Gemini - The Runtime Application Self Protection (RASP) Solution Combined With Deep Learning
Gemini-Self-Protector pioneers the fusion of Runtime Application Self Protection (RASP) and transformative Deep Learning. In today's evolving digital landscape, intelligent and adaptive application security is paramount. Enter Gemini-Self-Protector, ushering in a new era of proactive defense that revolutionizes application safeguarding amid ever-changing threats.
By seamlessly integrating RASP into your application's runtime fabric, Gemini-Self-Protector achieves unparalleled protection. It dynamically monitors and secures various aspects of functionality—database interactions, file operations, and network communications. This symbiosis with Deep Learning empowers Gemini-Self-Protector to adapt and evolve defenses in real-time, staying ahead of emerging threats.
👉 G-SP : gemini-self-protector
👉 G-WVD : gemini-web-vulnerability-detection
👉 G-BD : gemini-bigdata
The architecture of gemini-self-protector is composed of seven layers however it is optimized so as not to affect the performance on the application.
Language | Platform/ Framework |
---|---|
Python | Flask |
Gemini uses a deep learning model that combines Convolutional Neural Network (CNN) and a family of Recurrent neural network (RNN) techniques to detect and identify vulnerabilities.
For more details: G-WVD-DL
📜 All about Gemini-Self-Protector is in here
pip install gemini_self_protector
⚙️ See detailed installation instructions here
Gemini supports 3 modes and recommends sensitivity levels for the application to operate at its best state.
Mode | Sensitive |
---|---|
off | N/A |
monitor | 70 |
protector | 50 |
💪 You can implement your own G-WVD serve extremely simply and quickly. Details at gemini-web-vulnerability-detection (G-WVD)
Gemini-Self-Protector | Demo | Install - Configurate - Usage
![image](https://private-user-images.githubusercontent.com/31820707/270082193-068048ef-42cf-4032-b064-137d69abccb6.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.-Xjr8bEuqZBy6sHMb7LAkLj-P74nJfpOMOM7hpdloq0)
![image](https://private-user-images.githubusercontent.com/31820707/270082216-d8e4376f-72d1-4a7d-8a96-838b9436b0b1.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.pgoogbjqFn82SLB2GHl8ZDlqDH5gZgaJGBcAbuB1MiQ)
![image](https://private-user-images.githubusercontent.com/31820707/270082254-496033ec-e953-4ca4-9d16-73a402161f8a.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.UcAW8KFw34iLa_Y69iu-aE-81CGDkTUbcKsBCOJZHhY)
![image](https://private-user-images.githubusercontent.com/31820707/270082269-109717d9-aac2-4c97-8e36-133e2d6365cb.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmF3LmdpdGh1YnVzZXJjb250ZW50LmNvbSIsImtleSI6ImtleTUiLCJleHAiOjE3MzkzODYzMzUsIm5iZiI6MTczOTM4NjAzNSwicGF0aCI6Ii8zMTgyMDcwNy8yNzAwODIyNjktMTA5NzE3ZDktYWFjMi00Yzk3LThlMzYtMTMzZTJkNjM2NWNiLnBuZz9YLUFtei1BbGdvcml0aG09QVdTNC1ITUFDLVNIQTI1NiZYLUFtei1DcmVkZW50aWFsPUFLSUFWQ09EWUxTQTUzUFFLNFpBJTJGMjAyNTAyMTIlMkZ1cy1lYXN0LTElMkZzMyUyRmF3czRfcmVxdWVzdCZYLUFtei1EYXRlPTIwMjUwMjEyVDE4NDcxNVomWC1BbXotRXhwaXJlcz0zMDAmWC1BbXotU2lnbmF0dXJlPTRkMWZmYTJlYThhOGE2MzIyYzlhMzQ0YmMwNzljOWNlMmZiNjM3OTc0MjdlZTAyNzY1ZDliOGY2ZGM2NWFlOTYmWC1BbXotU2lnbmVkSGVhZGVycz1ob3N0In0.lAi-3FzPTm0GVjtJYmfDSehTic_GwmvJSF6dnBmJHY8)
Interested in contributing? Check out the contributing guidelines. Please note that this project is released with a Code of Conduct. By contributing to this project, you agree to abide by its terms.
gemini_self_protector
was created by lethanhphuc. It is licensed under the terms of the MIT license.
https://appseed.us/product/datta-able/flask/
Phuc Le-Thanh, Tuan Le-Anh, and Quan Le-Trung. 2023. Research and Development of a Smart Solution for Runtime Web Application Self-Protection. In Proceedings of the 12th International Symposium on Information and Communication Technology (SOICT '23). Association for Computing Machinery, New York, NY, USA, 304–311. https://doi.org/10.1145/3628797.3628901