In instructions it's said that "there will always be exactly one king". I coded assuming that was true and was hit by an exception on a (Haskell) random test case that has no king:

 King should not be in check:
  |---|---|---|---|---|---|---|---|
  |   |   |   |   |   |   |   |   |
  |---|---|---|---|---|---|---|---|
  |   | ♟ |   |   |   |   |   |   |
  |---|---|---|---|---|---|---|---|
  |   |   |   |   |   |   |   |   |
  |---|---|---|---|---|---|---|---|
  |   | ♞ |   |   |   |   |   |   |
  |---|---|---|---|---|---|---|---|
  |   |   | ♜ |   |   |   |   |   |
  |---|---|---|---|---|---|---|---|
  |   |   |   |   |   |   |   |   |
  |---|---|---|---|---|---|---|---|
  |   |   |   |   |   |   |   |   |
  |---|---|---|---|---|---|---|---|
  |   |   |   |   | ♝ |   |   |   |
  |---|---|---|---|---|---|---|---|

I assume this is a bug. If it's not, you should clarify in the instructions that "there will always be at most one king".