnpoint (even for a higher total cost). Then a Warlock player could use each card when it made sense to do so.