AML.CS0049
Supply Chain Compromise via Poisoned ClawdBot Skill
- Type
- Exercise
- When
- 2026-01-26
- Actor, as ATLAS names it
- Not stated
- Target
- ClawdBot (now OpenClaw)
Techniques, in order
- 1Develop CapabilitiesResource Development · AML.T0017
- 2Acquire Infrastructure: DomainsResource Development · AML.T0008.002
- 3LLM Prompt CraftingAI Attack Adaptation · AML.T0065
- 4Publish Poisoned AI Artifacts: AI Agent ToolsResource Development · AML.T0115.002
- 5AI Supply Chain Reputation InflationDefense Evasion · AML.T0111
- 6Supply Chain Compromise: AI Agent ToolInitial Access · AML.T0010.005
- 7AI Agent Tool Poisoning: Definition and InstructionsPersistence · AML.T0110.000
- 8User Execution: Poisoned AI Agent ToolExecution · AML.T0011.002
- 9LLM Prompt Injection: IndirectExecution · AML.T0051.001
- 10MasqueradingDefense Evasion · AML.T0074
- 11AI Agent Tool InvocationPrivilege Escalation · AML.T0053
- 12External HarmsImpact · AML.T0048
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.