Great stuff!

Small suggestion I'd like to make: instead of "unlikely to likely " classification, would be clearer imo some kind of ranking instead, as in 0 to 5 or 0 to 100 (max value indicating the higher possibility of the token being a security)