At the moment, the quality of the sentences in the Traditional Chinese collection is rather poor. When you look at Simplified (Mandarin) Chinese, it has a nice Fluency Fast Track with words in order of increasing difficulty, and a set of HSK groupings, but such features are unfortunately missing for Traditional Chinese.
There is only the “Random Collection”, in which the algorithm used to choose the cloze word seems to have gone wrong. It happens extremely often (> 50% of the time) that the cloze word is very common, e.g. (the Chinese for) “he”, “I”, “she”, “just”, and that this word re-occurs in many, many sentences - even though there would be much better choices for the cloze word in many of these sentences. When you are at the Chinese level where using Clozemaster starts to make sense, it is not very useful to repeat the word “he” all the time.
Would it be possible to create a Fast Track and “Pro Groupings” like for simplified Chinese? My suggestion would be to use the word lists of the TOCFL, which is the Taiwanese version of the HSK.
In the past, I have done some work on the Tatoeba sentence corpus for Traditional Chinese, so I can help if you like!