AML.CS0016
Achieving Code Execution in MathGPT via Prompt Injection
- Type
- Exercise
- When
- 2023-01-28
- Actor, as ATLAS names it
- Not stated
- Target
- MathGPT (https://mathgpt.streamlit.app/)
Techniques, in order
- 1Search Open AI Vulnerability AnalysisReconnaissance · AML.T0001
- 2AI-Enabled Product or ServiceAI Model Access · AML.T0047
- 3LLM Prompt Injection: DirectExecution · AML.T0051.000
- 4Verify AttackAI Attack Adaptation · AML.T0042
- 5Prompt Infiltration via Public-Facing ApplicationInitial Access · AML.T0093
- 6AI Agent Tool InvocationExecution · AML.T0053
- 7Unsecured CredentialsCredential Access · AML.T0055
- 8External Harms: Financial HarmImpact · AML.T0048.000
- 9Denial of AI ServiceImpact · AML.T0029
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.