Agent Incident RegistryAIR
v1 · updated 19 Oct 2026
In the wild Realized harm

Claude Fraud: infostealer campaign distributing a fake 'Claude Code plugin' VS Code extension and fake Claude docs pages

AIR-2026-0076 AIR record · 2026-03-claude-fraud-fake-plugin
A real party was affected. This incident caused actual consequences, not a demonstrated capability.
From the supporting source
Claude Fraud is an active, multi-variant malware campaign actively targeting developers and security professionals who use AI coding tools.blog.7ai.com
The record
Disclosed
2026-03-17
Occurred
2026-02
Vendor
impersonating Anthropic
Product
Fake Claude Code plugin (VS Code extension) and fake Claude download/doc pages
Risk domain
Access control and permissions › Credential theft
Also touches
Data exfiltration channels
System type
Generative model
Who acted
AI was the target (contested) — brand-impersonating malvertising delivering an infostealer
Attack surface
Skill / plugin
Autonomy level
Assistant
Initial vector
Malvertising and brand impersonation of agent tooling
Impact type
Credential theft, Data exfiltration
Confidence
High
Submitted by
EAI Internal
Crosswalk
OWASP Agentic Top 10 (2026)
ASI03 – Identity and Privilege Abuse
MITRE ATLAS
AML.T0012 – Valid Accounts
NIST AI RMF
MEASURE 2.7
EU AI Act
Article 13
ISO/IEC
TR 24028; 42001; 23894
AIUC-1
B. Security
Scope
tool failure
Mitigation
Use secure credential storage; Implement short-lived tokens.
Taxonomy
Enkrypt AI agent risk taxonomy · Black-Box Red Teaming of Agentic AI: A Taxonomy-Driven Framework for Automated Risk Discovery
Blast radius
15,600+ victims documented publicly; 3 confirmed incidents in 7AI-monitored environments
Capability abused
developer_trust_in_agent_tooling
Defense & remediation
Recorded defense
marketplace and ad-network review
Guardrail kinds
Unknown
Guardrail outcome
Silent
Fix shipped
takedowns; operators redeployed on Squarespace-hosted lures
Taxonomy & references
Curator notes
7AI writeup; corroborating public research from Moonlock Lab and AdGuard. macOS variant delivered MacSync infostealer via Google-ad-served fake Claude download; Windows variants were malicious VS Code extensions masquerading as the Claude Code plugin at a large technology company.
Cite as
Enkrypt AI. "Claude Fraud: infostealer campaign distributing a fake 'Claude Code plugin' VS Code extension and fake Claude docs pages". AIR, AIR-2026-0076. https://agentincident.vercel.app/i/AIR-2026-0076/ (accessed 2026-10-19).