My name is Liu Tong, a second-year PhD student in Institute of Information Engineering, Chinese Academy of Sciences, advised by Prof. Kai Chen and Prof. Guozhu Meng. I am also a core player of the well-known CTF teams Nu1L and Straw Hat.
My current research interests mainly focus on AI security and software security, with particular attention to their intersection and exploring security issues within the AI software ecosystem. My work has been published in top-tier academic and industry conferences and journals such as CCS, USENIX Security, TOSEM, and BlackHat.
Besides research, I am an independent bug hunter who has discovered hundreds of vulnerabilities in both open-source and closed-source software, earning acknowledgments from leading tech companies including Apple, Google, Microsoft, Baidu, and Huawei.
🔥 News
- 2025.06: 🎉🎉 I’ll present my research as an invited speaker in ISSTA LLMSC workshop.
- 2024.07: 🎉🎉 One paper accepted by CCS 2024.
- 2024.05: 🎉🎉 One paper accepted by USENIX Security 2024.
- 2024.01: 🎉🎉 Our talk got accepted by BlackHat Asia 2024.
📝 Publications
-
Demystifying RCE Vulnerabilities in LLM-Integrated Apps [PDF] [Code]
Tong Liu, Zizhuang Deng, Guozhu Meng, Yuekang Li, Kai Chen.
ACM Conference on Computer and Communications Security (CCS), 2024 (CCF-A) -
Making Them Ask and Answer: Jailbreaking Large Language Models in Few Queries via Disguise and Reconstruction [PDF] [Code] [Video]
Tong Liu, Yingjie Zhang, Zhe Zhao, Yinpeng Dong, Guozhu Meng, Kai Chen.
USENIX Security Symposium (Security), 2024 (CCF-A) -
LLM4Shell: Discovering and Exploiting RCE Vulnerabilities in Real-World LLM-Integrated Frameworks and Apps [Slides] [Video]
Tong Liu, Yuekang Li, Zizhuang Deng, Guozhu Meng, Kai Chen.
BlackHat Asia, 2024 -
Differential Testing of Cross Deep Learning Framework APIs: Revealing Inconsistencies and Vulnerabilities [PDF] [Slides] [Video]
Zizhuang Deng, Guozhu Meng, Kai Chen, Tong Liu, Lu Xiang, Chunyang Chen.
USENIX Security Symposium (Security), 2023 (CCF-A) -
Attack as Detection: Using Adversarial Attack Methods to Detect Abnormal Examples [PDF]
Zhe Zhao, Guangke Chen, Tong Liu, Taishan Li, Fu Song, Jingyi Wang, Jun Sun.
ACM Transactions on Software Engineering and Methodology (TOSEM) (CCF-A)
🎖 Honors and Awards
🛡️ Bug Bounty & Security Research
- Bug Bounty and Acknowledgements from leading tech companies including Apple, Microsoft, Google, Huawei, and Baidu etc.
- Discovered and reported 40+ CVEs in AI infrastructures (Deep learning libraries, LLM-integrated frameworks, etc.)
🏆 Selected CTF & AI Competitions
| Achievement | Team |
|---|---|
| 4th Place, DEFCON CTF Qual 2024 | Straw Hat |
| 5th Place, DEFCON CTF Final 2024 | Straw Hat |
| 1st Place, ByteAI 2024 | suibianwanwan |
| 5th Place, DEFCON CTF Final 2023 | Straw Hat |
| 6th Place, DEFCON CTF Qual 2023 | Straw Hat |
| 1st Place, D3CTF 2022 | Nu1L |
| 3rd Place, 0CTF 2022 | Straw Hat |
| 7th Place, DEFCON CTF Final 2022 | Straw Hat |
| 8th Place, DEFCON CTF Qual 2022 | Straw Hat |
| 1st Place, CVPR 2022 Workshop-Track II Phase I | ShanghaiTech-S3L |
| 5th Place, CVPR 2022 Workshop-Track II Phase II | ShanghaiTech-S3L |
| 7th Place, 0CTF 2021 Final | Nu1L |
| 7th Place, DEFCON CTF 2021 Final | Nu1L |
📖 Educations
- 2023.09 - now, Ph.D, Institute of Information Engineering, Chinese Academy of Sciences; School of Cyber Security, University of Chinese Academy of Sciences.
- 2019.09 - 2023.06, ShanghaiTech University, SIST.
💬 Invited Talks
- 2025.06, Invited speaker, ISSTA LLMSC workshop 2025.
- 2025.03, LLM Security, Aliyun WhiteHat Conference 2025.
- 2024.08, New Security Exploitations in the Era of LLMs, CTFCON 2024.
- 2024.04, LLM4Shell, BlackHat Asia 2024.