Lets get it our of the way: Yet another mechanic. If this set has less than 300 cards, this is way too many.

What an odd little ability to keyword it is, too. Encourages playing a limited number of mana costs.

For pure "This turn is taking too long!" reasons, probably doesn't want to live at common. Which, again, means it probably shouldn't be a named mechanic.

It's not another mechanic, it's one of the 5 basic mechanics of Muraganda. Yes, it is a little complicated and I worry about it, but it fits kinda well for the described "mages studying ancient glyphs". That basically means to use the graveyard and the library to gain card advantage. If it proves to complicated I might need to come up with a new one.

The number of mechanics I suspect is fine. Khans of Tarkir had six - one for each of the clans, plus morph, and then plus themes like Warrior tribal on top of that. I think your Muraganda set is similar, and is fine from that point of view.

This specific mechanic though I worry isn't sensible at common. It'll lead to repetitive game states.

I do like the constraint on deckbuilding, though I suspect it'll lead to people making their decks worse: e.g. if I've got a great 4-mana removal spell or two, and even one archaeology card, this encourages me to play more 4-drops than I usually would. I think that's a trap - it'll hamper me in more games than I get the archaeology bonus.

Khans of Tarkir was the template, yes. Tricolor factions, shardbased, with each having its own mechanic plus a neutral mechanic. We felt it best to have not too many gold cards since the basic land theme would be hard to pull off.

This ability is also under observation, but I want to try it first. It might lead to a lot of feel bad moments if you never get to return a card, although the chances for that should be low enough in the late game.

Well, at least one time in three it hits land and fizzles (you don't have 0-cost stuff in the set)