Nebula, an open-source security platform developed by BerylliumSec, is transforming the landscape of penetration testing by integrating advanced AI models directly into the command-line interface. This innovative tool allows security professionals to automate the process of vulnerability assessments and generate exploit scripts efficiently. By maintaining a seamless workflow without the need to switch contexts, Nebula enhances the productivity of ethical hackers.
Flexible AI Backend Support
Nebula offers support for various AI backends, giving users the flexibility to choose models based on their infrastructure and privacy considerations. The platform supports multiple models, including OpenAI’s API-accessible options, Meta’s Llama-3.1-8B-Instruct, Mistral AI’s Mistral-7B-Instruct-v0.2, and DeepSeek-R1-Distill-Llama-8B. This versatility ensures that users can select the most appropriate model for their specific needs.
The platform employs Ollama for local inference, which can operate on both CPU and GPU, while remote models are accessible via API keys. This seamless integration with existing command-line tools and security utilities, such as Nmap and Metasploit, ensures that Nebula can complement rather than replace established workflows.
Key Features and Enhancements
Nebula boasts several features that enhance its usability and effectiveness in penetration testing. Among these are AI-driven search agents that provide real-time cybersecurity context, automated note-taking capabilities, and real-time exploitation suggestions based on terminal outputs. The tool also includes built-in screenshot capture and annotation, alongside a status feed that updates every five minutes to reflect ongoing testing activities.
Users can interact with Nebula’s AI by prefixing commands with “!” or toggling a dedicated AI/Terminal mode button, facilitating a smooth transition between manual and AI-assisted operations. Such features make Nebula a versatile and powerful tool for cybersecurity professionals.
Installation and Future Developments
To install Nebula, users need at least 16GB of RAM and Python 3.10–3.13.9 for CPU-based inference via Ollama. The installation process is straightforward via pip, and users can deploy the tool locally or through Docker, with options for GUI support.
Complementing Nebula, BerylliumSec has launched the Deep Application Profiler (DAP), a malware analysis service that uses neural networks for zero-day malware detection, moving beyond traditional signature-based methods. Looking ahead, BerylliumSec plans to develop custom models tailored to penetration testing tasks, enhancing Nebula’s capabilities.
The integration of AI into penetration testing is indicative of a broader trend towards the inclusion of large language models in offensive security tools. Nebula, with its support for both local and cloud-based models, addresses diverse operational security requirements, offering a comprehensive solution for modern cybersecurity challenges.
