Omg, dude. this is awesome! And it's your first game on your profile, congrats!

For a first HTML game this is really coherent (I can tell you're a web dev). I like that you've drawn everything inside a canvas, and are forcing a low res pixel grid. Feels like a lazier version of you could've skipped that step, thrown some DOM elements over the top of the canvas, like the buttons, emoji etc. - and called it a day. But keeping everything bound to the canvas on the same pixel grid produces a really clean, consistent style. I know your pixel art is obviously a larger pixel size, but I think it works because it's larger by a fixed multiple. Anyway, thats just me hyper-nerding out over specifics. :)

The first level is really well done. I'm a big fan of teaching mechanics without a written tutorial. I bet most players are going to drag the filters onto the list from top to bottom, and solve it without even realising why. lol. For me, it wasn't until the next level where I did the same thing that I realised the order matters. I will say, the second level is quite a lot harder than the first for this reason, and you probably could introduce a softer gradient by requiring more obvious patterns at first. E.g. the first level was A,B,C, what about the second level being C,B,A. Then the third being A,C,B. Then the player can brute force it a few times while they start to see the patterns. Otherwise asking the player to immediately understand cryptographic logic is a pretty tall order. I will say also, it would be really nice to click a filter to remove it - maybe on hover you see a little red X or the text goes red, and clicking it removes it. The drag to move functionality is great, but I think drag-n-drop almost always should be accompanied by a click-and-click alternative.

The fire burning makes it super hard! Perhaps consider a way to extend the timer? E.g. for each filter you place in its correct position, you get an extra, I dunno, 10 seconds. That way you can also kind of game the system at first by trying every slot until one works. But later when you have loads more filters and slots, that won't work and you'll run out of time. So the player is still forced to actually learn the patterns to look for, but they're given the opportunity to take their time and get invested first. You could even tell a little narrative underneath about how cheating doesn't work in the long run, haha.

Trying to get the player to learn how your puzzle system works without burning them out immediately is the kind of thing people spend their whole game design careers learning. I only just recently got a taste for it in my game LIGHTS OUT. In that game, you're turning window lights on/off, but every adjacent window state flips. So you have to closely watch the patterns to learn how to actually get them all off. In the first level, I made it that the player just has to tap ONE window to win. Then in the second level, two windows. Then the third, four. But I made an error not doing three in the third level, and you can totally see how - it's so much harder by the third level.

Anyway, awesome entry bro, hope you keep releasing these TINY games and I'll be looking forward to a submission from you for TINY Jam October! P.S. if you do a next one, would you be open to posting a thread in the community tab when you first get the idea to start? I'd like to turn this thing into a space for sharing ideas and getting early playtesting before a game's release.

Cheers!