It is possible for a string to be single-script but still have a combination of characters that is not possible for a single language.
Useful to Chrome would be a spoof check that flags strings where there is no language that covers all characters in the string.
Special handling may be required for English since it is common to mix English with other languages.
Another possible issue: CLDR exemplar characters cover only a limited number of languages. Strings written in a language not covered by CLDR would be flagged by this test. Example: Shan (shn), which shares some code points with Burmese (my) but is not encoded into CLDR.
Needs an API proposal and implementation.
Could be a candidate for UTS 39 as well.