Public instances

• https://cysectool.prem.moe (always up to date)
• http://hegel.eecs.qmul.ac.uk:5006/visualiser

Research

Read the paper here: https://arxiv.org/abs/2204.11707

CySecTool is a tool that finds a cost-optimal security controls portfolio in a given budget for a probabilistic attack graph. A portfolio is a set of counter-measures, or controls, against vulnerabilities adopted for a computer system, while an attack graph is a type of a threat scenario model. In an attack graph, nodes are privilege states of the attacker, edges are vulnerabilities escalating privileges, and controls reduce the probabilities of some vulnerabilities being exploited. The tool builds on an optimisation algorithm published by Khouzani et al. (2019), enabling a user to quickly create, edit, and incrementally improve models, analyse results for given portfolios and display the best solutions for all possible budgets in the form of a Pareto frontier. A case study was performed utilising a system graph and suspected attack paths prepared by industrial security engineers based on an industrial source with which they work. The goal of the case study is to model a supervisory control and data acquisition (SCADA) industrial system which, due to having the potential to harm people, necessitates strong protection while not allowing the use of typical penetration tools like vulnerability scanners. Results are analysed to show how a cyber-security analyst would use CySecTool to store cyber-security intelligence and draw further conclusions.

Requirements

• Python 3.7+
• virtualenv & pip OR conda

virtualenv way

Setup

git clone https://github.com/przemub/security-graph-visualiser
cd security-graph-visualiser
virtualenv3 venv
source venv/bin/activate
pip install -r requirements.txt

Run

source venv/bin/activate
python main.py

conda way

Setup

git clone https://github.com/przemub/security-graph-visualiser
cd security-graph-visualiser
conda env create -f environment.yml

Run

conda activate visualiser-env
python main.py

Test

Once you have a working environment:

python -m unittest

For testing the optimization routine, see test/test_model.py. The way to test a model is to subclass TemplateTestCase (when testing a built-in model) or FixtureTestCase (when testing an extra model used only for testing). In the first case, set model_number to the number of the model. In the latter, put the model in doc/fixtures folder and set fixture_name to the name of the model (without extension).

Afterwards, take a look at CustomAdjustmentTestCase to see the parameters that can be tested and set them in your test case.

For testing the GUI, see test/test_gui.py. These tests are run using headless Firefox and Chromium browsers instrumented by Selenium framework. Refer to Selenium with Python docs in order to learn how to use it.

Structure

• main.py - Bokeh server runner
• src - Python server-side source code
  • visualiser.py - analyzer/visualiser application code
    • drawing.py - graph drawing routines
    • api.py - API routes
  • data.py - attack graph data structures
    • optimization.py - optimization routines
  • edit.py - graph editor application code (see my_static/js/edit.js)
• my_static/js - JavaScript client-side code
  • edit.js - graph editor code
  • main.js - some fix-ups to Bokeh
  • callbacks.js - events handling
• templates - HTML templates of the sites, written in Jinja2
• tests - see Test
• doc - all kinds of research output and analysis tools
  • doc/templates - graph templates used in the application (probably should be moved)
  • doc/fixtures - graph templates used for automated testing