Given a list of strings (words with less than 20 characters), return a new list containing only the words whose length is unique within the original list. In other words, if two or more words in the original list have the same length, all words of that length must be removed from the resulting list. Only words whose length appears exactly once in the original list should be included.