BreachSeek, AI-Based Automated Multi-Platform Penetration Testing Tool
- С сайта: Vulnerability(cybersecuritynews.com)
- Вернуться к списку новостей
С сайта: Vulnerability(cybersecuritynews.com)
AI is significantly evolving penetration testing by enhancing automation, accuracy, and adaptability.
AI-driven tools can simulate sophisticated attack techniques, analyze vast datasets for vulnerabilities, and determine genuine threats from false positives, allowing security teams to focus on critical risks.
The following cybersecurity analysts from King Fahd University of Petroleum and Minerals (KFUPM) recently developed BreachSeek, it’s the first AI platform that provides safety and allows websites and network penetration testing to serve as breach assessment tools:-
The authors implemented a multi-agent system with the help of LLMs using LangChain and LangGraph with Python.
Such self-sufficient agents are able to search for vulnerabilities, simulate cyberattacks, and exploit them with as little help from humans as possible.
BreachSeek Penetration Testing Tool
Within the platform’s architecture, it contains various individual specialized AI agents that are hosted in separate containers which resolves the context window limitations of LLMs and guarantees extension in case of different network sizes.
Merging artificial intelligence, natural language processing, and security intelligence, BreachSeek provides an all-inclusive approach that is more effective than manual pen testing in terms of time taken, accuracy, and response to new threats.
As a result, this technique is particularly beneficial for companies dealing with information of a confidential nature such as finance, medicine, and government in which time-consuming vulnerability scanning is unacceptable.
Large language models (LLMs) are rapidly transforming the cybersecurity landscape, and this is happening more specifically in penetration testing automation.
Tools like PentestGPT make use of the LLMs to perform tasks traditionally done by human testers.
While the PentestGPT outperformed GPT-3.5 and GPT-4 on a benchmark of 182 sub-tasks aligned with OWASP’s top 10 vulnerabilities.
Besides this, other tools like Mayhem use fuzzing and symbolic execution to quickly identify vulnerabilities.
The general workflow of such models (Source – Arxiv)
BreachSeek uses several AI agents to deal with the context windows before proceeding to interface with the target environment.
These developments are quite useful in increasing the efficiency of vulnerability detection and test scenarios.
Breachseek UI
But here the main barriers still lie within the area of context over extended interactions and adapting to specific organizational needs.
However, future developments focus on improving the continuous learning aspect of the LMLs for preparedness in case of sudden changes in the situation.
The introduction of these LLMs within the cybersecurity landscape represents a significant step forward.
However, this requires continued research that is proactive and responsive to the challenges faced in order to exploit these technologies for defensive cyber warfare, which is not basic.
#Cyber_Security #Cyber_Security_News #Penetration_Testing #Vulnerability #ai_in_cybersecurity #automaated_vulnerability_detection #cyber_security_news #penetration_testing
Оригинальная версия на сайте:
BreachSeek, AI-Based Automated Multi-Platform Penetration Testing Tool
Author: Tushar Subhra DuttaAI is significantly evolving penetration testing by enhancing automation, accuracy, and adaptability.
AI-driven tools can simulate sophisticated attack techniques, analyze vast datasets for vulnerabilities, and determine genuine threats from false positives, allowing security teams to focus on critical risks.
The following cybersecurity analysts from King Fahd University of Petroleum and Minerals (KFUPM) recently developed BreachSeek, it’s the first AI platform that provides safety and allows websites and network penetration testing to serve as breach assessment tools:-
- Ibrahim AlShehri
- Adnan AlShehri
- Abdulrahman AlMalki
- Majed Bamardouf
- Alaqsa Akbar
The authors implemented a multi-agent system with the help of LLMs using LangChain and LangGraph with Python.
Such self-sufficient agents are able to search for vulnerabilities, simulate cyberattacks, and exploit them with as little help from humans as possible.
BreachSeek Penetration Testing Tool
Within the platform’s architecture, it contains various individual specialized AI agents that are hosted in separate containers which resolves the context window limitations of LLMs and guarantees extension in case of different network sizes.
Merging artificial intelligence, natural language processing, and security intelligence, BreachSeek provides an all-inclusive approach that is more effective than manual pen testing in terms of time taken, accuracy, and response to new threats.
As a result, this technique is particularly beneficial for companies dealing with information of a confidential nature such as finance, medicine, and government in which time-consuming vulnerability scanning is unacceptable.
Large language models (LLMs) are rapidly transforming the cybersecurity landscape, and this is happening more specifically in penetration testing automation.
Tools like PentestGPT make use of the LLMs to perform tasks traditionally done by human testers.
While the PentestGPT outperformed GPT-3.5 and GPT-4 on a benchmark of 182 sub-tasks aligned with OWASP’s top 10 vulnerabilities.
Besides this, other tools like Mayhem use fuzzing and symbolic execution to quickly identify vulnerabilities.
The general workflow of such models (Source – Arxiv)BreachSeek uses several AI agents to deal with the context windows before proceeding to interface with the target environment.
These developments are quite useful in increasing the efficiency of vulnerability detection and test scenarios.
Breachseek UIBut here the main barriers still lie within the area of context over extended interactions and adapting to specific organizational needs.
However, future developments focus on improving the continuous learning aspect of the LMLs for preparedness in case of sudden changes in the situation.
The introduction of these LLMs within the cybersecurity landscape represents a significant step forward.
However, this requires continued research that is proactive and responsive to the challenges faced in order to exploit these technologies for defensive cyber warfare, which is not basic.
#Cyber_Security #Cyber_Security_News #Penetration_Testing #Vulnerability #ai_in_cybersecurity #automaated_vulnerability_detection #cyber_security_news #penetration_testing
Оригинальная версия на сайте:
