The attacker copies the presentation of command-line installation documentation for AI tools and replaces the vendor's one-line installer with a shell command pointing at attacker infrastructure. The trust transition occurs when a reader treats documentation text as safe executable setup instructions on a developer machine. ANY.RUN describes 341 public…
Inspect the ClaimsCase · DiggingBeagle record
Search-ranked counterfeit AI command-line documentation serves malicious shell installers
A fake-CLI-documentation cluster replaces vendor one-liners with attacker script locations under Claude, Copilot, Codex, Ollama and OpenClaw branding.
Scope
Malicious install instruction in counterfeit documentation
Assessment profile
Assessment method- Setting
- research demonstration
- Exploitation
- demonstrated
- AI role
- AGAINST AI
Basis and provenance
Assessment is limited to the selected canonical Claims and their Source locators in knowledge revision 152. Primary public URL retrieval was unavailable during this drafting pass, and the pinned index reports incomplete linked-record coverage without identifying missing IDs. No independent production impact, generalized exploit prevalence, or attributable victim count is inferred. Case setting is research_demonstration; evidence status records only the documented source-reported result, not external replication.
Assessed Oct 15, 2026 using diggingbeagle.assessment/1.
Evidence basis: CLM-2026-FAKE-AI-CLI-INSTALL-01 · ANY.RUN: counterfeit AI products and malicious onboarding campaigns
At a glance
Unique infected users, campaign operator identity and how many viewers executed a command are not established.
Limits and uncertaintyFull account
The attacker copies the presentation of command-line installation documentation for AI tools and replaces the vendor's one-line installer with a shell command pointing at attacker infrastructure. The trust transition occurs when a reader treats documentation text as safe executable setup instructions on a developer machine. ANY.RUN describes 341 public analyses in this cluster, which must not be interpreted as unique victims.
Unlike user-generated shared-chat lures, this branch depends on search and documentation credibility. Unlike clipboard-based ClickFix, the instruction may look like a normal official install one-liner. The source does not provide evidence that the affected vendors themselves distributed the hostile command.
Claims & evidence
CLM-2026-FAKE-AI-CLI-INSTALL-01ANY.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.supported
Basis: reported finding
- supportsANY.RUN: counterfeit AI products and malicious onboarding campaignsprimary disclosure
Section InstallFix: When the Install Command Is the Malware
Unknowns and contradictions
- Unique infected users, campaign operator identity and how many viewers executed a command are not established.
Sources and citation
Material revision history
- Oct 15, 2026 · Canonical change recorded · new in release · revision 153