Safety for Agents — ADR layer for OpenClaw
Guardrail plugin for OpenClaw that blocks prompt injection, jailbreak attempts, and sensitive credential exfiltration before agent dispatch.
