Skip to content

katrinewi/Docker-image-analyzing-tools

Repository files navigation

Docker image analyzing tools

This repository contains scripts that can be used for performing an analysis of Docker Hub images. The web scraper collects images from the Docker Hub web page along with the image type, the ImageInfo scripts gather additional image information, and the automate script runs the Anchore Engine vulnerability scanner automatically on images with the latest tag.

Web Scraper

Requires the following dependencies:

In the scraper script, we specify the driver on line 34. The path must be changed to the location of the Gecodriver executable on the user's computer. Alternatively, it can be added to the PATH by placing it in the /usr/bin or /usr/local/bin folder, and remove everything inside the brackets on line 34 in the script. It is important to use versions of Selenium, Geckodriver, and Firefox that are compatible.

The scraper creates two files: image-names.txt and image-info.csv, and writes the gathered data to them. If these files already exist, the content inside will be overwritten.

ImageInfo scripts

These scripts create two files each: results_apiv1.csv and failed_apiv1.txt, and results_apiv2.csv and failed_apiv2.txt. The gathered data is written to these files. If they already exist, the content inside will be overwritten. Additionally, the scripts take the image_names.txt file as input, which constitutes of image names separated by line shift. If this file is not to be found, the scripts will not run.

Automate script

Requires the following dependencies:

In order to run the script, Docker Compose is required to run. Docker Compose is started with the following command: docker-compose up -d. The script will first try to rename the vuln.csv and the failed.txt files. Because of the possible long run time of the script, it is important that this content is not overwritten by mistake. Thus, an error will be outputted if these files are not found, but the script will continue to run as expected. The script takes the image_names.txt file as input, which constitutes of image names separated by line shift. If this file is not to be found, the script will not run. Docker is by default running as root, and thus, needs to be run using the SUDO command. We highly recommend running Docker as a non-root user.

Referencing our work

If you use our scripts in your work please use the following reference:

@mastersthesis{WistHelsem2020, title={{An Extensive Analysis of the Current Vulnerability Landscape in Docker Hub Images}}, author={Wist, Katrine and Helsem, Malene}, year={2020}, school={Norwegian University of Science and Technology (NTNU)} }

About

Scripts for performing analysis of Docker images

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published