For play-testing, have you heard of Magic Set Editor? It can format custom magic cards and print sheets of proxies. It's a windows program, but there are ways to get it to run on mac.
Two favorite cards. Wash Out and Cut the Thread. Cut is perfect. Wash Out is just the right hosing effect for hags, but it's weak. It needs to be cheaper, tap two creatures, or come with a body. Maybe a creature that taps a target, and the target stays tapped for a turn if its a hag.
Besides that there are plenty of little balance and color-pie issues. Some of the cards would not normally see print at common if that is something you want to worry about it. Don't worry about it too much at this point.
The cyclopes cards seem to have variations on the reveal mechanic: When this cyclops is revealed, reveal another cyclops, count how many cyclops are revealed, etc... This is good for testing. Eventually you may want to make them work more consistently, especially at common.
It can be iffy to have both +1/+1 counters and sprout counters in the same set since players can get confused about what the counters on a creature represent (although in this set there is generally a creature type delineation). I had an idea for elemental tribal a while back that might apply well to plants. Put most off the tribal support for plants on lands, then give some of the plants effects that dig for lands.
Planning to test? You almost have enough to throw together some rough limited decks.
Yeah, I built a pauper cube a few years ago and I put Eye Gouge + a bunch of cyclops in it.
If you do make two sets, I'd recommend trying out a bunch of tribes (10+), before narrowing it down to the ten that interest you the most and dividing them into two color-balanced sets of 5. That way you can pair tribes together that best compliment and contrast each other.
I'd think about how each tribe works together as a deck. The plants having a split aggro/control focus seems like it might discourage you from putting green and black plants together. I like the way you described the cyclopes better, where red and blue cyclopes contribute their colors effects to a unified game plan. This is more something to do on the level of individual rare cards that push a tribal archetype in a specific direction.
Plants are another tricky tribe because it is hard to distinguish them thematically from treefolk, elementals, and fungus. If you wanted a weird take on plants, what about moving them to green/red? You know how some forest ecosystems are adapted to burn down periodically and then regrow rapidly? The plant people use their own dead leaves and twigs to power their pyromancy. See Creeping Trailblazer from the M2020 spoiler.
Not a big fan of Graze mechanic for goats. For one thing, it let's mono-red and mono-white cards do things they are not supposed to. I wouldn't worry about finding a keyword mechanic for each tribe, and I'd be careful about defining each tribe too narrowly around a mechanic, because it limits the types of creature card you can make (For example, its hard to do tribal with subtypes that by convention are expected to have flying, deathtouch, or some other keyword).
Will let you know if I think of any more ideas/solutions.
The beauty of custom sets is you can do what you want
Wizards has never printed a cyclops smaller than a 3/3 because they have always adhered to the mythological depiction of cyclopes as a type of giants. Making smaller cyclopes would make things a lot easier. You just say on this plane that some cyclops are smaller, or some are big but defensive (2U 1/4) or they are big but have drawbacks like "can't attack or block alone." The important thing for a tribal deck is that is that it's mana curve isn't too skewed.
The other challenge is the popular conception of cyclops that stems from the Odyssey. The "community" of cyclopes in the Odyssey were totally unhelpful when Polyphemus gets his eye poked out. But now that I think about it, there is another other depiction of cyclopes in Greek mythology–Arges, Brontes, and Steropes–who worked for Hephaestus and forged thunderbolts for Zeus. That depiction would fit well with a blue-red instants and sorceries theme or maybe a white-red equipment smithing theme.
The only other thing I can think of is that Polyphemus kept goats and sheep. I've always found dual-tribe themes interesting (Ogres and Demons in Kamigawa), so how about Goat/Cyclops mutualism.
For drafting, it's good to have each tribe present in at least two colors. This also gives you a more range in the kinds of effects and creatures you can make.
Hags are definitely an unusual tribe. I like your "Power of Three" concept, though its exactness might be difficult to get it to play well in limited. The best comparison would be the artifact density of 33% in Scars of Mirrodin to support metalcraft. Your deck does nothing when you draw too few hags or your opponent saves removal to keep you off three. If your deck is full of hags and you draw a bunch, most of will sit dead in your hand (though your coven on the board should be winning you the game).
As is, I'd recommend making the hags better at baseline and reducing the potency of the coven effects to reduce the swingy-ness of the mechanic.
It might be better to have effects scale up to a limit of three hags. For example "Put a +1/+1 counter on up to three target hags." or "Draw a card and lose 1 life for each hag you control up to three." This wording might be a little confusing on some effects.
What if hags had a keyword "Toil" (Tap this and two other creatures you control to toil), and then an additional rider "When this toils [do something]." So a hag can tap with any two other creatures to trigger its effect, but if the other creatures are hags, their toil effects will also trigger. You could have assembly-line synergies among effects: "Create a 1/1 frog creature token" -> "Opponents lose 1 life for each hag and frog you control" -> "Sacrifice a creature. Draw a card."
As for cyclops, their main challenge is that they are medium-to-large creatures, which means they must be handled differently from smaller tribes like goblins more elves.
