My anxiety levels skyrocket when my hypothesis is actually verified in experiments. It gives me that haunting feeling of "I must have done something wrong at some point."

Ye olde "you only need one counterexample to prove false, but a full elaborate proof covering all cases for true"

I get obsessive about checking stuff and usually try to find another way to test it

When things seem to work, my first instinct is to make the tests harder - especially when it works right away.

@Fast-Nop In this case the easy test is the hardest case. I did the experiment in the simplest settings and some model failed. I assumed that it actually would still work in these settings, and that I will have to find complex scenarios where it does fail to justify why I want to use my method.