AI security scanner for OpenClaw - powered by AgentTinman. Discovers prompt injection, tool exfil, context bleed, and other security issues in your AI assistant sessions, then proposes mitigations mapped to OpenClaw's security controls.