OpenClaw has a process of flagging skills as potentially malicious. AI developers may still be exposed by malicious skills, currently only handled by an imperfect moderation and screening process.