Reminded a lot of the experimentation I had with wandcrafting in Noita and the alchemy type of games (Water+earth = mud, goblin + arrow = archer, etc.

The health/blood potion is a must, fire is suicidal (died twice, once with a fire potion (whyy?!), and once to a neverending fire I caused to myself with a fire arrow), single Ex(plosion) is the most simple, instantaneous, direct and efficient damaging spell (blow an small area very fast).
Compared to arrows (that can miss), swords (that can bounce off walls, a nice detail) and fire (that can damage you a lot over time), I supose explosion could be more expensive in the wand cooldown (IF you want some balance. Balance and fun are not the same, and fun is more important for a game.

Still have no Idea what the rune does, or if gold have any use besides be tossed like stone and wood. Took some time to understand that the feather increases the speed of the arrows. Was kinda of sad to not have found any combination with 3 materials, so that made me avoid getting the 4th material slot. 

In the end I had 1 [Explosion],  2 [blood][flask], 3 [feather][arrow-homing][empty], 4 [arrow-homing][blue-goblin], 5 [sword].

I like the prototype. This may be mentally demanding to keep expanding while still being interesting and not just more of the same, but with so very few "spell-crafting" games out there, this could be a good adition.
Spell-crafting games I remember from the top of my head:  Noita, Magicmaker, Mages of Mystralia, Magicka (1/2), heard about Two worlds 2 but still haven't played, Lichdom:Battlemage. There was another one that each spell had to be programmed with drag and drop mechanics, but can't recall the name. I hope this helped =)

Glad to be helpfull 😁

Well, noita revolves around gathering materials/spells and building into appropriate wands until you end the run (death or one of the win conditions). There is a reasonable team behind it, lots of time spent, and social interaction/content creators promoting it. If the idea was doing a roguelike just like noita, then it will be very close to a "worse noita". But there are some significant diferences that I think you should focus to serve as an alternative and not a competitor of noita (in my opinion, of course 😅): 
Your game is top-side (the map reminded me of the very old Gauntlet game, maybe a fusion between that one and the spell crafting could be an option).
You can always change the spells (in noita this only happens with a perk, or a mod).
A spell with different materials is not the sum of it's parts: in noita, besides the obscure stuff that I still haven't found and most players won't even bother, if you join the magic missile and chainsaw "materials", the end result is a better(faster) magic missile. In Sage & Summonion, if you join the wood and the fire materials, you summon a torch, that is neither a wood nor a fire (the alchemy stuff mechanic I said before). This needs more work (and brainstorming with people that support you) to get new ideas, verify what ideas have more potential, and fill in with more content. I was more hooked into finding out what combinations I could make than getting to the last floor. And these alchemy combination games are good even as mobile games (simple to understand, complex enough to keep interesting for those who want to keep finding out more). 

Ideas that came to me while writing this:
- Maybe let the materials be combined like the alchemy games?
- Probably making the last floor of each dungeon have a unique/special material to add to the collection
- If a roguelike/lite is what you are aiming for, maybe distinguish consumable materials ("spell ammo") and "eternal/runic/X" materials, a special type that you get awarded and can experiment at will, without consuming it. You can use this as a developer to limit the player to getting overpowered and making the challenge meaningless in some situations (besides the energy cost).
- If above is true, maybe add a system (more work code-wise 🥲🤷‍♂️) to "study" combinations, giving hints of possible combinations, or expected results as to not waste materials when studying unknown spells, only when casting them. This would protect the exploring and experimenting side of the game from the consumable materials. Finding a rare/new material would be very interesting and not "just one random spell test"
- Consider what will be the cost of losing. This game had 10 floors and one dungeon. If it had 10 dungeons with 10 floors each, death at the 8th-9th dungeon because a spell test went wrong might not add good things to the player experience.