Question

I want to write an online application that:

1. reads the URL from address bar of the browser
2. extracts its lexical features (like n-grams)
3. extracts its host based features (fetch DNS records online, its A, PTR, TTL fields)
4. classify the URL into malicious or benign (using machine learning)

Can anyone help me with 1 and 3?

Answer 1

I don’t believe this (application) is a task you can accomplish, as you can’t really determine site content based on url.

See something like Mozilla Phishing Protection Design Documentation and Google Safe Browsing spec instead