Seeing your code here and in some recent threads, I cannot forbear to say that you shouldn't triffle "too much" with regexes
For example, a channels name may contain the | char. Same goes for nicks. To play more safe, you'd have to escape (threat literally) text that may contain "metachars" with \Qtext\E. And to play even more safe, possible occurances of \E in the text itself with: $+(\Q,$replacecs(text,\E,\E\\E\Q),\E).
I forgot about this in another recent thread as well, and gave:
on $*:text:$($+(/\b,$me,\b/Si)):*: { stuff }
instead of the better:
on $*:text:$($+(/\b\Q,$me,\E\b/Si)):*: { stuff }
or respectively:
on $*:text:$($+(/\b\Q,$replacecs($me,\E,\E\\E\Q),\E\b/Si)):*: { stuff }
One could say that's nit-picking and the scenario is unlikely. Maybe it is... but then, missing a highlight is less dreadful than kicking or being kicked for the wrong reasons.

(In the concrete example here, I'd use $istok instead)