github.com/Charcoal-SE/SmokeDetector/blob/zero-length-link/… if it's whitespace or not printable it'll flag as a zero length link detection. if the length of the text in the anchor (a tags) is zerolength there's a catch for that as well. AND it handles obfuscation too, strips out all the tags and gets plain text in there. Assuming the raw_body is proper HTML