Home

Awesome

Nebula

Welcome to Nebula.

nebula

Galaxy

Acknowledgement

First i would like to thank the All-Mighty God who is the source of all knowledge, without Him, this would not be possible.

<span style="color:green">📢 Announcement📢 </span>

🌐 Introducing Nebula Pro: A New Era in Ethical Hacking 🌐

🚀 We're thrilled to unveil a sneak peek of Nebula Pro, our latest innovation designed to empower ethical hackers with advanced, AI-driven capabilities. After months of dedicated development, we have launched the preview version. Some of the exciting features are:

📺 Click Here to Get Access To Nebula Pro Now 🚀

DISCLAIMER

Please do not use this tool in a production environment, we are still experimenting and it is currently only suitable for testing environments

At the moment, to utilize the tools provided in this repository effectively, users are expected to possess a basic understanding of nmap, nuclei, zap, and crackmap.

For an AI assistant that supports almost any tool, consider neutron and Nebula Pro

Please be advised that this repository is currently in its beta phase; as such, occasional discrepancies or bugs might be encountered. We are diligently working towards refining and resolving all known issues.

Why nebula?

The youtube video below provide a quick example of how Nebula can speed up the process of identifying vulnerabilities

Nebula Usage Youtube Video

Disclaimer: Only NMAP is currently supported for smart analysis of tool output.

Overview

nebula Nebula is an AI-powered assistant specifically designed for the field of ethical hacking. It provides a unique capability for users to input commands using natural language processing, facilitating a seamless transition from intent to execution.

Additionally, Nebula offers a command search engine. Ethical hackers can effortlessly search for services, ports, or specific terms. In response, Nebula provides curated suggestions on commands that can aid in identifying potential vulnerabilities.

Currently, Nebula is integrated with and supports the following ethical hacking tools:

Our roadmap envisions Nebula's continuous expansion to incorporate the majority of the tools leveraged by ethical hackers globally. This commitment ensures that our users remain at the cutting edge of cybersecurity endeavors.

Compatibility

Non-Docker versions of Nebula have been extensively tested and optimized for Linux platforms. As of now, its functionality on Windows or macOS is not guaranteed, and it may not operate as expected.

System dependencies

Docker based distribution requirement(s)

PYPI based distribution requirement(s)

You can install exploitdb via APT or GIT, note that that will affect what you set the --exploit_db_base_location as. Linux (debian based):

APT based install:

sudo apt -y install exploitdb

GIT based install:

sudo git clone https://gitlab.com/exploit-database/exploitdb.git /opt/exploitdb
sudo ln -sf /opt/exploitdb/searchsploit /usr/local/bin/searchsploit

sudo apt install -y libreadline-dev

Linux (debian based):

sudo apt install -y wget

Installation

The easiest way to get started is to use the docker image. Please note that the ZAP model is NOT supported in the docker image. If you would like to use ZAP please install the package using pip.

PRO TIP: Regardless of if you are using the docker or pip version, always run nebula in the same folder so that it doesn't have to download the models each time you run it.

Docker:

Pulling the image:

docker pull berylliumsec/nebula:latest

Running the image without GPU:

docker run --rm -it berylliumsec/nebula:latest

To avoid downloading the models each time you run the docker container, mount a directory to store the models like so:

docker run --rm -v "$(pwd)":/app/unified_models_no_zap -it berylliumsec/nebula:latest

Running the model with all GPU(s)

docker run --rm --gpus all -v "$(pwd)":/app/unified_models_no_zap -it berylliumsec/nebula:latest

docker run --rm --gpus all -v "$(pwd)":/app/unified_models -it nebula:latest

For autonomous mode and to use the default vulnerability scan:

docker run --rm --gpus all -v "$(pwd)/targets.txt":/app/targets.txt -v "$(pwd)"/unified_models:/app/unified_models -v $(pwd)/targets:/app/targets -it nebula_testn:latest --autonomous_mode True --targets_list /app/targets.txt

For autonomous mode and to use your own nmap vulnerability scan:

docker run --rm --gpus all -v "$(pwd)/targets.txt":/app/targets.txt -v "$(pwd)"/unified_models:/app/unified_models -v $(pwd)/targets:/app/targets -it nebula_testn:latest --autonomous_mode True --nmap_vuln_scan_command="nmap -Pn -sV --exclude-ports 21 --script=vulscan/vulscan.nse" --targets_list /app/targets.txt

PIP:

pip install nebula-ai

To run nebula simply run this command:

nebula

For performing operations that require elevated privileges, consider installing via sudo

sudo pip install nebula-ai

Then run:

sudo nebula

Kali Linux Specific Post Installation Steps

By default, nebula is installed in the following path on Kali

"$HOME/.local/bin/.local/bin

To ensure that nebula is in your system path, add this to your .zshrc file

export PATH="$HOME/.local/bin:$PATH"

