Common vulnerability scoring system prediction based on open source intelligence information sources

This repository hosts the code used in the paper Common vulnerability scoring system prediction based on open source intelligence information sources [1], which was originally a master's thesis by David Relke.

[1]: Kuehn, P., Relke, D. N., & Reuter, C. (2023). Common vulnerability scoring system prediction based on open source intelligence information sources. Computers & Security. https://doi.org/10.1016/j.cose.2023.103286

Requirements

Python 3.10
Pipenv
Gecko or Chrome Driver
For Remote Driver: Docker

Web Scraping

Define Sources In main.py with sources = ['zerodayinitiative.com', 'ibm.com', 'tools.cisco.com', 'support.f5.com', 'www.qualcomm.com', 'www.intel.com', 'talosintelligence.com', 'snyk.io']

MongoDB will return all referenced URLs that match the sources. By default, all URLs are returned, even if there is already text from a previous run. This behavior can be adjusted with the parameters of MongoHandler.get_reference_list_for_url. Possible options: only_scraped (only references that have already been visited), only_not_scraped (only references that have not yet been visited) and only_scraped_but_no_text (only references that have already been visited and for which there is still no text).

Number of Threads

The default is NUMBER_OF_THREADS = 4 and can be adjusted there. It corresponds to the number of Selenium Firefox WebDrivers that are started.

Classifier

cvss_predictor.py supports two tasks: train and eval

The following parameters must be set:

--target: The target variable, i.e. the CVSS component. One from ['av', 'ac', 'pr', 'ui', 's', 'c', 'i', 'a']

--checkpoint_path: The file path under which the models are located or should be stored

--model_checkpoint: The checkpoint to be used for further training or evaluation. 'latest' uses the most recent, 'base' the untrained base model, otherwise the iteration number of the checkpoint.

--base_model: The base model. Provided are ['distilbert-base-uncased', 'prajjwal1/bert-small', 'prajjwal1/bert-medium'].

--eval_file: The file in which the results of the evaluation are saved. .json is automatically attached.

--dataset_path: Path to the folder containing test.json and train.json.

--epochs: Optional number of epochs to be trained for.

Examples:

python cvss_predictor.py eval --target av --checkpoint_path "path/to/models" --model_checkpoint final --base_model "distilbert-base-uncased" --eval_file "my/eval/dir/res_distil_av", --dataset_path "my/distilbert/dataset/path"

Will execute evaluation for AV on the last checkpoint of the DistilBERT model in path/to/models and write the result to my/eval/dir/res_distil_av.json. The data set my/distilbert/dataset/path/test.json is used for the evaluation.

python cvss_predictor.py train --target ui --checkpoint_path "path/to/models" --model_checkpoint base --base_model "prajjwal1/bert-medium" --eval_file "my/eval/dir/res_bert-medium_ui", --dataset_path "my/bert_medium/dataset/path" --epochs 6

Will train the untrained Bert-medium model for UI with the dataset my/bert_medium/dataset/path/train.json for 6 epochs and store the checkpoints in path/to/models.

Citing

If you make use of this code in your work, please cite the following paper:

@article{Kuehn_Relke_Reuter_2023,
    title={Common vulnerability scoring system prediction based on open source intelligence information sources},
    rights={All rights reserved},
    ISSN={0167-4048},
    url={https://www.sciencedirect.com/science/article/pii/S0167404823001967},
    DOI={10.1016/j.cose.2023.103286},
    journal={Computers & Security},
    author={Kuehn, Philipp and Relke, David N. and Reuter, Christian},
    year={2023}
}

Contributors

Philipp Kühn
David Relke
Christian Reuter

Acknowledgements

This work was supported by the German Federal Ministry for Education and Research (BMBF) in the project CYWARN (13N15407) and German Federal Ministry of Education and Research and the Hessian Ministry of Higher Education, Research, Science and the Arts within their joint support of the National Research Center for Applied Cybersecurity ATHENE.

License

MIT License

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

PEASEC/Open-Information-CVSS-Prediction