A Multi-Agent System for Cybersecurity Threat Detection and Correlation Using Large Language Models
Open accessThe DOI matches its Crossref record; 45 of the 45 references it cites were found in a registry or an index.
Abstract
As cyber-attacks rapidly evolve across communication, infrastructure and data layers, traditional security solutions such as rule-based intrusion detection systems (IDS) or signature-based antivirus programs are effective at detecting known threats, but they often lack the contextual understanding and semantic interpretation necessary to detect complex or evolving attacks. For example, spear-phishing campaigns, advanced persistent threats (APTs), and multi-stage attacks often escape detection due to their subtle and context-dependent nature. This limitation creates a critical gap in detecting coordinated or subtle attack patterns that span multiple systems and domains. The need for semantic understanding, cross-domain visibility, and adaptive detection is increasingly urgent, particularly as threat actors employ polymorphic and AI-driven strategies that traditional systems cannot interpret or correlate effectively. This paper presents a modular multi-agent architecture that integrates established cybersecurity analysis tools with large language models (LLMs) to achieve intelligent, explicable and highly accurate detection of threats across diverse data types. Three specialized agents: (1) email verification, (2) log analysis, and (3) IP address scanning each operate independently with tailored detection pipelines that combine domain-specific tools and LLM-powered semantic analysis components to identify, characterize, and report threats specific to their domain. At the core o…
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“The second is the development of agentic AI: software systems that wrap LLM reasoning in perception, planning, memory, tool-use, and multi-agent communication loops, enabling them to autonomously execute multi-step security tasks that previously required human analysts”
From Agentic AI and Large Language Models for Autonomous IoT Cybersecurity: A Systematic Survey, Taxonomy, and Research Roadmap (2026) · cites this paper as [23]
In the citing paper
“cooperative threat-detection agents reach over 90% classification accuracy on multi-attack IoT benchmarks [23,92]”
From Agentic AI and Large Language Models for Autonomous IoT Cybersecurity: A Systematic Survey, Taxonomy, and Research Roadmap (2026) · cites this paper as [23]
In the citing paper
“The collaborative IoT intrusion detection framework of Liu et al. [23] reports over 90% classification accuracy across three benchmark datasets, attributable to specialization gains.”
From Agentic AI and Large Language Models for Autonomous IoT Cybersecurity: A Systematic Survey, Taxonomy, and Research Roadmap (2026) · cites this paper as [23]
Citations
Cites
A survey on large language model (LLM) security and privacy: The Good, The Bad, and The Ugly
Yao, Y., Duan, J., Xu, K., Cai, Y., Sun, Z., & Zhang, Y. (2024) · High-Confidence Computing · doi:10.1016/j.hcc.2024.100211 · Cited by 1,083 · cited 3×
“In addition, our system stands out for its modular design, structured workflow and transparent semantic layer, which enable consistent interpretability across logs, emails and network analysis [2], [25], [31].”
When LLMs meet cybersecurity: a systematic literature review
Zhang, J., Bu, H., Wen, H., Liu, Y., Fei, H., Xi, R., Li, L., Yang, Y., Zhu, H., & Meng, D. (2025) · Cybersecurity · doi:10.1186/s42400-025-00361-w · Cited by 383 · cited 4×
“However, many of these frameworks still suffer from a lack of deep semantic reasoning or natural language understanding, making them hard to interpret and limited in terms of adaptability when dealing with polymorphic or AI-driven threats [25], [27].”
InfluentialLog-based anomaly detection with deep learning: how far are we?
Le, V.-H., & Zhang, H. (2022) · Proceedings of the 44th International Conference on Software Engineering · doi:10.1145/3510003.3510155 · Cited by 257 · cited 1×
“Le and Zhang [53] provided a comprehensive review of deep learning techniques applied to system logs, highlighting both their effectiveness and limitations in real-world environments.”
A survey of large language models for cyber threat detection
Chen, Y., Cui, M., Wang, D., Cao, Y., Yang, P., Jiang, B., Lü, Z., & Liu, B. (2024) · Computers & Security · doi:10.1016/j.cose.2024.104016 · Cited by 227 · cited 5×
“…among actual positives), F1 score (harmonic mean of accuracy and recall), false positive rate (FPR), correlation accuracy (accuracy of the link between inter-agent events), and reduction in analyst response time (time saved thanks to the explanation and synthesis provided by LLMs) [9], [18], [35].”
InfluentialDeep learning for anomaly detection in log data: A survey
Landauer, M., Onder, S., Skopik, F., & Wurzenberger, M. (2023) · Machine Learning with Applications · doi:10.1016/j.mlwa.2023.100470 · Cited by 202 · cited 3×
“Performance was measured using standard metrics from artificial intelligence and cybersecurity, as shown in Table 4 [8], [17].”
