Agents-MCP-Hackathon/ExposureGPT

ExposureGPT - Simplified OSINT Intelligence

Single MCP tool for comprehensive security intelligence using Shodan + OpenAI

 

Features

• —Single Tool: One comprehensive OSINT intelligence gathering function
• —Shodan Integration: Real infrastructure and device discovery
• —AI Analysis: GPT-4o-mini powered security insights
• —MCP Server: Built-in Model Context Protocol server for AI assistants
• —Risk Assessment: Automated security scoring and recommendations
• —Simple Interface: Single input, comprehensive output

Configuration

Required: Set these environment variables in your Space settings:

• —SHODAN_API_KEY - Your Shodan API key (get from https://shodan.io)
• —OPENAI_API_KEY - Your OpenAI API key (get from https://openai.com)

MCP Integration

This Space automatically serves as an MCP server that AI assistants like Claude can use!

MCP Endpoint: https://acloudcenter-exposuregpt.hf.space/gradio_api/mcp/sse

Claude Desktop Configuration:

{
  "mcpServers": {
    "exposuregpt": {
      "command": "npx",
      "args": ["mcp-remote", "https://acloudcenter-exposuregpt.hf.space/gradio_api/mcp/sse"]
    }
  }
}

Available Tool

`intelligence_gathering(target: str)`

• —Comprehensive OSINT analysis for any domain, IP address, or organization
• —Uses Shodan for infrastructure discovery and vulnerability detection
• —AI-powered analysis with actionable security recommendations
• —Returns detailed security report with risk assessment

Usage Examples

Web Interface

• —Domain: google.com - Analyze domain infrastructure
• —IP Address: 8.8.8.8 - Scan specific IP for services
• —Organization: Microsoft Corp - Corporate intelligence gathering

Via AI Assistant (Claude)

"Analyze the security posture of example.com"
"What are the security risks for tesla.com?"
"Perform OSINT analysis on 1.1.1.1"

Security & Ethics

This tool is designed for:

• —Security awareness and education
• —Authorized penetration testing
• —Risk assessment for your own organization
• —Academic research

Not for:

• —Unauthorized reconnaissance
• —Malicious activities
• —Privacy violations

How It Works

1. 1.Input Analysis: Automatically detects if target is domain, IP, or organization
2. 2.Shodan Query: Searches for exposed infrastructure and services
3. 3.Risk Assessment: Analyzes vulnerabilities and calculates risk scores
4. 4.AI Analysis: GPT-4o-mini generates security insights and recommendations
5. 5.Comprehensive Report: Formatted intelligence report with actionable findings

Perfect for security researchers, penetration testers, and AI assistants needing OSINT capabilities.

Built for the 2025 Gradio Agents & MCP Hackathon