From poking around using built in element inspectors in browsers, I believe the following portion of one of the style sheets is the issue:

pre.CodeMirror-line-like.cm-mw-section-3,pre.CodeMirror-line.cm-mw-section-4,pre.CodeMirror-line-like.cm-mw-section-4,pre.CodeMirror-line.cm-mw-section-5,pre.CodeMirror-line-like.cm-mw-section-5,pre.CodeMirror-line.cm-mw-section-6,pre.CodeMirror-line-like.cm-mw-section-6,.cm-mw-section-3 ~ *,.cm-mw-section-4 ~ *,.cm-mw-section-5 ~ *,.cm-mw-section-6 ~ *{font-weight:bold;}

In particular, from testing things out, I think the * following .cm-mw-section-3, .cm-mw-section-4, etc. at the end are what is causing the issue as when I delete the * in the developer tools window of my browser the issue immediately resolves.

Following up on my previous comment, and apologies for the ping, but @Jdforrester-WMF, could you point me at the general place at which support for language codes is implemented?

If possible, I might be able implement support for cmn if I was pointed in the general direction of where the relevant code is. I am fairly competent when it comes to coding.

Follow up on ISO 639-3 code recognition: the way {{#babel}} handles ISO 639-3 codes is different than the way it handles inputs which are not part of ISO 639-1 or ISO 639-3. As an example, {{#babel:foo}} renders no box and shows only a broken template link, while cmn, and seemingly the vast majority of ISO 639-3 codes, lead to a box proclaiming "This user has a native understanding of English." and displaying the passed code. This suggests to me that some basic back end implementation was done, but that it was not completed. This is not entirely surprising to me given that ISO 639-3 codes are much less common as well as much more numerous. This idea is additionally supported by the fact that the ISO 639-3 code for German (deu), and some other languages, display differently most ISO 639-3 codes, such as cmn, which are undifferentiated from each other. This means that there does exist a capability for such codes to be integrated into the extension.

I cannot confirm that has been anytime in the past in which "cmn" was correctly recognized as a language code on Meta. I only recently learned about the code "cmn" and that it might be able to be used to generate a Babel box. Theoretically it should have been/be since it is a valid ISO 639-3 code and the extensions Mediawiki page states all that ISO 639-3 codes are usable.

Got it. Thank you for takimg the time to respond.

What would the cause of the signifcant cost be?

@Cyberpower678 Hey, I submitted this a little while ago as suggested by on IABot's Talk Page as a feature request, its appears that you have been active since its creation (most recently as I write this on January 21st), but you have yet to take any action in relationship to it so I wanted to ask if you are aware of this request or not? Thanks.

The bug also appears when the '[' is inside the the text but not at the end of the text, examples include "[[Foobar|te[xt]]". Ending text in a opening bracket does not cause the bug, an example would be [[Foobar|text[]].