A.I.G (AI-Infra-Guard)

What is A.I.G (AI-Infra-Guard)?

A.I.G (AI-Infra-Guard) is a powerful AI Red Teaming platform developed by Tencent Zhuque Lab. It is designed to help users identify and manage security risks in their AI infrastructure. A.I.G offers a range of features, including AI infra vulnerability scans, MCP server risk scans, and jailbreak evaluations, making it a comprehensive solution for AI security self-examination.

Benefits

A.I.G provides several key benefits for users looking to secure their AI infrastructure:

• Comprehensive Security Scans: A.I.G can identify over 30 AI framework components and cover nearly 400 known CVE vulnerabilities, ensuring a thorough security assessment.
• Intelligent Risk Detection: Powered by AI agents, A.I.G detects nine major categories of MCP security risks, providing users with advanced threat detection capabilities.
• User-Friendly Interface: The platform is designed to be easy to use, with a straightforward web interface that makes it accessible for both technical and non-technical users.
• Rapid Jailbreak Evaluation: A.I.G quickly assesses prompt security risks and generates jailbreak evaluation reports, helping users identify and mitigate potential threats.

Use Cases

A.I.G is suitable for a variety of use cases, including:

• AI Infrastructure Security: Organizations can use A.I.G to scan their AI infrastructure for vulnerabilities and ensure compliance with security standards.
• MCP Server Risk Management: A.I.G's MCP scan feature helps users detect and manage security risks in their MCP servers, ensuring robust protection.
• Prompt Security Assessment: A.I.G's jailbreak evaluation feature allows users to quickly assess the security of their prompts and identify potential risks.

Quick Start

Getting started with A.I.G is straightforward. Users can deploy the platform using Docker with one of the following methods:

• One-Click Install Script: This method automatically installs Docker and launches A.I.G with a single command.
• Run with Pre-Built Images: This method pulls pre-built images from Docker Hub for a faster start.
• Build from Source and Run: This method builds a Docker image from local source code and starts the service.

Once the service is running, users can access the A.I.G web interface athttp://localhost:8088.

Contribution Guide

A.I.G encourages community contributions through its extensible plugin framework. Users can contribute by adding new fingerprint rules, vulnerability rules, MCP plugins, and jailbreak evaluation datasets. Contributions can be submitted via a Pull Request.

Additional Information

A.I.G is licensed under the MIT License. For collaboration inquiries or feedback, users can contact the team at zhuque(at)tencent.com. Users can also join the A.I.G community group by scanning the WeChat QR code provided on the platform's website.