Hackers Abuse Claude.ai to Distribute Malicious Software
Attackers leverage Claude's legitimate domain to trick users into manually installing credential-stealing malware.
Updated on Oct. 27, 2026 in Cybersecurity
Live Poll
Do you trust that advertisements appearing at the top of search results are always safe?
Hackers are using Anthropic’s official claude.ai website to host malicious advertisements for malware that target both macOS and Windows systems. By leveraging the platform's legitimate domain to host fraudulent content, attackers are tricking users into installing credential-stealing software.
Why it matters
This campaign exploits the inherent trust in legitimate, secure domains to bypass standard security filters. By hosting malicious pages directly on the claude.ai site, attackers weaponize familiar branding to deceive users into executing dangerous commands.
The macOS attack utilizes a 4-step manual command execution process to install the Atomic macOS Stealer. On Windows, the campaign hides a malicious .NET payload within a standard system process to scrape cryptocurrency wallet data.
The players
Anthropic
An AI research company that develops the Claude large language model and provides the claude.ai platform.
Gen Digital
A cybersecurity corporation that manages a portfolio of consumer security brands and conducts threat research.
The details
Attackers leverage the ClickFix technique, which instructs users to manually copy and paste scrambled commands into system terminals to 'fix' a supposed error. The malicious pages are hosted through Claude's Shared Chat or Artifacts features, allowing attackers to exploit the claude.ai domain and its HTTPS padlock for credibility. Once triggered, the Windows malware swaps legitimate Ledger or Trezor apps for compromised versions designed to capture recovery phrases.
Timeline
October 26, 2026: Gen Digital researchers publish a report detailing the active malware campaigns.
The Tech Race
This campaign follows the pattern set by the 2024 ClickFix malware campaign, demonstrating a sophisticated shift toward abusing legitimate SaaS platforms for distribution. It reflects a trend where attackers prioritize host domain reputation over traditional phishing infrastructure to bypass security filters.
Users should exercise caution when clicking Google Ads that direct them to claude.ai pages featuring unexpected desktop installation instructions. Never copy and paste unknown commands into a Terminal or Run prompt, especially when prompted by unofficial instructions on third-party sites.
The takeaway
This campaign highlights how threat actors now treat legitimate, high-reputation domains as a primary attack surface. Users should verify that any software installation instructions originate strictly from official, verified company download pages.
Further reading
For broader trends in domain-abuse and credential theft, visit Cybersecurity.
Source note: This article includes information reported by Cyber Security News.
Live Poll
Do you trust that advertisements appearing at the top of search results are always safe?
