The other crx checkers and extension scanners mostly just check permissions and for eval() calls in the code, but fail to actually find malicious behavior. I built extensionsafe around LLMs that gives an evidence-based abuse and risk profile based on the extensions actual functionality from the code.