OpenClaw Threat Model: MITRE ATLAS Security Analysis and Formal Verification
OpenClaw uses the MITRE ATLAS framework to analyze AI system threats, identifying three Critical risks (prompt injection, malicious skills, credential theft), and employs TLA+ formal verification for security properties.