It's really clever! I'll hold off on "if furthered developed" comments, but I really like the idea of the haphazard mage. The decreased trigger probability was scary and I never tried clicking on it cause I felt the flat randomness of the cards was going to serve my luck better than decreasing the chances.
The idea of tweaking the chances is to allow you to have higher control of the situation.
For example, you can lower the trigger chance of your defensive cards and then you can target the enemies and hope that your higher chance attack cards will trigger on them and the heals will be left for the final turns and you can cast them on yourself with 100% certainty.
Or, if you lower your character's target chance, you can play your attack cards first in hopes that they'll target the higher chance enemies.
I was working on an update to make it more visible through color-coding the chances, but couldn't upload it in time. In the uploaded version it is slightly annoying to have to read all the chances to come up with the optimal plan.
("haphazard" - first time I'm seeing that word, but I think it describes my game jam dev very well :D )
Oh yeah I felt the strategy of basically saving up heals, however on my 2 runs I would use heal on turn one (if I drew them at all) and then spam my damage spells on a single target. I felt that if I seeded my deck with only damage and never leveled up healing, then it wouldn't matter if the heals missed because the enemy would be dead. It was a glass cannon approach but it got me up to the ghoul with 2 skeletons at least.
If you consider developing this idea further, I think a lot of the trouble will be balancing the experience. The other issue is the massive influx of card based deck building roguelikes games on steam. You'd be in stiff competition with everyone, including granddaddy Slay the Spire.
There is so much creative room for new spells and flavor, but honestly good job!