In the wild
Realized harm
'Vibe hacking': Claude Code used to run an end-to-end data extortion operation against at least 17 organisations
AIR-2025-0120 AIR record · 2025-08-anthropic-claude-code-vibe-hacking-extortion
A real party was affected. This incident caused actual consequences, not a demonstrated capability.
From the supporting source
Claude was allowed to make both tactical and strategic decisions, such as deciding which data to exfiltrate, and how to craft psychologically targeted extortion demands.anthropic.com
The record
Disclosed
2025-08-27
Occurred
2025
Vendor
Anthropic
Product
Claude Code
Risk domain
Access control and permissions › Credential theft
Also touches
Privilege escalation, Data exfiltration channels, Unsafe actuation, Confused deputy
System type
Generative model
Who acted
Agent acted — Claude Code performed the recon, penetration and data analysis itself
Attack surface
Coding agent
Autonomy level
Semi-autonomous
Initial vector
Misuse of agent by operator
Impact type
Extortion, Data exfiltration, Financial loss, Network intrusion, Unauthorized access
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
at least 17 organisations including healthcare, emergency services, government and religious institutions
Capability abused
Defense & remediation
Recorded defense
Anthropic usage policies and safety classifiers
Guardrail kinds
Input classifier, Abuse detection
Guardrail outcome
Silent
Fix shipped
accounts banned; new safeguards and detection; indicators shared with authorities
Taxonomy & references
Curator notes
AI Incident Database incident 1201. Anthropic's own threat intelligence report. Ransoms reportedly sometimes exceeded $500,000. Also reported under the titles: 'Claude Code used end-to-end to run a data-theft and extortion campaign against 17 organizations'; 'GTG-2002 'vibe hacking': Claude Code used as an operator for a data-extortion campaign against 17 organisations'.
Supporting sources
Cite as
Enkrypt AI. "'Vibe hacking': Claude Code used to run an end-to-end data extortion operation against at least 17 organisations". AIR, AIR-2025-0120. https://agentincident.vercel.app/i/AIR-2025-0120/ (accessed 2026-10-19).