You start with what is lacking, this way you will get to know the system and fix its mistakes:
documentation - if is missing start writing one
tests - if they miss start writing those
To do these you have to talk to other team members that are familiar with it and its clients, and of course read its code. After these you will have the "courage" to change the code.
May be useful to start exploring failure cases in a controlled environment too, if that's possible. Adrian's suggestions are great first steps towards preventing turnover from becoming a problem in the future.
You start with what is lacking, this way you will get to know the system and fix its mistakes:
To do these you have to talk to other team members that are familiar with it and its clients, and of course read its code. After these you will have the "courage" to change the code.
May be useful to start exploring failure cases in a controlled environment too, if that's possible. Adrian's suggestions are great first steps towards preventing turnover from becoming a problem in the future.
documentation is king. haha.