OPTIONAL nebula-watcher installation

PIP:

To install nebula-watcher:

pip3 install nebula-watcher

Docker

Pulling the image:

docker pull berylliumsec/nebula_watcher:latest

Running the docker image :

docker run --network host -v directory_that_contains_nmap_results/nmap_plain_text_or_xml:/app/results -v where/you/want/the/diagram:/app/output  berylliumsec/nebula_watcher:latest

To change the diagram name from the default:

docker run --network host -v directory_that_contains_nmap_results/nmap_plain_text_or_xml:/app/results -v where/you/want/the/diagram:/app/output  berylliumsec/nebula_watcher:latest python3 nebula_watcher.py --diagram_name /app/your_diagram_name

Upgrading

For optimal performance and to ensure access to the most recent advancements, we consistently release updates and refinements to our models. Nebula will proactively inform you of any available updates to the package or the models upon each execution.

PIP:

pip install nebula-ai --upgrade

pip3 install nebula-watcher --upgrade

Docker:

docker pull berylliumsec/nebula:latest

docker pull berylliumsec/nebula_watcher:latest

Usage.

In this beta release, there are three primary applications for Nebula:

As an auto-exploitation Engine

Using the autonomous mode, ethical hackers can supply a list of targets in a file named targets.txt. Nebula will run an NMAP vulnerability scan and then attempt to exploit the vulnerabilities using a combination of scripts and AI.

As a search engine:

Within the search engine capability, ethical hackers can input port numbers or specific service names. In return, they will receive recommended commands to assist in the identification of potential vulnerabilities.

Pro Tip: For optimal results, search using service names or port numbers. This approach is more effective than entering a full sentence or broad query.

As an AI-driven assistant.

DISCLAIMER: The results provided by this tool may contain inaccuracies or may not be suitable for all scenarios. We highly recommend users to review and, if necessary, modify the suggested commands before executing them. Proceed with caution and always ensure you are acting within legal and ethical boundaries Queries can be presented naturally to the AI-driven assistant, which then translates them into specific commands.

As a command suggestion engine.

Nebula can process results from NMAP scans (plain text or XML format) and suggest commands to run to detect vulnerabilities on services running on open ports.

As an ethical hacking coverage tool

Using the optional Nebula-Watcher, ethical hackers can automatically monitor the IP addresses and ports that they have engaged with during a penetration test to ensure maximum coverage .

DocuNebula.

Options

To view Nebula's options run

nebula -h

usage: nebula.py [-h] [--results_dir RESULTS_DIR] [--model_dir MODEL_DIR] [--testing_mode TESTING_MODE] [--targets_list TARGETS_LIST] [--autonomous_mode AUTONOMOUS_MODE]
                 [--attack_mode ATTACK_MODE] [--nmap_vuln_scan_command NMAP_VULN_SCAN_COMMAND] [--lan_or_wan_ip LAN_OR_WAN_IP] [--vuln_db_base_location VULN_DB_BASE_LOCATION]

Interactive Command Generator

options:
  -h, --help            show this help message and exit
  --results_dir RESULTS_DIR
                        Directory to save command results
  --model_dir MODEL_DIR
                        Path to the model directory
  --testing_mode TESTING_MODE
                        Run vulnerability scans but do not attempt any exploits
  --targets_list TARGETS_LIST
                        lists of targets for autonomous testing
  --autonomous_mode AUTONOMOUS_MODE
                        Flag to indicate autonomous mode
  --attack_mode ATTACK_MODE
                        Attack approach
  --nmap_vuln_scan_command NMAP_VULN_SCAN_COMMAND
                        Nmap vulnerability scan command to run
  --lan_or_wan_ip LAN_OR_WAN_IP
                        Your lan or wan ip for metasploit tests
  --vuln_db_base_location VULN_DB_BASE_LOCATION
                        the base location of vuln_db files

Modes.

Nebula can be run in autonomous mode or manual mode.

Autonomous Mode (Experimental).

Nebula can be run in autonomous mode or manual mode.

To activate the autonomous mode, run:

nebula --autonomous_mode True

By default it will run the vuln, exploit, vulnscan, service detection -sV, and treat the host as online -Pn. You can override this by passing in your own nmap command and scripts:

nebula --autonomous_mode True --nmap_vuln_scan_command 'nmap -Pn -sV --script=your_script'

There is no need to add output flags such as -oX and -oN as those will be automatically added.

