🔒 Security

Plugin Approval Guard

VirusTotal for OpenClaw skills — static analysis trust scanner. Audits skill code for exfiltration, shell injection, eval/exec abuse. Generates A-F trust grades.

What it does

Plugin Approval Guard is a static analysis security scanner for OpenClaw skills — think VirusTotal, but for agent code. It audits every .py file in a skill for data exfiltration patterns, shell injection vectors, eval/exec abuse, obfuscated payloads, and hidden network calls, then assigns an A-F trust grade with detailed findings.

For teams managing shared skill libraries or installing third-party skills from ClawHub, Plugin Approval Guard integrates with OpenClaw 3.28's approval hooks to automatically block low-trust skills before they can execute. No more blindly trusting code you didn't write.

Use cases

• Enterprise teams managing shared skill libraries across multiple agents
• Security-conscious individuals who install third-party skills from ClawHub
• OpenClaw administrators who want automated pre-install vetting in their approval workflow

Requirements

• OpenClaw 3.20+ (3.28+ for approval hook integration)
• Python 3.10+
• No API key required — pure static analysis

How to install

After purchase you'll receive a download link by email. Extract the files and copy the plugin-approval-guard/ folder into your OpenClaw skills/ directory, then restart OpenClaw.