Request for Reproduction Instructions and Missing Components

Dear Professor Gervais and MAPTA authors,

I hope you are doing well. My name is [Your Name], and I am currently working on a research project related to LLM-based automated penetration testing. I recently read your paper on MAPTA, Multi-Agent Penetration Testing AI for the Web, and I found the multi-agent architecture with Coordinator, Sandbox, and Validation agents very relevant to my research.

I found the public repository at:

https://github.com/arthurgervais/mapta

However, after reviewing the repository, it seems that the current release mainly contains main.py, analyze_logs.py, and experiment logs. Some important components for reproducing the system appear to be missing or redacted, such as the sandbox factory implementation, the function_tool wrapper, complete system prompts, dependency setup, and the workflow for reproducing the real-world local white-box assessment.

I am writing to ask whether it would be possible to obtain more complete reproduction materials for academic experimentation, including any of the following if available:

1. The full runnable version of MAPTA or an artifact branch;
2. The sandbox factory implementation required by SANDBOX_FACTORY;
3. The missing function_tool.py or equivalent tool-wrapper code;
4. The complete or reproducible version of the Coordinator / Sandbox / Validation prompts;
5. Setup instructions, dependency files, or Docker configuration;
6. The workflow used for local white-box assessment of open-source GitHub projects;
7. Any scripts used to generate metrics.json files for analyze_logs.py.

My purpose is strictly academic. I plan to run experiments only on locally deployed, authorized, and isolated test environments, such as vulnerable Docker targets or self-hosted open-source web applications. I am not requesting any undisclosed vulnerability details, private target identities, or exploit information related to ongoing responsible disclosure cases.

If the full source code cannot be shared publicly, I would also greatly appreciate any partial instructions or guidance that could help reproduce the MAPTA architecture in a safe local environment.

Thank you very much for your time and for making your work available to the research community. Your paper is very helpful for my study of LLM-driven penetration testing systems.

Best regards, [Your Name] [Your University / Institution] [Your Email]