Currently, we have a few unit tests for malice detection (such as these https://github.com/maidsafe/parsec/pull/295/commits/4a574fd774103699d8098f9e5b1f8212a72783fb), but they don't cover a very varied range of possible attack, so they are not enough to make us feel comfortable every edge case will be handled properly.
Think of more edge cases and unit test them to ensure the code is correct.
Unit tests on their own will never give us 100% confidence, but they will be a good way to catch low hanging bugs. For more thorough testing, we will move also write more integration tests to be soak tested in a separate task.
Keeping in the backlog for now as we are first prioritising the Fleming work on the routing side. This is something we will come back to before testnets, though.