AML.CS0036
AIKatz: Attacking LLM Desktop Applications
- Type
- Exercise
- When
- 2025
- Actor, as ATLAS names it
- Not stated
- Target
- LLM Desktop Applications (Claude, ChatGPT, Copilot)
- ATLAS attributes it to Lumia Security (security evaluation).
- On ShieldMap: ChatGPT (matched by name in the target).
Techniques, in order
- 1Valid AccountsInitial Access · AML.T0012
- 2Enterprise Environment DiscoveryDiscovery · AML.T0089
- 3OS Credential DumpingCredential Access · AML.T0090
- 4Use Alternate Authentication Material: Application Access TokenLateral Movement · AML.T0091.000
- 5AI-Enabled Product or ServiceAI Model Access · AML.T0047
- 6LLM Prompt Injection: DirectExecution · AML.T0051.000
- 7AI Agent Context Poisoning: ThreadPersistence · AML.T0080.001
- 8AI Agent Context Poisoning: MemoryPersistence · AML.T0080.000
- 9Manipulate User LLM Chat HistoryDefense Evasion · AML.T0092
- 10External Harms: Financial HarmImpact · AML.T0048.000
- 11External Harms: User HarmImpact · AML.T0048.003
- 12Denial of AI ServiceImpact · AML.T0029
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.