It works like a red team within your system. You describe the task in plain language — then it plans the attack itself, selects tools, and proceeds through the entire process: from reconnaissance to reporting. Without manual fiddling and endless commands.
What it can do in practice:
➡️ Checks everything: code, business logic, network traffic, protocols.
➡️ Analyzes the found vulnerabilities and explains where the problem is and how to fix it.
➡️ Works autonomously — you just launch it and get a full-fledged research.
➡️ Connects MCP servers and tools (nmap, metasploit, ffuf, etc.) itself.
➡️ Uses Pentesting Task Trees for meaningful decision-making, not just brute force.
➡️ Supports ready-made workflows for comprehensive checks.
➡️ Generates detailed reports in Markdown with facts and recommendations.
➡️ Remembers the dialogue context and doesn't "get lost" after a couple of requests.
➡️ Sees real files: wordlists, payloads, configs — and uses them in its work.
➡️ Allows you to choose an AI model and customize its behavior.
➡️ No registration and no restrictions.
⚙️ Installation:
git clone https://github.com/GH05TCREW/ghostcrew.git
cd ghostcrew
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
▶️ Usage:
python main.py
⚠️ The information is provided solely for informational purposes. And it encourages to pay attention to security issues.
♎️ GitHub/Instructions
#python #soft #github