First things first—great little engine! Minimalist, simple, and to the point. Love the UI simplicity!
Now, some long winded notes and suggestions:
Skills
Allow for the enabling and disabling of individual skills. This way, you can create a "do-it-yourself" character build. Rename skills to have non class related names. For example, a skill like "Fire Mage" makes no sense if you're playing as a thief. Call it "Fire Walker" and it becomes universal. Some skill suggestions below (for illustrative purposes only):
- Undying: Automatically revive once after death.
- Silver Tongue: Unlock new dialogue options.
- Backstabber: 50% chance for double damage from the back.
- Alchemist: Potions heal full life.
- Fire Walker: You can walk on lava.
- Locksmith: Open locks without a key.
- Cautious: Pass through traps without taking damage.
- Booksmart: XP scrolls give you double experience.
- Weapon Expert: Your weapons have +1 on stats.
Dialogues
Have branching dialogue options. If nothing else, at least have a ✗ or ✓ as options to answer, to create a semblance of dialogue.
Tilesets
The tileset is suboptimal. It covers just one genre (fantasy), has way too many NPC assets (considering the small scope), and way too few environment assets (even for the small scope).
The bigger problem is that right now, it stands somewhere between symbolic and literal representation. I would propose you move further into symbolic representation, with the optional import of other tilesets (modern, sci-fi, western, cyberpunk, etc.). To do this, you will need to implement name tags in dialogue and allow for duplicate (reused) NPCs and the renaming of NPCs in a scene. I know the asset import goes against the idea of having everything ready to export in a (short) string, but so does redrawing half or all of the NPCs to get what you need. If nothing else, at least think of including tilesets for the other genres I mentioned.
Variables
You probably need more. 16–20 sounds about right for a micro-game.
World
A bigger world map would be nice (perhaps have a switch between 3x3 and 5x5). The bigger option would allow for overworld, one big(ish) and one small dungeon, which is perfect for a small game. Right now, the 3x3 map is more of a proof of concept than anything really useful.
Objects
- Enemy Stats: Allow for changing enemy stats (0–9 range)—click on the boxes and write a value. This way, you don't need to use a fixed enemy progression.
- Weapon Stats: Allow for changing weapon stats (0–9 range)—click on the boxes and write a value. This way, you can easily create armor suits if needed(low attack, high defense) or different weapon kinds.
- Rename Tabs: Rename the "Swords" tab to "Weapons."
- Traps: Add a Trap—basically a static enemy that has 1 life and adjustable attack(see above). It can easily be bypassed by advanced characters or characters with certain skills. Great for temporary closing off parts of the map among other things.
- Chests: Add a Chest. It may require a key. When opened, it disappears, dispersing one item (scroll, potion, weapon, or enemy) of your choice. Surprise!
- End Game: The object should work similar to gates- activated by a variable. This way you can set a dialog as a condition to end game(great for narrative games).
With the above changes to the objects and the addition of the trap and chest, you cover all the basic pillars of a good micro-RPG: 2–3 location-appropriate enemies, one boss, at least one cool weapon and item, one trap, one secondary location, an element of surprise, and some story.
Ideally, you should also have one resource to dump into upgrades, so think about implementing money and traders as a resource drop and mechanic; however, you can probably skip this if you make skills a binary choice at level up. This way experience becomes the resource you dump into upgrades.
EDIT: Noticed the sign post and speech bubble behave identically- they set a variable and display information. Would make more sense if the speech bubble disappears after it's interacted with. Otherwise it just hangs there.
EDIT2: After playing some more with the engine I realized I was wrong on some points, so i crossed some things out.