agentguard

Scan codebases for vulnerabilities and block unsafe commands.

26|5|Updated Feb 19, 2026

Search Tags:#security #vulnerability #agent #command execution #guard #data leak

Ready to run

npx skills add https://github.com/kunwl123456/zeroclaw-skills-collection --skill agentguard-kunwl123456

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agentguard
Source: https://github.com/kunwl123456/zeroclaw-skills-collection/tree/main/security
Command: npx skills add https://github.com/kunwl123456/zeroclaw-skills-collection --skill agentguard-kunwl123456

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @goplus/agentguard-sdk, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill acts as an AI agent security guard, automatically blocking dangerous commands, preventing data leaks, and protecting secrets to ensure safe AI operations.

Core Features & Use Cases

Command Blocking: Prevents the execution of malicious or unsafe shell commands.
Data Leak Prevention: Scans for and blocks attempts to exfiltrate sensitive information like API keys or secrets.
Vulnerability Auditing: Analyzes code and actions for potential security risks.
Use Case: When reviewing a new third-party skill, use AgentGuard to scan its codebase for vulnerabilities and evaluate the safety of its proposed actions before allowing it to run.

Quick Start

Scan the 'my-new-skill' directory for security risks.

Frequently Asked Questions about agentguard

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I prevent AI agents from executing dangerous shell commands?▼

To prevent AI agents from executing dangerous shell commands, you need an AI agent security framework that automatically blocks malicious command execution. This Skill scans actions and blocks unsafe shell commands before they run.

How do I audit third-party code for security vulnerabilities before execution?▼

You audit third-party code for security vulnerabilities by scanning its codebase to evaluate action safety. This Skill analyzes code for potential security risks and provides reports on proposed actions before allowing them to run.

What is the best way to scan AI skills for data leaks and exposed secrets?▼

The best way to scan AI skills for data leaks and exposed secrets is to use a security guard framework that detects exfiltration attempts. This Skill scans for sensitive information like API keys and blocks attempts to leak data.

Can I configure different protection levels for command execution safety?▼

Yes, you can configure different protection levels for command execution safety. This Skill supports configuring protection levels, trust management, and viewing security logs to evaluate action safety dynamically.

Do I need the agentguard-sdk dependency to review action safety?▼

Yes, you need the agentguard-sdk dependency to review action safety. This Skill relies on the SDK to provide the core security framework for scanning, blocking dangerous commands, and generating security logs.

Why does my AI agent framework fail to block malicious code vulnerabilities?▼

Your AI agent framework fails to block malicious code vulnerabilities without a dedicated security guard. This Skill provides vulnerability auditing and action evaluation to identify and block risks before execution.