🛡️ CommandGuard

Never run a dangerous command again.

CommandGuard is an AI-powered terminal safety layer that intercepts risky commands before they execute. It leverages GitHub Copilot CLI to explain what a command does and suggests safer alternatives.

🚀 Features

• Real-time Interception: Hooks into your shell (bash, zsh) to catch commands on the fly.
• AI Analysis: Uses GitHub Copilot to understand the intent of a command, not just regex matching.
• Risk Assessment: Classifies commands by severity.
• Educational: Explains why a command is flagged and teaches safer habits.
• Cross-Platform: Works on macOS, Linux, and Windows (via PowerShell or WSL).

📦 Installation

Prerequisites:

• Node.js (v16+)
• GitHub CLI (gh) with Copilot extension installed (gh extension install github/gh-copilot)

From Source

git clone https://github.com/Aryakoste/commandguard.git
cd commandguard
npm install
npm run build
npm link
commandguard setup

Install via npm (Public Registry)

npm install -g commandguard
commandguard setup

Install via GitHub

npm install -g github:Aryakoste/commandguard
commandguard setup

Install from Source (For development)

🛠️ Usage

Once installed and set up, CommandGuard runs automatically in your terminal.

1. Type a command.

   rm -rf /

2. CommandGuard intercepts. It checks if the command matches known dangerous patterns.
3. AI Analysis. If flagged, it asks GitHub Copilot for a detailed explanation and risk assessment.
4. Decision Time. You get a beautiful prompt:
   • 🚫 Abort: Stop the command.
   • ✅ Proceed: Run it anyway (if you really mean it).
   • 💡 Alternative: Run a suggested safer command.

⚙️ Configuration

You can customize CommandGuard's sensitivity and rules.

commandguard config

📄 License

Distributed under the MIT License. See LICENSE for more information.