Thomaszhou22

Danger Guard Skill — AI skill for Claude Code

AI community

AI agent safety shield.

How to install Danger Guard Skill

This entry records only its repository, not the path inside it, so there is no exact command to give. Open Thomaszhou22/danger-guard-skill and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Danger Guard Skill does

AI agent safety shield. Intercepts dangerous commands, requires password verification, sends alerts. Works with OpenClaw, Claude Code, Cursor, and any agent platform.

Alternatives in AI

Golutra — Multi-agent AI orchestration platform for automation, workflows, and developer tools 3.8k ★
Kru — Local-first MCP password and credential manager for AI agents 125 ★
Anthropic Proxy Rs — A proxy server that intercepts Anthropic API requests and converts them to OpenAI-compatible format, enabling 90 ★

README

Danger Guard

AI agent safety shield. Intercepts dangerous commands before execution, requires password verification, and sends alerts when someone tries to destroy your system through a compromised account.

[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE) [![OpenClaw](https://img.shields.io/badge/OpenClaw-compatible-blue)](https://openclaw.ai) [![Claude Code](https://img.shields.io/badge/Claude%20Code-compatible-purple)](https://claude.ai) [![Cursor](https://img.shields.io/badge/Cursor-compatible-orange)](https://cursor.sh) [![Platform](https://img.shields.io/badge/platform-all%20agent%20platforms-green)](https://github.com/Thomaszhou22/danger-guard-skill)

[中文文档](README-CN.md)

🔌 **New: Gateway-level plugin version now available.** [Danger Guard Plugin](https://github.com/Thomaszhou22/danger-guard-plugin) moves the check into OpenClaw's `before_tool_call` hook — pattern matching in code, outside the model, impossible to forget or bypass. Skill + plugin are complementary: install both. ([ClawHub](https://clawhub.ai))

The Problem

Your AI assistant has full access to your machine. If someone compromises your messaging account — Feishu, Telegram, Discord, WhatsApp, whatever — they can tell your AI to `rm -rf /`, `format C:`, or `git push --force`. The AI just does it.

Danger Guard stops that.

How It Works

Dangerous command detected
        ↓
Immediate intercept (nothing executes)
        ↓
Warning + risk level displayed
        ↓
Requires sudo/admin password verification
        ↓
3 failed attempts → lockout + alert
        ↓
Password correct → execute
Password wrong → block + alert