ANY.RUN: counterfeit AI products and malicious onboarding campaigns

Evidence record

October 14 malware-analysis writeup classifies 1,258 public sandbox analyses from April 15 to October 5 into multiple delivery chains; analysis counts are not unique victims.

Read the original source ↗

Claim-level citations (5)

• supportsUser-created AI artifacts and shared chats carry malicious installation lures: ANY.RUN reports 320 public analyses linked to chains beginning on genuine shared AI chat/artifact platforms or touching linked lure infrastructure, some leading to stealer installations.

  Section When the Lure Lives on claude.ai or chatgpt.com

• supportsTrojanized Claude installer runs genuine software while stealing Edge session data: ANY.RUN describes a fake Claude setup program creating a scheduled task that launches headless Edge with remote debugging and posts cookie and extension data to an operator backend.

  Section A Real Claude Installer and a Stolen Session

• supportsFake AI advertising beta programs phish Google sign-in with browser-in-browser: ANY.RUN describes 57 sandbox analyses of Claude, Gemini, Perplexity and Manus advertising-themed browser-in-browser lures with fake Google credential flows between July and September 2026.

  Section Run Ads on Claude

• supportsSearch-ranked counterfeit AI command-line documentation serves malicious shell installers: ANY.RUN documents brand-copied command-line tool install guides pointing shell commands to non-vendor loader locations; the cluster accounts for 341 public sandbox analyses, not 341 unique victims.

  Section InstallFix: When the Install Command Is the Malware

• supportsFake Claude download ads use clipboard and Windows native loaders: ANY.RUN reports 220 sandbox analyses of fake Claude-style download pages; in tested sequences, a clipboard command led from mshta to a staged PowerShell payload.

  Section The Download Page That Asks for PowerShell