I added a site-wide feed.xml at the root, alongside the existing blog feed. It pulls the ten most recent posts, sorts them by date, and renders a standard RSS 2.0 document with categories drawn from each post’s tags.

I also stopped shoving entire post bodies into the feed descriptions. A new getExcerpt helper now looks for a <!--more--> marker and uses everything before it, falling back to the first paragraph when there’s no marker. The text runs through a small stripMarkdown pass that flattens links, strips HTML tags, and removes Markdown punctuation so the excerpts read cleanly in a reader. The existing blog feed picked up the same treatment.