TA0001
Reconnaissance
Gathering information about the target AI system
TA0002
Initial Access
Gaining initial unauthorized interaction with the AI system
TA0003
Execution
Techniques for executing unauthorized actions through the AI system
TA0004
Persistence
Maintaining unauthorized influence over the AI system across interactions
TA0005
Privilege Escalation
Gaining elevated capabilities or bypassing permission boundaries
TA0006
Defense Evasion
Avoiding detection and bypassing safety mechanisms
TA0007
Collection
Extracting sensitive data and model information
TA0008
Exfiltration
Extracting data from the AI system to external destinations
TA0009
Impact
Causing harm through or to the AI system