Skip to main content
Panguard AI is an open-source security platform that protects both AI agents and endpoints. As the security layer for the OpenClaw skill ecosystem, Panguard scans AI skills for prompt injection, tool poisoning, and hidden threats before they reach your agents. It also brings enterprise-grade endpoint protection to developers and small businesses — without requiring a security team. The core philosophy is simple: when something happens, Panguard tells you. When nothing happens, you do nothing.

Skill Auditor

Scan AI skills for security threats before installing. The App Store review process for the OpenClaw ecosystem.

Quick Start

Get from zero to your first security scan in under 5 minutes.

Product Overview

Explore the full product suite: Skill Auditor, Scan, Guard, Chat, Trap, Report, and more.

OpenClaw Integration

How Panguard fits into the OpenClaw AI skill ecosystem as the trust layer.

​
Who is Panguard AI for?

AI Agent Developers

Vet third-party skills before installing them. Use Skill Auditor as a pre-install gate in your agent pipeline, or integrate it into CI for skill repositories.

Developers and SMBs

Secure your servers and development machines with zero configuration. Auto-generate ISO 27001, SOC 2, and Taiwan TCSA compliance reports.

Managed Service Providers

Monitor and protect multiple client endpoints from a single dashboard. Enforce skill audit policies across your agent fleet.

​
Why Panguard AI?

AI agents are gaining powerful capabilities through skill ecosystems like OpenClaw — but every installed skill is a trust decision. Meanwhile, traditional endpoint security tools require deep expertise and cost six figures. Panguard AI takes a different approach:
  • AI skill security — scan skills for prompt injection, tool poisoning, hidden Unicode, and encoded payloads before installation
  • One command install — no configuration files, no tuning, no jargon
  • Plain language notifications — receive alerts via Telegram, Slack, or Email in language you understand
  • AI auto-judges severity — the system triages, responds, and reports without human intervention
  • Gets smarter over time — behavioral baselines adapt to your specific environment

​
Quick install

curl -fsSL https://get.panguard.ai | bash
Then authenticate and run your first scan: 
panguard login
panguard scan --quick
The Community plan is free and includes unlimited scans plus Layer 1 rule-based protection. No credit card required.

​
Next steps

Installation

Detailed platform-specific installation instructions for macOS, Linux, and Windows.

Account Setup

Create your account, choose a plan, and configure your CLI session.

Concepts

Understand the three-layer AI architecture that powers Panguard.

Guides

Step-by-step guides for every feature, from scanning to honeypots.