Metraipolis › archive › dc_eb90c6636e4f

A Multi-Agent Large Language Model Framework for Autonomous Web Application Penetration Testing (2026) · DOI 10.21203/rs.3.rs-11114403/v1

Source
Metraipolis archive
Original
https://doi.org/10.21203/rs.3.rs-11114403/v1
License (as archived)
CC-BY-4.0
Archived
2026-10-18T14:30:00.998Z
SHA-256 of the archived text
240b6a2c64d1d80b33195653a860eac756ac7ce13c857a127e1353b93c9afec8
Stable URL of this page
https://metraipolis.com/documento/dc_eb90c6636e4f

How to cite

A Multi-Agent Large Language Model Framework for Autonomous Web Application Penetration Testing (2026) · DOI 10.21203/rs.3.rs-11114403/v1 — Metraipolis archive, https://doi.org/10.21203/rs.3.rs-11114403/v1 (CC-BY-4.0); archived copy: https://metraipolis.com/documento/dc_eb90c6636e4f (Metraipolis archive dc_eb90c6636e4f, sha256 240b6a2c64d1d80b33195653a860eac756ac7ce13c857a127e1353b93c9afec8)
BibTeX
@misc{dceb90c6636e4f,
  title = {A Multi-Agent Large Language Model Framework for Autonomous Web Application Penetration Testing (2026) · DOI 10.21203/rs.3.rs-11114403/v1},
  howpublished = {Metraipolis archive},
  url = {https://doi.org/10.21203/rs.3.rs-11114403/v1},
  note = {License CC-BY-4.0; archived copy https://metraipolis.com/documento/dc_eb90c6636e4f; sha256 240b6a2c64d1d80b33195653a860eac756ac7ce13c857a127e1353b93c9afec8},
  year = {2026}
}

Archived text

Reproduced under CC-BY-4.0. Attribution: Metraipolis archive, https://doi.org/10.21203/rs.3.rs-11114403/v1. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Articolo recente con licenza aperta: scheda dai metadati Crossref (posted-content).
Titolo: A Multi-Agent Large Language Model Framework for Autonomous Web Application Penetration Testing
Autori: Tapasvi Madhak, Vidhi Desai, Shubham Bholwankar, Jeet Moradiya, Anand Singh Rajput, Mehulkumar Dalwadi
Pubblicato: 2026-10-08 · editore: Springer Science and Business Media LLC
DOI: https://doi.org/10.21203/rs.3.rs-11114403/v1
Licenza dell'articolo: CC-BY-4.0

Abstract (testo degli autori, CC-BY-4.0):
Abstract As cyberattacks targeting web applications are becoming progressively more sophisticated, traditional penetration testing finds it difficult to keep pace with the evolving and adaptive tactics of modern attackers. Some manual tools are effective for detecting standard vulnerabilities, but they lack semantic reasoning or logical flaws, limiting their ability to discover complex exploit chains. While recent AI-based approaches and Large Language Models (LLMs) show potential for automating vulnerability analysis, most of them still lack integration with practical, real-world testing environments. Furthermore, reinforcement learning has been examined for attack simulation planning, but its integration within operational red-teaming applications is still immature.This study presents ARES (Automated Red-Teaming Evaluation System) --- an AI-driven, multi-agent penetration testing framework built for autonomous, ethical and adaptive vulnerability assessment by integrating the use of Large Language Models (LLMs) and reinforcement learning. ARES is built uponthe open-source Hexstrike-AI, along with two key modules: ARES v1, a domain-specific fine-tuned LLM trained with OWASP and ExploitDB datasets for thorough understanding of exploit reasoning; and ARES-RL, a Proximal Policy Optimization (PPO) based module selection and attack sequencing. Tested on controlled web environments such as DVWA, OWASP Juice Shop, and bWAPP, ARES demonstrated that it could achievehigher vulnerability findings and efficiency than traditional scanners, while enhancing its reasoningand exploit accuracy over practice iterations. ARES represents a novel convergence of LLM-based reasoning, reinforcement learning, and modular automation, advancing the field toward safe, ethicalred-teaming frameworks for web application security.

Fonte: https://api.crossref.org/works/10.21203%2Frs.3.rs-11114403%2Fv1 — metadati Crossref, consultati il 2026-10-18. Interrogazione: «multi-agent systems large language models».