This is a known limitation. Markup always is considered to be a word
breaker by the spell checking code. I will file an enhancement CR for
this. We may have to do this using the DCF file, so you would have to
list every tag that should not be a word breaker. The issue is some tags
can generate content. As an example think about an inline generating a
following mdash character, it should probably be considered a word
breaker. But this is hard to determine automatically inside the
software.
John Dreystadt
Software Development Director
Arbortext - PTC
734-327-6079
-