AML.CS0041
Rules File Backdoor: Supply Chain Attack on AI Coding Assistants
- Type
- Exercise
- When
- 2025-03-18
- Actor, as ATLAS names it
- Not stated
- Target
- Cursor, GitHub Copilot
- ATLAS attributes it to Pillar Security (security evaluation).
- On ShieldMap: GitHub Copilot, Cursor (matched by name in the target).
Techniques, in order
- 1Stage CapabilitiesResource Development · AML.T0079
- 2LLM Prompt CraftingAI Attack Adaptation · AML.T0065
- 3LLM Prompt ObfuscationDefense Evasion · AML.T0068
- 4Supply Chain Compromise: SoftwareInitial Access · AML.T0010.001
- 5Modify AI Agent ConfigurationPersistence · AML.T0081
- 6LLM Prompt PropagationLateral Movement · AML.T0143
- 7LLM Prompt Injection: DirectExecution · AML.T0051.000
- 8LLM JailbreakDefense Evasion · AML.T0054
- 9LLM Trusted Output Components ManipulationDefense Evasion · AML.T0067
- 10External Harms: User HarmImpact · AML.T0048.003
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.