Zscaler flags Claude chat malware campaign on Macs

Zscaler has identified a malware campaign that used Anthropic's Claude shared chats to target Mac users. The attack exploited a trusted artificial intelligence service to make malicious content appear legitimate.

Researchers at the cybersecurity group said the attackers used the ClickFix technique, which aims to persuade users to run harmful commands on their own devices. In this case, shared Claude chats carried instructions that appeared genuine but were designed to infect Apple computers.

The campaign also used Google search advertising to reach potential victims. People searching for Claude-related terms were directed through paid ads to malicious shared chats presented as legitimate material.

Once users followed the instructions, the attack installed MacSync Stealer, according to Zscaler's threat-hunting team. The malware is designed to collect credentials, browser data, sensitive files, and cryptocurrency wallet information from compromised devices.

The findings point to a shift in how ClickFix attacks are delivered. Such attacks have more commonly relied on compromised websites or fake software downloads, but this campaign used a feature on a recognised AI platform instead.

Changing tactics

The use of Claude shared chats suggests attackers are adapting social engineering methods to match the growing use of generative AI tools in everyday work and research. By placing harmful instructions in a familiar online setting, the campaign aimed to reduce suspicion and increase the chances that users would follow the steps.

The operation shows how trusted online services can become part of malware distribution chains, even when the platform itself is not presented as malicious. The approach relied on the credibility of the platform's brand and on users arriving through search results linked to paid advertising.

Together, search ads and hosted chat content gave the attackers a route to users already looking for Claude-related information. It also reduced the need for more traditional lures such as suspicious downloads or hijacked web pages.

Mac focus

The campaign centred on Mac devices, an area attracting more criminal attention as Apple systems become more common in business settings. MacSync Stealer is built to harvest information that can be monetised or used in later attacks, including login details and wallet data.

Credential theft remains one of the most direct paths to financial fraud, account takeover, and wider network compromise. Access to browser data and stored files can also help attackers build profiles of victims or move further into personal and corporate systems.

The cryptocurrency element reflects the enduring value of digital wallet access for cybercriminals. Stolen wallet information can enable immediate theft, while passwords and files may support longer-running fraud or extortion attempts.

Industry response

The research adds to broader concern across the security industry about the misuse of mainstream digital platforms for criminal activity. As AI services become more widely adopted, defenders are watching for cases in which attackers use familiar interfaces and trusted brands in social engineering campaigns.

Zscaler linked this incident to a wider pattern in which threat actors quickly adjust to consumer and workplace technology trends. Rather than relying only on technical exploits, such campaigns often succeed by convincing users to ignore warnings and carry out the harmful step themselves.

Ruchna Nigam, Principal Security Researcher at Zscaler, said the campaign reflects a broader trend. "Cybercriminals are always looking for new ways to earn users' trust. In this Claude campaign, attackers used a legitimate AI platform to make their malicious instructions appear credible. As AI tools become more popular and widely used, we expect threat actors to continue exploiting trusted platforms and services as part of their social engineering and malware attacks."