Right now, it just looks for new lines and tests them; but it fails in the case lines are deleted, for instance, or in simply other cases where the underlying software is going to be changed.
In that sense, the testing script should also become a module that should be tested independently for all possible cases that might arise. Right now it's not totally clear it's testing correctly, and in many cases (for instance, external dependencies) we just "let it pass" (we have to locally test, anyway). That case is going to be difficult to cover, but in many others it shouldn't