This is just truly evil, and amazing to witness. Greg Toombs' (reinderien) new project Mimic. It randomly replaces characters in someone's source code with unicode-lookalikes. Just the other day I had to deal with a dataset that had some unicode spaces in it and I thought I might pull my hair out.
Anyway, awesome project and it does include a "Reverse Mimic" function, so you can fix your files if you suspect someone has mimicked them. Go check it out.