@seanbudd, you write:

The sanitization deletes any HTML tags that are not recognized. This includes using angle brackets around words, e.g.: <minor>. If these are wrapped by code formatting these are correctly escaped: e.g. `<minor>` While english files are mostly correct, many translated files do not wrap text with angle brackets with code formatting. This means certain parts of translated documentation will be stripped, i.e. <major>.<minor>.<patch> becomes ..

I can see that you have fixed such issues in the Russian change log. But I guess other languages are impacted.
Is there an easy way for translators to figure which part of the file has been modified/deleted so that they can fix it? Else, we will probably end up with many information removed from the translated documents.

Though, I understand the need to sanitize documentation's HTML; my remark is not a criticism of this approach.

Is there an easy way for translators to figure which part of the file has been modified/deleted so that they can fix it?

Searching for angle brackets in markdown files should make it easy to find these issues.
Some early bad news: Translators will be needing to re-check the entire translated documents line by line when these files go up in crowdin. This means every line will need to be manually checked anyway.

The iframe tag in the Russian source caused the rest of the document to be escaped - i.e. half the document appeared as HTML source code in the browser. For other instances, it is just the fake "tag" that was removed.