I once inherited code that “worked” but operated on uninitialized values and out of bounds data unintentionally. I also couldn’t get it to complete a run without crashing. Even when I fixed bugs, I couldn’t reproduce the results. My own mentor told me to look the other way and move on as if this was known and totally usual.
It was also exceedingly hard to reproduce results from papers and authors seldom responded. Few would open data and even fewer would detail enough about their methods in the paper.
I’m sure others saw exceptional levels of fraud and other unethical behavior. I don’t, for one second, doubt that Claudine could have forged her way to the top. In fact, I think it could be the norm.