Attackers use CSS to fool anti-phishing systems

Email and collaboration security company Avanan has spotted an attack that uses multiple techniques to fool natural language scanners in anti-phishing systems.

The attack uses a mix of techniques to hide content from people while making it visible to machines.

The first obfuscation technique uses cascading style sheet (CSS) information to hide links. CSS is unseen metadata that tells a browser how to display text and images.

Attackers also hide links inside the HTML tag and use it to reduce the font size to one, said researchers at Avanan.

These attacks make it possible to get credential harvesting pages through to a victim's inbox, explained the company. The CSS hacking creates "gibberish" for natural language filters while displaying a fully rendered email to victims, it explained.

Many anti-phishing scanners use natural language processing to spot signs of fraudulent emails. They might treat an email with suspicion if it did not come from Apple but included text such as "© 2018 Apple Corporation. All rights reserved" in the body text, explained the company in another blog post.

While the users won't see the embedded links in these latest attacks, they fool anti-phishing scanners. "This combination can confuse text semantic analysis, which leads many to treat it as a marketing email rather than a phishing email," the company added.

This is the latest iteration in a series of techniques that Avanan has seen phishing criminals use to get past email scanners. These include setting the font size to zero and inserting hidden text that would break up text visible to victims. This is an old attack spammers used in the early days to circumvent anti-spamming software.

The company recommends using a multi-layered approach to phishing protection, including domain and sender analysis tools, to increase their chances of spotting malicious emails. Companies should also train users to confirm with IT before changing any passwords, it concluded.

Avanan made headlines last month when it discovered phishing attackers impersonating communications from DocuSign.