Currently the Trados (2024 but also the older ones) default seems to be to pass hyphenated terms to Hunspell dictionaries are separate tokens i.e. in my case, it passes a-rithist 'again' to the dictionary as a and rithist. Before you ask, a-rithist is in the .dic file and yes, it works fine elsewhere, such as LibreOffice.

By itself, a happens to be a standalone word but rithist isn't, so I get a red underline under ALL words with a word hyphen that don't consist of tokens that are standalone words - and there's a lot of these. Now there may be a use case for tokenizing stuff like this but as an across-the-board approach, it's not great.

Is there a way I can tweak the settings to turn tokenization of word hyphens OFF when it comes to spell checking?

English doesn't really have good examples of this, I guess the closest might be co-operation, where co isn't really a word by itself so passing co-operation as two tokens to Hunspell might result in the same problem.