AML.CS0038
Planting Instructions for Delayed Automatic AI Agent Tool Invocation
- Type
- Exercise
- When
- 2024-02
- Actor, as ATLAS names it
- Not stated
- Target
- Google Gemini
- ATLAS attributes it to Embrace the Red (security evaluation).
Techniques, in order
- 1LLM Prompt CraftingAI Attack Adaptation · AML.T0065
- 2Prompt Infiltration via Public-Facing ApplicationInitial Access · AML.T0093
- 3LLM Prompt Injection: IndirectExecution · AML.T0051.001
- 4Delay Execution of LLM InstructionsDefense Evasion · AML.T0094
- 5AI Agent Tool InvocationPrivilege Escalation · AML.T0053
- 6Data from AI Services: AI Agent ToolsCollection · AML.T0085.001
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.