Fake Claude Code sites spread malware through cloned install guides targeting developers

Fake Install Guide Websites Target Developers Searching for Claude Code

Developers searching for Claude Code installation instructions are falling victim to a sophisticated malware campaign that exploits the growing popularity of AI tools. Cybersecurity researchers at Push Security have identified a new social engineering scheme called InstallFix, which uses cloned versions of the legitimate Claude Code website to distribute malware through fake installation commands 1

1

.

The fake Claude Code installers appear in Google search results as sponsored links when users search for terms like "Claude Code," "Claude Code install," or "Claude Code CLI" 2

2

. This malvertising approach places malicious sites at the top of search results, increasing the likelihood that unsuspecting developers will click on them before reaching the legitimate Claude Code page.

Cloned Sites Look Nearly Identical to Legitimate Installation Pages

The cloned websites replicate every detail of the authentic Claude Code installation site, including the layout, branding, documentation sidebar, and even functioning links that redirect to the legitimate Claude Code platform. The only malicious element is the one-line command used to install Claude Code for macOS, Windows PowerShell, and Windows CMD 2

2

. When developers copy and paste this command into their terminal, it downloads malware instead of the intended AI coding assistant.

This InstallFix variant modifies the ClickFix social engineering scam by inserting instructions to download malware during the installation process. The attack bypasses many standard malware protections because the user initiates it themselves 1

1

. The curl-to-bash command that appears legitimate points instead to an attacker-controlled server that can download anything without tripping anti-malware software.

Infostealer Malware Harvests Developer Credentials and Session Tokens

The malicious installation commands deliver a multipronged attack. When developers run the download command, it installs an executable that then downloads additional malware from a remote URL. The campaign appears to be related to Amatera Stealer, an infostealer malware that primarily targets user data 1

1

.

Amatera Stealer can harvest browser data including saved passwords, cookies, session tokens, autofill data, cryptocurrency wallets, and credentials, as well as system information 2

2

. The malware is also difficult to delete once installed. For developers, this means their code repositories, API keys, and development environment credentials could all be compromised, potentially giving attackers access to sensitive projects and company systems.

Attackers Exploit Trust in AI Tools and Installation Processes

The Claude Code scam is particularly effective because it exploits the trust developers place in installation guides. The legitimate Claude Code installation site asks users to copy and paste a curl-to-bash command into their terminal, making it difficult to distinguish between authentic and malicious instructions 1

1

. Attackers may further avoid detection by hosting malicious sites on legitimate platforms like CloudFlare Pages and Squarespace 2

2

.

This campaign reflects an increasing trend where threat actors use trending AI tools as bait for targeting developers 3

3

. As AI coding assistants attract new users looking to experiment with these technologies for the first time, InstallFix scams are likely to proliferate further, targeting developers who may be less familiar with security best practices.

Developers Should Verify URLs and Commands Before Execution

Security experts recommend several precautions when searching for coding tools or installation instructions. Always verify URLs carefully, as threat actors use tricks to make fake web addresses look legitimate at first glance. Developers should bookmark trusted sources to avoid relying on search results, and consider hiding sponsored results in Google search to reduce exposure to malvertising 2

2

.

Never run commands copied from emails, forums, social media posts, or websites unless you've independently verified their legitimacy. When possible, type commands manually from verified sources to ensure you're not copying and executing hidden malicious code. The targeting developers through fake installation guides represents a shift in how attackers are approaching the AI tools ecosystem, making vigilance essential for anyone working with these platforms.