Malicious link classifier
Inspired by publicly reported work at Cloudflare
Tell phishing, malware and defacement links from safe ones using only the URL text.
The brief
Cloudflare has written about using machine learning to score web traffic and attacks so they can be blocked before they reach a site. Start with a small, practical piece: classify links as safe, phishing, malware or defacement from the URL alone, explain what gives a bad link away, and test on fresh links. Good looks like: per-class precision and recall, features you can explain, and a browser-style demo that warns about a link.
Read the case first
What Cloudflare has said in public about this kind of work, on its own pages:
- Cloudflare Bot Management: machine learning and more (Cloudflare Blog) ↗
- Improving the WAF with machine learning (Cloudflare Blog) ↗
- Bot score (Cloudflare docs) ↗
Make it your own
Focus on scams people around you get: fake job offers, KYC update SMS links or lottery messages. Collect 100 examples from public warnings (never click them) and test the model on them.
AITeky has no link with Cloudflare, which has not reviewed or endorsed this brief. The brief only draws on what the company has published, uses open data rather than its data, and is a learning exercise: don't present your project as Cloudflare's work or system.
What you will build
- Explore URL classes and patterns
- Engineer lexical and host features
- Train and compare classifiers
- Evaluate per class and inspect mistakes
- Build a link-checking demo
Free, open data to build it with
Sign up free to get the data and links
It takes a minute with your email. AiTeky Pass (₹999/month or ₹9,999/year) is free for your first 30 days.