Congratz on tackling procedural generation! It can be a tough nut to crack. I did get an unwinnable situation on the 4th level I think.
As criticism, I would make the levels smaller/tighter. That would in turn make the graphics stand out more and would create more focused puzzles. Some different enemy types would also spice things up. But, overall, a dungeon crawler puzzler with static enemies can definitely work as a game.
Unwinnable levels are possible, a downside of the random generation (10 rooms - collisions, random puddle & gold, 3 mobs/room - collisions, a tunnel from room 1->2->..., and from entrance/exit inward to hit something. The only "restriction" was that the puddle & gold can't be in the same room. The only "guarantee" a path from entrance to exit.)
Crunch time ate into the graphics (a little more time and the attacks would have been animated.) Hmm, smaller/tighter levels . . . My knee-jerk reaction is that that would make it too hard, there being less possible paths. That is something to consider though (hmm, jack room insertion force. I'll play with it.) I had thought of including other fiends, such as ones that attacked only one way, shot arrows, or patrolled. Unfortunately, there was no time to balance that with random placement.
Thanks for the thought-provoking feedback.
Smaller levels would definitely put more strain on your level generation, but, as an upside, it's most likely easier to calculate if a level is solvable. Your maps are 50x35 tiles I think. You can compare them to the ones of my entry, where each map is only 16x16 tiles, and I'm very happy with how those turned out.
