I don't know what special features "interactive conflict adaptation" adds to facial image-based Stroop task, so I cannot definitely say whether your task can be programmed (easily) in scriptingrt. what special features does it have?
apart from that: If you have already an inquisit version, you probably already considered simply using an inquisit web edition version? But the problem is that this will not run on Macs?
unfortunately, you have to start from scratch. there is no conversion routine.
For new projects, it is recommendable that you iron out the structure of one item, and then work with copy paste from there (because you don't have the table/script structure of e.g. eprime). if your tasks has a lot of trials, you can also assemble the items in a spreadsheet (excel with concatenate).
if you have a complex randomization / counterbalancing design, consider taking apart the experiment into multiple parts, and do the randomization in the html questionnaire (e.g., in qualtrics). There is no need to put everything into one scriptingrt study, you can assemble the data parts later.
best,
thomas