Show HN: I spent 8 months trying to make LLMs Hack

I'm excited to share with you my latest project, Peneterrer, a custom-built AI-powered penetration testing tool that uses Large Language Models (LLMs) to identify vulnerabilities in web applications.

For the past 8 months, I've been working tirelessly to create an AI security tester that can simulate real-world attacks on web applications. The system consists of 29 agents, each with its own unique capabilities and strengths. These agents work together seamlessly to identify potential vulnerabilities and provide a comprehensive report to users.

The core of Peneterrer is a custom LLM orchestration framework, designed using the task-subtask architecture. This approach allows the agents to break down complex tasks into smaller, manageable chunks, making it easier for them to complete their objectives. The framework also includes a custom agent calling mechanism, which enables the agents to communicate effectively and work together as a team.

One of the key features of Peneterrer is its ability to learn from user feedback. After each test, users receive a full Vulnerability Management portal and a comprehensive Pentest report, sometimes exceeding 30 pages in length. This detailed report provides users with a clear understanding of the identified vulnerabilities and recommended remediation steps.

Each test runs in an isolated Kali Linux environment on AWS Fargate, giving the agents full access to the environment to conduct their simulated attacks. The agents also have unrestricted access to the internet through Tavily, allowing them to search for and research relevant content as needed during the testing process.

What's next for Peneterrer?

I'm proud of what I've accomplished so far, but I know that there's still much work to be done. My next steps include refining the framework, adding more agents, and exploring ways to integrate Peneterrer with existing security tools.

As a community-driven project, I encourage anyone interested in contributing to Peneterrer or providing feedback on its performance to reach out to me directly. Your input will help shape the future of this innovative tool and ensure that it continues to provide the best possible experience for users.

Test Peneterrer today!

You can try out Peneterrer for yourself by visiting our website: https://peneterrer.com/

Feedback appreciated!

I'm grateful for the opportunity to share my passion project with you and look forward to seeing what the future holds for Peneterrer.