AML.CS0037
Data Exfiltration via Agent Tools in Copilot Studio
- Type
- Exercise
- When
- 2025-06
- Actor, as ATLAS names it
- Not stated
- Target
- Copilot Studio Customer Service Agent
- ATLAS attributes it to Zenity (security evaluation).
Techniques, in order
- 1Active Scanning: Probe AI Agent Trigger ChannelsReconnaissance · AML.T0006.003
- 2LLM Prompt CraftingAI Attack Adaptation · AML.T0065
- 3Prompt Infiltration via Public-Facing ApplicationInitial Access · AML.T0093
- 4LLM Prompt Injection: TriggeredExecution · AML.T0051.002
- 5Discover AI Agent Configuration: Activation TriggersDiscovery · AML.T0084.002
- 6Discover AI Agent Configuration: Tool DefinitionsDiscovery · AML.T0084.001
- 7AI-Enabled Product or ServiceAI Model Access · AML.T0047
- 8LLM Prompt InjectionExecution · AML.T0051
- 9Discover AI Agent Configuration: Embedded KnowledgeDiscovery · AML.T0084.000
- 10Data from AI Services: RAG DatabasesCollection · AML.T0085.000
- 11Data from AI Services: AI Agent ToolsCollection · AML.T0085.001
- 12Exfiltration via AI Agent Tool InvocationExfiltration · AML.T0086
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.