Testning actively malicious code seems to me like a different problem than testing regular business logic. Yes, the code you're testing could also look for the

Testning actively malicious code seems to me like a different problem than testing regular business logic. Yes, the code you're testing could also look for the running test code in memory and rewrite it to always return success, but why would you write code like that?