AML.CS0069
GTG-1002 Claude Code Espionage Campaign
- Type
- Incident
- When
- 2025-09
- Actor, as ATLAS names it
- Not stated
- Target
- 30 entities in the technology, financial, chemical, and government sectors
Techniques, in order
- 1AI Model Inference API AccessAI Model Access · AML.T0040
- 2LLM Prompt Injection: DirectExecution · AML.T0051.000
- 3LLM JailbreakDefense Evasion · AML.T0054
- 4Obtain Capabilities: Software ToolsResource Development · AML.T0016.001
- 5Compromise InfrastructureResource Development · AML.T0128
- 6Deploy AI AgentExecution · AML.T0103
- 7Autonomous Attack-Path AdaptationAI Attack Adaptation · AML.T0117
- 8Autonomous ReconnaissanceReconnaissance · AML.T0116
- 9Active ScanningReconnaissance · AML.T0006
- 10Develop Capabilities: Autonomous Exploit DevelopmentResource Development · AML.T0017.001
- 11Exploit Public-Facing ApplicationInitial Access · AML.T0049
- 12Enterprise Resource DiscoveryDiscovery · AML.T0075
- 13Enterprise Environment DiscoveryDiscovery · AML.T0089
- 14Unsecured CredentialsCredential Access · AML.T0055
- 15Valid AccountsLateral Movement · AML.T0012
- 16Create AccountPersistence · AML.T0125
- 17Data from Information RepositoriesCollection · AML.T0036
- 18Data from Local SystemCollection · AML.T0037
- 19Automated CollectionCollection · AML.T0126
- 20Data StagedCollection · AML.T0127
- 21Exfiltration via AI User InterfaceExfiltration · AML.T0136
How these attacks work
From MITRE ATLAS 2026.10 (Apache-2.0). ShieldMap lists who and what was involved and the techniques used, and links to ATLAS for the write-up; it does not copy the steps. Attribution is ATLAS's, not ShieldPi's.