Nebula will run an initial vulnerability scan, parse the results, attempt to discover more vulnerabilities. Your targets should be placed in a plain text file titled targets.txt. This is also customizable by using the --targets_list arg:

If Nebula recognizes any CVEs, it will try to exploit them. By default, the commands are limited to 1 commands per service/port. You can set the number of commands like this:

For up to 5 commands per service with open ports:

nebula --autonomous_mode True --attack_mode raid

To run every possible command per service with open ports:

nebula --autonomous_mode True --attack_mode war

If you want to do a dry-run of autonomous mode so that it does not attempt to exploit any vulnerabilities, set testing mode to True:

Note that testing mode will still perform a vulnerability scan, but it will not attempt to discover more vulnerabilities or attempt to exploit them

nebula --autonomous_mode True --testing_mode True

For bruteforce/password spraying attacks (if available or recommended by AI), provide these files usernames.txt and passwords.txt in the directory where you run nebula from.

Depending on how many IP addresses you provide, you may have several files to review. After Nebula is done in autonomous mode, it will drop into manual mode where you can view the results.

Nebula will try to execute available Metasploit modules using the standard remote ports, your local IP (or whatever ip you specify via args), and a random port. Even if you provide incorrect arguments to Metasploit, it won't shut down; it'll just notify you of the error. Hence, we've included placeholders for certain arguments, whether or not they're required.

{{ LHOST }}
{{ RHOST }}
{{ RPORT }}
{{ LPORT }}

In upcoming versions, we aim to improve this feature. As with other features, you can check out the outcomes in the view previous results menu option. For a deeper dive, launch the msfconsole and customize as needed. Our primary objective with the autonomous mode is to pinpoint exploits and demonstrate a quick proof of concept (POC) whenever possible using.

Note that results will only generally be written to a file if it is not empty.

Manual mode.

Upon initial access to Nebula, users are greeted with several options:

Enter a New Command: This prompt allows users to input commands using natural language. Subsequently, the system predicts and suggests a command for execution. Users have the discretion to either execute the generated command as is or modify it. After initiating the command, they can choose to await its completion or proceed with other tasks. Queries should be formulated as commands rather than questions, see the natural language examples section for more information.

Pro Tip: In the beta version, for optimal performance, limit your queries to a combination of up to two switches.

Enter a New Command

In the above screenshot, the user asks the NMAP model to perform a top 10 port-scan on 192.168.1.1.

View Previous Results: After a command's execution, users can review the output via this option.

View Previous Results

In the above screenshot, the user reviews the output of the top 10 port-scan.

Process Previous Results: Currently optimized for NMAP results (in plain text or XML format), Nebula enables users to select previous scan findings and obtain command suggestions to assess potential vulnerabilities of uncovered ports. Users have the flexibility to select a suggested command, modify it if needed, and execute. Support for results from other tools is planned for future releases.

Process Previous Results

In the above screenshot, the user asks Nebula to process the results of the top 10 port scan, Nebula provides suggestions on what commands to run, and the user chooses suggestion number #8 and runs the corresponding command after editing it.

Run Suggested Command

In the above screenshot, the user views the results of running the suggested command. In this case, the result is a list of HTTP methods supported by 192.168.1.1.

Select a Model: Users can choose from one of the three available natural language processing models (note that the actual name of the models/version may vary from what is seen in this screenshot).

Model Selection

Search by keywords: By leveraging this feature, users can input keywords—such as port numbers or service names—and obtain command suggestions to identify vulnerabilities related to that specific service.

Pro Tip: For optimal results, search using service names or port numbers. This approach is more effective than entering a full sentence or broad query.

Search

Results Storage: A folder named results is created (if it does not already exist) in the working directory where nebula is invoked.

Escape prompt screen without entering a prompt: To escape the prompt screen without entering a prompt, simply hit the enter key.

Examples of natural language queries

NMAP:

Note that for nmap commands, -oX and -oN are automatically appended. The plain text version is for you to be able to easily read while the xml version is for processing, please do not remove them

Examples:

CRACKMAPEXEC:

For crackmap, always include a username and a password in your prompt or indicate that you would like to use a null session:

ZAP:

The ZAP model currently only supports "full scan". Be sure to use the terms full scan in your commands.

Nuclei

Links to videos:

📺 Watch the Demo: Latest NMAP Model 🚀

🔍 Watch How to Use the Search Feature 🎥

Contributing and troubleshooting

Should you encounter inaccuracies in the model's responses to your natural language prompts, or face any other challenges, we kindly request that you create an issue to document the specifics. Your feedback is invaluable to our continuous improvement efforts.

You can also find error logs in command_errors.log in the folder where you run Nebula from