Prompt Engineering or Fine-Tuning? A Case Study on Phishing Detection with Large Language Models
Trad, F. S., & Chehab, A. (2024) · Machine Learning and Knowledge Extraction · doi:10.3390/make6010018 · Cited by 174 · cited 2×
“Altwaijry et al. [19] and Atawneh and Aljehani [20] benchmarked deep learning-based detectors, while Trad and Chehab [23] explored LLM prompt engineering versus fine-tuning for this task.”
Enhancing Autonomous System Security and Resilience With Generative AI: A Comprehensive Survey
Lopez, M. A., Lunardi, W. T., Lawton, G., & Thakkar, S. (2024) · IEEE Access · doi:10.1109/access.2024.3439363 · Cited by 171 · cited 1×
“Such systems decentralize detection logic into domain-specific intelligent agents, each one responsible for analyzing a particular threat layer [43].”
Advancing Phishing Email Detection: A Comparative Study of Deep Learning Models
Altwaijry, Ν., Al-Turaiki, I. M., Alotaibi, R., & Alakeel, F. (2024) · Sensors · doi:10.3390/s24072077 · Cited by 135 · cited 3×
“Altwaijry et al. [19] and Atawneh and Aljehani [20] benchmarked deep learning-based detectors, while Trad and Chehab [23] explored LLM prompt engineering versus fine-tuning for this task.”
Evaluation of LLM-based chatbots for OSINT-based Cyber Threat Awareness
Shafee, S., Bessani, A., & Ferreira, P. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.125509 · Cited by 112
Alhuzali, A., Alloqmani, A., Aljabri, M., & Alharbi, F. (2025) · Applied Sciences · doi:10.3390/app15063396 · Cited by 97 · cited 1×
“Alhuzali et al. [52] conducted an exhaustive comparative analysis of fourteen machine learning and deep learning models, including BERT and LSTM, on ten publicly available datasets.”
Research works printed but not linked: 10
- Ruzickova, AI and LLM models to analyze and identify cybersecurity incidents, 1
- 2006, Public Corpus
- Sharafaldin, Toward generating a new intrusion detection dataset and intrusion traffic characterization, 108
- 2023, National Vulnerability Database (NVD)
- Marantos, 2024, Leveraging LLMs for dynamic cyber-threat detection and training, IEEE Big Data
- Patil, 2024, ForenSift: Gen-AI powered integrated digital forensics and incident response platform using LangChain framework, Digit. Forensics Secur. Appl., 2024, 1
- Khan, 2025, LLMs for Malware detection: Review, framework design, and countermeasure approaches, SSRN
- Li, KnowPhish: LLMs meet multimodal knowledge graphs for phishing detection, 1
- Balasubramanian, CYGENT: A conversational GPT-based agent for log anomaly detection and cyber incident explanation, 1587
- ZeMicheal, LLM agents for vulnerability identification and CVE verification, 3562
Cited by
Mamodiya, U., Kishor, I., Naz, R., Almaiah, M., & Alqutaish, A. (2026) · Journal of Cybersecurity and Privacy · doi:10.3390/jcp6030085 · Cited by 9
“Certain models seek to automate mitigation based on security orchestration applications and policy engines [36].”
Alqahtani, H., & Kumar, G. (2026) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2026.077367 · Cited by 14
Nageshwaran, V., & Ezekiel, S. (2026) · Electronics · doi:10.3390/electronics15122740 · Cited by 4
“The collaborative IoT intrusion detection framework of Liu et al. [23] reports over 90% classification accuracy across three benchmark datasets, attributable to specialization gains.”
Nageshwaran, V., & Ezekiel, S. (2026) · doi:10.20944/preprints202606.0142.v1 · Cited by 4
Lan, K., Li, G. Y., Huang, W., Li, J., & Yu, Y. (2026) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2026.3687554 · Cited by 4
“HMIMOU et al. [53] proposed a Multi-Agent System for Cyber security Threat Detection and Correlation Using LLM, this method requires multiple agents and is still in the model exploration phase, where its maturity is not yet high.”
Metadata
- Authors
- Hmimou, Y., Tabaa, M., Khiat, A., & Hidila, Z.
- Affiliations
- École Marocaine des Sciences de l'Ingénieur (MA)
- University of Hassan II Casablanca (MA)
- Venue
- IEEE Access
- Published
- 2025
- Volume
- 13
- Pages
- 150199-150215
- ISSN
- 2169-3536 (ISSN-L)
- OpenAlex
- W4413556821
- Cited by
- 25
- References
- 44
- Citations by year
- 2026: 25
- FWCI
- 10.49
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- IEEE Access (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › Network Security and Intrusion Detection
- Physical Sciences › Computer Science › Hardware and Architecture › Network Packet Processing and Optimization
- Physical Sciences › Computer Science › Information Systems › Information and Cyber Security
- Keywords
- multi-agent systems, large language models, cyber threat detection, semantic analysis, email verification, log analysis, advanced persistent threats, spear phishing, multi-stage attacks, stealthy attacks, false positive reduction, explainable security
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.