Not bad!  Your main character is cool~  maybe give him animations after the jam or use him in another game with animations.  Great start of a game, good job!  Please check out https://neatgames.itch.io/only-one-arrow if you have time.

I really liked your premise of making an adventurous game where you can only carry one item at a time, it is really interesting and creative, the mechanics for grabbing and throwing stuff alongside the movement feel really good, and I loved your visuals. And please don't take this the wrong way, your game displays good effort on your part, but I must say though, that there are some things in your level's design which can be improved a fair bit, and I tell you this because it really shows that you like making games. Because you care, I would like to help you improve and reach your full potential. As I see it, there are three major problems that you should  try to work on for your next game, and some minor ones that you can work on after that; having honed your skills as a game maker. Here they are and I left some suggestions on what and how you could improve with the wishes that it is helpful to you:

1. Backtracking: This is important for your game, because holding only one item at a time makes the player need to go back and pick something up that they left behind, like a sword or a key of a particular color. Backtracking is usually seen as a bad thing because when you go for the first time into some new place, there is this sense of awe and discovery: a new challenge, a new level, a new area, a new mechanic, and so on. When you go back from an area you alreay beat, there is no new discovery or new challenge awaiting, it is just dead time for the player to walk back to the beginning. But you can either design around it, or use it for your advantage. A way to design around it would be to use one-way portals of sorts, where you have a main area that branches off into smaller areas, where at the end of them, a portal takes you back to the beginning. Or maybe this portal has to be activated from one side so that you can quickly go back to the end of the branch if you leave anything useful behind. An arguably more elegant approach to this, is to do what Dark Souls made with their levels; build your areas so that they end again on the beginning, with some locked door that you may have to unlock from the other side, or something similar. It requires a lot of thinking it through and it is hard to always be conscious about it, but keep making levels and keep trying to build stuff. It may not work as you intended, but know that you're getting better the more you experiment. Also, don't leave the things you make behind, iterate them, improve upon them, see how they can be fixed.

2. Use of space: This is pretty relevant to any game you make where there is navigation. Always know why are you making a level the size that you're making it. Is it to encourage the player to discover? Is it because the enemies in it demand the space? Is it because it is a boss arena? Is it a long hallway building up for the climax of the game?  This video is super relevant for this case, you have a lot of areas that are big but empty, and when there's empty space, the player will expect to find something there; maybe a treasure, a secret, a boss, a joke; something. As a thought experiment, see your map and shrink, block or outright cut the areas or the map that are dead space. Dead space are areas that offer no new/different challenge or  just don't have anything in there. Use space consciously, keeping track on why and acting upon it, will do wonders. Give areas a purpose.

3. Variation, contrast and repetition: I like to think about this as if I was writing a piece of music. Talking about music, in music theory there is something that is called phenomenology. Phenomenology is about how we percieve music; or <copy-pasting some article about it> "phenomenology of music is an investigation or inquiry into the direct perception and influence of sound and how the sound contributes to the musician's ability to reach a transcendent performance". What I'm trying to get at here, is don't only make things be factually different, but make them feel different; make the player feel different things; strive for variety of kind. On a top-down shooter you may have ten different shooting galleries, but maybe they feel more like one or two, because even if they absolutely are different, they convey the same; imagine having a level with a screen size of 256x256 and moving it a single pixel for each new level after that: yeah, you have a lot of levels, and they are factually different, but phenomenologically they're the same. On the flip side, too much variation makes your game feel like a gimmick or out of place, because repetition legitimizes (try reading between the lines in the video).

The biggest issues with your game link back to your level. Remember that your levels are the window to your mechanics. Without a good level, a great mechanic might be looked down upon because the levels did not show us that mechanic at its full potential, or didn't teach us how to use it properly. And levels are pacing/flow as well. Bad flow can make a very insightful story, or very good moments of gameplay be tarnished by everything else around it. A good experiment to try, is designing multiple cool moments based on what you want the player to feel, and writing reasons why it would achieve it, and placing them before the thing happen: Yoko Taro has an interesting GDC about it. Maybe make a game based around a single important moment / climax with a mechanic, and then make the rest of the levels that teach the player how to use the mechanics in interesting ways.  Maybe for another experiment, try making your levels using this structure or this, that are mechanic focused (don't get stuck with them though, always try to go beyond). 

Don't throw what you made away or brush it to the side! Use the assets you already have to create something new, little by little make more engaging levels and put them out there for people to give you feedback. A great way to start learning, would be Super Mario Maker 2 and Gamestar Mechanic, they both offer great insight into what makes a good level, have a supportive community that will gladly show you the way forward, and perhaps most importantly, you can make levels super quickly and improve them (iterate them) easily after testing and asking for feedback.

Hope that the information above is useful to you, and I'm excited for what you can come up with as you keep moving forward!

Keep creating and putting your stuff out there!