---
title: "prompt-guard"
type: agent skill (OpenClaw)
category: "Security"
curated_by: Agentmatik (https://agentmatik.ai)
source: https://agentmatik.ai/skills/prompt-guard
---

# prompt-guard

Advanced prompt-injection defense for OpenClaw - screens incoming content (emails, web pages) for instructions that try to hijack your agent.

**Why we recommend it:** An agent that reads your inbox will eventually read a malicious email. This is the seatbelt.

- Install: Install from ClawHub: seojoonkim/prompt-guard
- Security note: ClawHub-listed

More curated skills: https://agentmatik.ai/skills · Agentmatik is an AI automation studio (https://agentmatik.ai).