What Is Working Well
You are in a strong patch and it shows in your decision making. Your opening repertoire is practical: the English Opening with White and the Caro-Kann Defense with Black mean you often reach middlegames you know well, and that is a real strength.
Look at the way you handled your game against civilkeller. By trading on g6 you left White with doubled pawns, then you used the a-file with your rook and the c4 square for your knight. You kept asking questions and converted a solid edge into a full point. That is high class positional play.
You also have a sharp eye for the initiative. In the game against mahmoud-atef_2 you found a rook lift, sacrificed the exchange in the right way, and finished with a mating net. That attacking flow is hard to teach, and you are doing it well.
Your endgame work has been clean too. Against kivi-456, you won a pawn, then made a central breakthrough with a pawn to d7 and forced resignation. Good technique.
Main Area to Improve: Meeting Unusual Gambits
Your only loss in this batch came against the Diemer-Duhm Gambit, an offbeat line after 1.d4. That game is a valuable signal: when the opening goes outside your repertoire, especially into a sharp gambit, you are less comfortable. Against a gambit, the test is not whether you can grab a pawn, but whether you can survive the initiative your opponent receives.
Try to develop your pieces first, keep your king safe, and only then worry about the extra material. For example, against the Diemer-Duhm Gambit, solid moves like ...e6 and ...c6 before capturing on e4 can blunt the attack. If you review your game against the gambit, ask how many moves you spent preparing versus defending.
Spend a little time every week on a "gambit survival kit": pick the sharpest lines you meet, such as the Blackmar-Diemer, Smith-Morra, or King's Gambit as Black, and know the best squares for your pieces. You do not need deep theory, just the first 8 to 10 sensible moves so you are never improvising in time pressure.
Next Steps
Keep your English and Caro-Kann repertoire. Add some pre-planned answers to 1.d4 gambits, and practice converting won positions with less risk. In your win against tiger1998 there were a few moments where you could have traded more pieces or simplified earlier, but you still found the quickest win. Tightening those decisions will make you very dangerous above 2100.
You might also review this Caro-Kann game and this endgame to reinforce the good habits you already have.