There are a number of manufacturers of USB scopes.

This specific scope is a toy, not a real instrument.

There are USB scopes that do have reasonable sensitivity ranges and sampling rates. PicoTech is probably the largest manufacturer with their PicoScope series, but in my view they have never used a real oscilloscope - their PC software is really quite bad. I would expect that most smaller manufacturers would also have quite bad PC software.

I definitely recommend that you start by looking at a used scope from a major manufacturer and that isn't too very old. Almost everyone can design a USB scope, but if you can't quickly adjust range, sampling rates, DC-offset, trig levels, trig delay, ... then then scope isn't really useable.

Another thing is that quite a lot of scopes specifies bandwidth for repeatable signals. But a digital signal is normally not repeatable so the sampling rate must be high enough to catch the curve during a single trace. That means that you would like 10-20 times higher sampling rate than the frequency if the signal if you want to be able to see any suggestion of a curve form. Sampling at 2x the frequency is only applicable for a sinus signal. A square wave needs bandwidth enough for the overtones that forms the square wave or you would not see the raise and fall times or any ringing on the signal.

And the scope should have a square wave output signal so that you can adjust the capacitance in your scope probes.

These are just basic things with a scope - in reality, there are a large number of extras that are also needed to make the scope really useful for real-world problems. A digital scope with too short capture time (number of captured samples) don't allow you to zoom in on the individual bits of a serial message for example. Without a hold-off that works both with positive and negative time, you may not be able to catch the event you are interested in, because the event is very short but the natural trig point is a significant time before or after the event. Having the trig point after the event obviously requires the scope to have enough memory depth that it can constantly sample to memory until it sees the trig event.

The trig signal may be never stands still, in which case you might need a hold-off setting that says that you are only allowed to trig on a rising edge if there have been a pause longer than x time before the edge - this can allow you to trig on the first bit of a message, instead of trigging on a random bit within the message.

Some scopes have special logic to allow them to trig on very, very short pulses - way shorter than they can capture and show. This may be needed if you have a gated solution with timing glitches, where the delay in the individual gates makes a gate create a very short pulse representing the difference in gate delay between two gates.

Another thing when shopping for a scope: Check if the probes are supplied. High-frequency probes are very costly. But to look at faster signals, you do not want to work with a 1:1 probe, but with a probe with built-in damping together with a scope that have sensitivity enough to still display the signal. With a problem without damping, you can't compensate away the capacitive load of the probe, so the probe will both capture an incorrect waveform but also load the testpoint in a way that your circuit may fail.