AML.CS0008
ProofPoint Evasion
- Type
- Exercise
- When
- 2019-09-09
- Actor, as ATLAS names it
- Not stated
- Target
- ProofPoint Email Protection System
- ATLAS attributes it to Silent Break Security (security evaluation).
Techniques, in order
- 1Discover AI Model OutputsDiscovery · AML.T0063
- 2Outputs from AI ModelsCollection · AML.T0135
- 3Replicate AI Model Behavior: Replicate Behavior from Model OutputsAI Attack Adaptation · AML.T0005.001
- 4Craft Adversarial Data: Black-Box TransferAI Attack Adaptation · AML.T0043.002
- 5Evade AI ModelImpact · AML.T0015
How these attacks work
References
- National Vulnerability Database entry for CVE-2019-20634
- 2019 DerbyCon presentation "42: The answer to life, the universe, and everything offensive security"
- Proof Pudding (CVE-2019-20634) Implementation on GitHub
- 2019 DerbyCon video presentation "42: The answer to life, the universe, and everything offensive security"
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.