AUTOATTACKER

AUTOATTACKER IMPLEMENTATION

Table of Contents

• Features
• Prerequisites
• Installation on macOS
  • 1. Install Conda
    • For Intel Macs
    • For Apple Silicon Macs (M1/M2)
  • 2. Set Up the Conda Environment
  • 3. Install Project Dependencies
  • 4. Verify the Installation
• Usage
• Troubleshooting
  • Common Issues and Solutions
  • Getting Help

Features

• Automates penetration testing tasks using AI.
• Integrates with various tools like sqlmap and playwright.
• Provides an interactive command-line interface.

Prerequisites

• macOS (Intel or Apple Silicon)
• Python 3.10 (managed by Conda)
• Xcode Command Line Tools (for compiling packages)

Installation on macOS

1. Install Conda

Conda is an open-source package and environment management system that simplifies package installation and management. We'll use Miniforge, a community-driven minimal installer for Conda, which is preferred for its simplicity and compatibility.

For Intel Macs

Download Miniforge for Intel Macs:

curl -LO https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-MacOSX-x86_64.sh

## Install Miniforge

```bash
bash Miniforge3-MacOSX-x86_64.sh

• Follow the on-screen prompts.
• Accept the license agreement.
• Choose the installation location (default is recommended).
• Allow the installer to initialize Conda.
• Restart Terminal to apply changes.

For Apple Silicon Macs (M1/M2)

Download Miniforge for Apple Silicon Macs:

curl -LO https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-MacOSX-arm64.sh

Install Miniforge:

bash Miniforge3-MacOSX-arm64.sh

• Follow the on-screen prompts.
• Accept the license agreement.
• Choose the installation location (default is recommended).
• Allow the installer to initialize Conda.
• Restart Terminal to apply changes.

2. Set Up the Conda Environment

Create a New Conda Environment with Python 3.10:

conda create -n pentestgpt_env python=3.10

Activate the Conda Environment:

conda activate pentestgpt_env

Verify the Python Version and Architecture:

python --version
python -c "import platform; print(platform.machine())"

• Ensure the Python version is 3.10.x.
• For Intel Macs, platform.machine() should return x86_64.
• For Apple Silicon Macs, it should return arm64.

3. Install Project Dependencies

Before installing dependencies, ensure you have Xcode Command Line Tools installed, as some packages require compilers.

Install Xcode Command Line Tools

xcode-select --install

• Follow the prompts to complete the installation.

Navigate to the Project Directory

If you haven't cloned the repository yet:

cd pentestgpt

Install Dependencies with Conda and Pip

Some packages are available via Conda, and others via Pip.

Install Dependencies via Conda:

conda install -c conda-forge \
  cffi \
  cryptography \
  greenlet \
  orjson \
  pydantic \
  sqlalchemy \
  lxml \
  pillow \
  numpy

Install Remaining Dependencies via Pip:

pip install --upgrade pip setuptools wheel
pip install -r requirements.txt

If your project uses Poetry, install dependencies using Poetry:

pip install poetry
poetry install

Note: Ensure that pyproject.toml is present in the project directory.

4. Verify the Installation

Activate the Conda environment (if not already active):

conda activate pentestgpt_env

Run the Application:

pentestgpt-connection

If the command is not found, ensure that the package scripts are installed correctly. You might need to adjust your PATH or run the application using Python:

Test Specific Commands:

pentestgpt-connection

Usage

After successful installation, you can start using AutoAttacker by running:

python autoattacker.py

Follow the on-screen instructions to interact with the tool.

Troubleshooting

Common Issues and Solutions

1. pip Command Not Found Ensure that you have activated the Conda environment:

   conda activate pentestgpt_env

2. Architecture Mismatch Errors If you receive errors related to architecture mismatch (e.g., trying to install packages built for arm64 on x86_64), ensure that:

   • You're using the correct Miniforge installer for your Mac's architecture.
   • Your terminal is not running under Rosetta (for Apple Silicon Macs).

3. Compilation Errors During Package Installation Some packages require compilation from source. Ensure that:

   • Xcode Command Line Tools are installed:

     xcode-select --install

   • You have installed necessary build dependencies via Conda.
   • Rust is installed for packages like orjson and tiktoken:

     curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
     source $HOME/.cargo/env

4. Issues with playwright Installation For playwright, you might need to install browser binaries:

   playwright install

5. Environment Activation Issues If the Conda environment doesn't activate properly, try initializing Conda for your shell:

   conda init $(basename $SHELL)
   source ~/.bashrc  # or ~/.zshrc depending on your shell

Getting Help

• GitHub Issues: Open an issue at GitHub Issues for assistance.

Note: The setup and config is based on a windows laptop that i currently use and depending on your command line setup and other system setting the configuration and system setup might differ

Setup and Configuration

1. Set PowerShell Execution Policy:

   Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser

Setup in VSCode

• PS C:\Users\cghdhv\venv> cd venv
• PS C:\Users\cghdhv\venv> .\Scripts\Activate.bat
• PS C:\Users\cghdhv\venv> .\Scripts\activate
• pip3 install -e .

Setup in VSCode for environment variables

• $Env:OPENAI_KEY = ''
• $Env:PINECONE_API_KEY = ''

check connection in VS-code

• PS C:\Users\cghdhv\venv> pentestgpt-connection

Create a Pinecone account

• Get the API key
• Setup index, check for the settings in vectorDB.py, and update the parameters there if needed.

check Experience Manager working

• python TestExperienceManger.py
• This will check the working of Experience mainly dependent on the VectorDB.py implementation

AutoAttacker Pipeline

• There are two auto attacker pipelines one is using GPT-4 and the other using ThinkGPT, autoattacker.py, and autoattacker_thinkGPT.py
• They can be run after experience manager's working has been tested

The file logs will be stored in the pentestgpt/config/logs/ folder