How salah9003/automated-vulnerability-scanning-with-agentic-ai works
Main Components
• Controller/Main Process: The execution starts in main.py, which includes configuration options (target IP, scan description, API key) and orchestrates the workflow.
• AI Agents: The Agents directory contains specialized agent modules (e.g., ammar.py, hassan.py, kofahi.py, rakan.py, salah.py, sajed.py), each fulfilling a distinct role:
• Ammar: Strategy generation.
• Hassan: Reviewing strategies and outputs.
• Kofahi: Error handling.
• Rakan: Monitoring command execution.
• Salah: Command execution.
• Sajed: Report generation.
• Core Modules: agent.py and utils.py likely serve as the common interface and support functions for the agents and the scanning workflow.
• External Dependencies: External system tools from Kali Linux (for vulnerability scanning) and the OpenAI API (for natural language understanding and command synthesis).
• Reporting and Logging: Although not explicitly in the file tree, the README describes log generation in a Logs directory and a markdown report generation (findings_report.md generated by Sajed).
Interactions
• Main Process Coordination: main.py acts as the entry point and manages the scan’s flow. It likely invokes the agents (through agent.py or directly) and feeds necessary configuration.
• External Interactions: Connections to the OpenAI API for AI functionalities and the integration of external scanning commands (using Kali tools) should be highlighted.
Architectural Patterns & Design Principles
• Multi-Agent Architecture: The design follows a distributed, specialized agent model, where each agent operates independently but contributes to an overall scanning strategy.
• Modular Design: With distinct modules for each agent and utility functions, the system emphasizes separation of concerns, making it easier to maintain and extend.
Technology and Frameworks
• Python is the primary language.
• Use of external libraries for API interactions (OpenAI API) and possibly for executing system commands.
• Integration with Kali Linux’s penetration testing tools.
Final Recommendations
• Annotations or brief notes on each component about its role, especially concerning how agents work together to adapt and improve the scanning process.