Feed unicode/non-alphabet characters through a romanizer

Currently when a Korean heading ID is generated, the Hangul characters which are unicode are stripped, or in other markdown processors converted to useless dashes. In Ghost’s case they are stripped, so any headings that are hangul-only end up without IDs meaning you can’t create a table of contents with working links:

Usable heading ids could be generated by passing the heading through a romanizer like this one for example: GitHub - fujaru/aromanize-js: Korean transliteration tool for JavaScript

I’m also up to this issue. Any updates from Ghost team?