CrowdStrike published 156 security advisories for OpenClaw. 4 Critical, 52 High. Four attack vectors identified: direct prompt injection, indirect prompt injection, tool chain attacks, AI tool poisoning.

They also demoed blocking a live Discord exfiltration attack targeting an OpenClaw instance. Not theoretical.

https://clawhosters.com/blog/posts/crowdstrike-openclaw-security-assessment

#OpenClaw #InfoSec #CrowdStrike #AIAgents