yee-yore/DorkAgent

What it solves

DorkAgent automates the process of "Google Dorking"—using advanced search queries to find security vulnerabilities, sensitive information disclosures, and exposed panels—specifically for bug hunting and penetration testing.

How it works

The tool uses an LLM-powered agent framework (CrewAI) to execute search queries via search engine APIs (Serper or SerpApi). It allows users to select their preferred LLM provider (OpenAI, Anthropic, or Gemini) and search engine at runtime. The agent performs reconnaissance and analyzes the results to generate security reports detailing potential attack vectors and information disclosures.

Who it’s for

Security researchers, bug hunters, and penetration testers who need to automate the discovery of exposed assets and vulnerabilities using Google search operators.

Highlights

• Multi-LLM Support: Compatible with GPT-4, Claude, and Gemini models.
• Flexible Search Engines: Supports both Serper and SerpApi for fetching Google results.
• Customizable Search Parameters: Users can adjust the number of results and the time duration of the results (e.g., past week or month).
• Automated Setup: Auto-installs required packages and handles API key configuration via a .env file on first run.