Well, while I do agree, and honestly hate very much inline HTML on Markdown, that's something Steemians are already used to.

The first one is a quick hack, and for me, it qualifies as a real issue since the single line raw html + link/image parsing is the default behavior of Github (CommonMark)

As far the image tags, there's another plugin to be published soon, concerning auto-images.

You would be amazed to see how many posts contains plain steemitimages.com URLs without proper image or link markup wrapping.

My take on that matter is to allow images matching patterns and extensions, and auto-image then.

The whole process is not heavy on the tokenizer, as far as I've tested.

So, on the things pending commit and proper repository refactoring coming up:

• @schema parsing (detect usernames on post content and auto-link)
• auto images for predefined patterns or hosts.
• safe-link exits (you are leaving to a suspicious URL).
• extensive, but more strict HTML sanitization.
• Asynchronous markdown rendering.
• Parallel render executing by ready-to-use Web Workers

As a tutorial from my recent "sanitization" journey, I ended up even creating a Web Assembly parser with Rust (with some help from @tensor).

Thank you for your dedicated review of this post @espoem, keep up the great work!