AML.CS0027
Organization Confusion on Hugging Face
- Type
- Exercise
- When
- 2023-08-23
- Actor, as ATLAS names it
- threlfall_hax
- Target
- Hugging Face users
Techniques, in order
- 1Establish AccountsResource Development · AML.T0021
- 2ImpersonationDefense Evasion · AML.T0073
- 3Full AI Model AccessAI Model Access · AML.T0044
- 4External Harms: AI Intellectual Property TheftImpact · AML.T0048.004
- 5Manipulate AI Model: Embed MalwareAI Attack Adaptation · AML.T0018.002
- 6Publish Poisoned AI Artifacts: ModelsResource Development · AML.T0115.001
- 7Supply Chain Compromise: Model RepositoryInitial Access · AML.T0010.003
- 8User Execution: Unsafe AI ArtifactsExecution · AML.T0011.000
- 9MasqueradingDefense Evasion · AML.T0074
- 10Cyber Communication ChannelCommand and Control · AML.T0072
- 11Unsecured CredentialsCredential Access · AML.T0055
- 12Exfiltration via Cyber MeansExfiltration · AML.T0025
- 13Discover AI ArtifactsDiscovery · AML.T0007
- 14Obtain Capabilities: Adversarial AI Attack ImplementationsResource Development · AML.T0016.000
- 15Manipulate AI Model: Poison AI ModelAI Attack Adaptation · AML.T0018.000
- 16External HarmsImpact · AML.T0048
How these attacks work
Software used
- EasyEdit · Tool
- huggingface-cli · Tool
- Sliver · Tool
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.