I tested your issue on a windows server with SVN repository and I had no issues with Umlaute in the readme section / plugin. Please see screenshot below:
I had another look. The umlauts are already broken in the HTML source code so it is no issue of the browser settings. When changing the encoding of the readme file to ANSI, umlauts are rendered correctly. So I checked the encoding settings of the server. Turned out that the code page was good old 850. I activated the - experimental (sic!) - UTF-8 support and now everything works as expected.
Only wondering why Windows Server 2022 Datacenter comes with such outdated default settings