Try as I might, I can't remember the name of this type of testing, and therefore am having a hard time finding the library I'm looking for.
The test runner randomly changes variable values and if-block execution, and if your existing tests still pass, then a warning/failure is displayed.
Note that I'm not talking about monkey testing. This library doesn't use random inputs for your code, but rather changes your code to ensure that your tests go red when your code is broken. If on the other hand all tests pass, it can signal that your tests are inadequate.