Is the implied alphabet limited to the characters mentioned for each task ? For instance in task 1, is 'X' allowed to appear in a rule, even though it's not supposed to be in the input and the output ?
Yes, it's possible to use other characters.
Strange, it seems something else must be wrong with my solution for task 1 then, although it passes my tests. Is there a way to see the failed test ?
No, debugging your program without seeing the tests is a huge part of solving a problem.
I think that they should have defined "brackets" in task 1. Other than that, I think they defined alphabets for all tasks.