← Bibliography

OpenClaw agentic AI threat assessment

View original ↗

Note

Concluded OpenClaw maps to all 10 OWASP risks for agentic applications; lacks enforceable trust boundaries between untrusted inputs and high-privilege reasoning

Citation Key

unit42_2026_openclaw