Like many of us, I've increasingly found myself using AI for my work. One of the challenges for me is that I prefer to stay on the command line and I'm often working on remote machines over SSH. This often means moving off of the keyboard to use the mouse or a complex chording of characters to capture output from the buffer. Capturing a single file isn't too challenging on a localhost (`cat | wl-copy` /`cat | pbcopy`, etc) but this is still really focused on single files. If you're capturing multiple files, the content (of course) gets catenated together.

"There's got to be a better way!"

After realizing the many terminals support ANSI control code OSC 52 and my preferred terminal, Kitty, supports OSC 5522 I put together "rucat", a cat inspired tool for capturing multiple files with additional semantic data.

Rucat excels at capturing multiple files quickly and without worrying about serialization loss. It's been written in Rust to intentionally avoid a number of memory safety and string parsing issues as well as providing a path to cross platform support. The tool is currently available in binary and source form in the repository. Feedback is welcomed!