Ithkuil script is very tightly bound to that language's grammar. One cannot just take phonetic rendition of lojban text and write it in ithkuil characters. And while there are some common ideas in lojban grammar and ithkuil grammar, they are distinct enough to make mapping between them a quite difficult task.
For instance: in theory, there is no distinction in ithkuil corresponding to the distinction between sumti and selbri. In practice, most of time, main word of predicate have some verbal categories indicators, and other words have case category indicator, so we can take and use case category marker from the ithkuil script and use it to represent FA, modal and tense place tags. But there is two complications for that use: first, idiosyncratic semantics of FA places for every gismu; and second, that what lojban modal/tense system is even more complex and rich than ithkuil's dozens of cases. And that's only one example of transcription difficulties.
So, while one can devise a system for using ithkuil script to represent lojban text, it (most probably) will be highly unpractical and artificial. Better to take ithkuil script as a model or a lesson and devise a lojban semantic script based on similar principles.