Forest Global Training Partners

Cyber Risks · Unit 42 Palo Alto Networks

PhishingJS: A Deep Learning Model for Java Script-Based Phishing Detection

Summary: Palo Alto Networks' Unit 42 trained a deep learning model, PhishingJS, on roughly 120,000 phishing and 300,000 benign JavaScript samples to detect phishing pages by analyzing the JavaScript in their script tags rather than relying on login forms or brand logos in the HTML. The model, running in the cloud-delivered Advanced URL Filtering service, detects an additional 15,000 phishing URLs per week, including pages that use client-side cloaking to inject credential-stealing forms only after user interaction or page render.

Why it matters: Attackers are now hiding the credential-theft form until a user clicks, so crawler-based detection that reads only the HTML sees nothing to flag - and PhishingJS is finding 15,000 such URLs every week that those systems miss.

Read the original at Unit 42 Palo Alto Networks ↗

Topics
Malware
Related