When LLMs Meet Cybersecurity: A Systematic Literature Review
🔍 Explore 861+ Papers Across 11 Research Categories
📊 RQ1: Domain LLMs | 🎯 RQ2: Applications | 🤖 RQ3: Future Directions
Updates
📆[2026-08-20] We have updated the related papers up to 2026/08/20, with 105 new papers added.
📆[2026-06-15] We have updated the related papers up to 2026/06/15, with 108 new papers added.
📆[2026-02-09] We have updated the related papers up to 2026/01/31, with 80 new papers added (2025.08.31-2026.01.31).
📆[2025-11-17] We have updated the related papers up to 2025/08/31, with 176 new papers added (2025.03.01-2025.08.31).
📆[2025-03-03] We have updated the related papers up to 2025/02/28, with 33 new papers added (2025.01.01-2025.02.28).
📆[2025-01-21] We have updated the related papers up to 2024/12/31, with 74 new papers added (2024.09.01-2024.12.31).
📆[2025-01-08] We have included the publication venues for each paper.
📆[2024-09-21] We have updated the related papers up to 2024/08/31, with 75 new papers added (2024.06.01-2024.08.31).
Brief Introduction
📊 Paper Distribution by Research Question
| RQ | Category | Papers | Percentage | Trend |
|---|---|---|---|---|
| RQ1 | 📊 Evaluation Benchmarks | 93 | 10.8% | 📊 Stable |
| 🔧 Fine-tuned LLMs | 37 | 4.3% | 📊 Stable | |
| RQ2 | 🛡️ LLM Assisted Defense | 150 | 17.4% | 📈 Growing |
| 🐛 Vulnerability Detection | 120 | 13.9% | 📈 Growing | |
| ⚔️ LLM Assisted Attack | 126 | 14.6% | 🔥 Hot | |
| 🔨 Program/Vulnerability Repair | 78 | 9.1% | 📊 Stable | |
| 📦 Others | 72 | 8.4% | 📊 Stable | |
| 🎯 Threat Intelligence | 54 | 6.3% | 📊 Stable | |
| 🔍 FUZZ | 26 | 3.0% | 📊 Stable | |
| ⚠️ Insecure Code Generation | 34 | 3.9% | 🚫 Unmaintained | |
| RQ3 | 🤖 Agent4Cybersecurity | 71 | 8.2% | 🚀 Emerging |
📜 Literatures
👉 View Full Literature Review →
The complete literature review with 861+ papers has been moved to a separate file for better readability.
Quick Links by Research Question:
-
- Cybersecurity Evaluation Benchmarks (93 papers)
- Fine-tuned Domain LLMs (37 papers)
-
RQ2: LLM Applications in Cybersecurity
- Threat Intelligence (54 papers)
- FUZZ (26 papers)
- Vulnerability Detection (120 papers)
- Program/Vulnerability Repair (78 papers)
- Insecure Code Generation (34 papers)
- LLM Assisted Defense (150 papers)
- LLM Assisted Attack (126 papers)
- Others (72 papers)
-
RQ3: Future Research Directions
- Agent4Cybersecurity (71 papers)
BibTeX
@article{zhang2025llms,
title={When llms meet cybersecurity: A systematic literature review},
author={Zhang, Jie and Bu, Haoyu and Wen, Hui and Liu, Yongji and Fei, Haiqiang and Xi, Rongrong and Li, Lun and Yang, Yun and Zhu, Hongsong and Meng, Dan},
journal={Cybersecurity},
volume={8},
number={1},
pages={1--41},
year={2025},
publisher={SpringerOpen}
}