Because cyclops are associated with the Izzet there are several existing cyclops cards that care about instants and sorceries. You could try doing cyclops and in red and blue, with mechanics built around counting instants and sorceries in the graveyard. The play pattern of the deck would be to cast cheap instant and sorcery spells in the early turns, which strengthens the cyclops that you play in the mid and late game. The explicitly tribal effects could go on some of the instants and sorceries, maybe asking you to reveal a cyclops from your hand (like dragon tribal in Dragons of Tarkir).
http://magicseteditor.boards.net/page/downloads
It makes testing a lot easier. Use generic core set cards to fill holes in your mana curve and just throw some decks together to see how things play.
Besides that there are plenty of little balance and color-pie issues. Some of the cards would not normally see print at common if that is something you want to worry about it. Don't worry about it too much at this point.
The cyclopes cards seem to have variations on the reveal mechanic: When this cyclops is revealed, reveal another cyclops, count how many cyclops are revealed, etc... This is good for testing. Eventually you may want to make them work more consistently, especially at common.
It can be iffy to have both +1/+1 counters and sprout counters in the same set since players can get confused about what the counters on a creature represent (although in this set there is generally a creature type delineation). I had an idea for elemental tribal a while back that might apply well to plants. Put most off the tribal support for plants on lands, then give some of the plants effects that dig for lands.
Planning to test? You almost have enough to throw together some rough limited decks.
If you do make two sets, I'd recommend trying out a bunch of tribes (10+), before narrowing it down to the ten that interest you the most and dividing them into two color-balanced sets of 5. That way you can pair tribes together that best compliment and contrast each other.
I'd think about how each tribe works together as a deck. The plants having a split aggro/control focus seems like it might discourage you from putting green and black plants together. I like the way you described the cyclopes better, where red and blue cyclopes contribute their colors effects to a unified game plan. This is more something to do on the level of individual rare cards that push a tribal archetype in a specific direction.
Plants are another tricky tribe because it is hard to distinguish them thematically from treefolk, elementals, and fungus. If you wanted a weird take on plants, what about moving them to green/red? You know how some forest ecosystems are adapted to burn down periodically and then regrow rapidly? The plant people use their own dead leaves and twigs to power their pyromancy. See Creeping Trailblazer from the M2020 spoiler.
Not a big fan of Graze mechanic for goats. For one thing, it let's mono-red and mono-white cards do things they are not supposed to. I wouldn't worry about finding a keyword mechanic for each tribe, and I'd be careful about defining each tribe too narrowly around a mechanic, because it limits the types of creature card you can make (For example, its hard to do tribal with subtypes that by convention are expected to have flying, deathtouch, or some other keyword).
Will let you know if I think of any more ideas/solutions.
Wizards has never printed a cyclops smaller than a 3/3 because they have always adhered to the mythological depiction of cyclopes as a type of giants. Making smaller cyclopes would make things a lot easier. You just say on this plane that some cyclops are smaller, or some are big but defensive (2U 1/4) or they are big but have drawbacks like "can't attack or block alone." The important thing for a tribal deck is that is that it's mana curve isn't too skewed.
The other challenge is the popular conception of cyclops that stems from the Odyssey. The "community" of cyclopes in the Odyssey were totally unhelpful when Polyphemus gets his eye poked out. But now that I think about it, there is another other depiction of cyclopes in Greek mythology–Arges, Brontes, and Steropes–who worked for Hephaestus and forged thunderbolts for Zeus. That depiction would fit well with a blue-red instants and sorceries theme or maybe a white-red equipment smithing theme.
The only other thing I can think of is that Polyphemus kept goats and sheep. I've always found dual-tribe themes interesting (Ogres and Demons in Kamigawa), so how about Goat/Cyclops mutualism.
Hags are definitely an unusual tribe. I like your "Power of Three" concept, though its exactness might be difficult to get it to play well in limited. The best comparison would be the artifact density of 33% in Scars of Mirrodin to support metalcraft. Your deck does nothing when you draw too few hags or your opponent saves removal to keep you off three. If your deck is full of hags and you draw a bunch, most of will sit dead in your hand (though your coven on the board should be winning you the game).
As is, I'd recommend making the hags better at baseline and reducing the potency of the coven effects to reduce the swingy-ness of the mechanic.
It might be better to have effects scale up to a limit of three hags. For example "Put a +1/+1 counter on up to three target hags." or "Draw a card and lose 1 life for each hag you control up to three." This wording might be a little confusing on some effects.
What if hags had a keyword "Toil" (Tap this and two other creatures you control to toil), and then an additional rider "When this toils [do something]." So a hag can tap with any two other creatures to trigger its effect, but if the other creatures are hags, their toil effects will also trigger. You could have assembly-line synergies among effects: "Create a 1/1 frog creature token" -> "Opponents lose 1 life for each hag and frog you control" -> "Sacrifice a creature. Draw a card."
As for cyclops, their main challenge is that they are medium-to-large creatures, which means they must be handled differently from smaller tribes like goblins more elves.
Because cyclops are associated with the Izzet there are several existing cyclops cards that care about instants and sorceries. You could try doing cyclops and in red and blue, with mechanics built around counting instants and sorceries in the graveyard. The play pattern of the deck would be to cast cheap instant and sorcery spells in the early turns, which strengthens the cyclops that you play in the mid and late game. The explicitly tribal effects could go on some of the instants and sorceries, maybe asking you to reveal a cyclops from your hand (like dragon tribal in Dragons of Tarkir).
Tribal suggestions/requests: Centaur, Clamfolk, Egg, Eye, Gamer, Goat, Hatificer, Kangaroo, Salamander, Weird.