Deep learning-based vulnerability detection tool for C/C++ smart contracts at function-body slice level

doi:10.11772/j.issn.1001-9081.2024121858

Journal of Computer Applications ›› 2025, Vol. 45 ›› Issue (11): 3493-3501.DOI: 10.11772/j.issn.1001-9081.2024121858

• The 7th CCF China Conference on Blockchain Technology • Previous Articles

Deep learning-based vulnerability detection tool for C/C++ smart contracts at function-body slice level

Yushu LI¹^,², Ying XING¹^,², Siqi LU³, Heng PAN¹^,²(), Senchun CHAI⁴, Xueming SI¹^,²

^1.School of Cyberspace Security，Zhongyuan University of Technology，Zhengzhou Henan 450007，China
^2.Henan International Joint Laboratory of Blockchain and Data Sharing （Zhongyuan University of Technology），Zhengzhou Henan 450007，China
^3.Henan Key Laboratory of Network Cryptography Technology （Information Engineering University of PLA Strategic Support Force），Zhengzhou Henan 450001，China
^4.School of Automation，Beijing Institute of Technology，Beijing 100081，China

Received:2025-01-02 Revised:2025-03-06 Accepted:2025-03-13 Online:2025-04-18 Published:2025-11-10
Contact: Heng PAN
About author:LI Yushu， born in 1998， M. S. candidate. His research interests include cybersecurity， blockchain.
XING Ying， born in 1985， Ph. D.， lecturer. Her research interests include malicious code detection， smart contract vulnerability detection.
LU Siqi， born in 1990， Ph. D.， lecturer. His research interests include smart contract security， cryptographic protocol analysis and design， formal methods.
CHAI Senchun， born in 1978， Ph. D.， professor. His research interests include intelligent information processing and control.
SI Xueming， born in 1966， Ph. D.， professor. His research interests include blockchain， cryptography.
Supported by:
Key Scientific Research Project of Higher Education Institutions in Henan Province(24A520059);Henan Province Science and Technology Research Project in 2024(242102210136);Discipline Strength Enhancement Program of Zhongyuan Institute of Technology(GG202425);Graduate Research Innovation Program of Zhongyuan University of Technology(YKY2024ZK27)

基于深度学习的函数体切片级C/C++智能合约漏洞检测工具

李浴淑¹^,², 邢颖¹^,², 陆思奇³, 潘恒¹^,²(), 柴森春⁴, 斯雪明¹^,²

^1.中原工学院网络空间安全学院，郑州 450007
^2.河南省区块链与数据共享国际联合实验室（中原工学院），郑州 450007
^3.河南省网络密码技术重点实验室（中国人民解放军战略支援部队信息工程大学），郑州 450001
^4.北京理工大学自动化学院，北京 100081

通讯作者: 潘恒
作者简介:李浴淑（1998—），男，河南开封人，硕士研究生，主要研究方向：网络安全、区块链
邢颖（1985—），女，河南沈丘人，讲师，博士，主要研究方向：恶意代码检测、智能合约漏洞检测
陆思奇（1990—），男，山东济宁人，讲师，博士，主要研究方向：智能合约安全、密码协议分析与设计、形式化方法
柴森春（1978—），北京人，教授，博士，主要研究方向：智能信息处理与控制
斯雪明（1966—），男，浙江诸暨人，教授，博士，主要研究方向：区块链、密码学。
基金资助:
河南省高等学校重点科研项目(24A520059);2024年度河南省科技攻关计划项目(242102210136);中原工学院优势学科实力提升计划项目(GG202425);中原工学院研究生科研创新计划项目(YKY2024ZK27)

Abstract

Abstract:

With the frequent occurrence of security incidents caused by smart contract vulnerabilities， existing detection tools lack sufficient support for multiple programming languages， particularly in terms of their inability to detect vulnerabilities at the source code level in C/C++ smart contracts. To address this issue， a deep learning-based vulnerability detection method for C/C++ smart contracts was proposed and a function-body slicing-level detection tool CDFSentry was designed. Starting from the perspective of source code， the concept of target regions in deep learning applications within the field of image processing was applied to smart contract vulnerability detection. The implementation of the tool involved four steps： first， extracting function-body slices related to vulnerabilities to obtain complete function-body information； second， annotating the extracted slices； third， encoding the slices into vectors to convert them into input formats suitable for deep learning； four， completing vector labeling and model training. Besides， by analyzing the causes of vulnerabilities in C/C++ smart contracts， five types of vulnerabilities were defined： integer overflow， permission control， token transfer， memory management， and transaction delay， and a dataset containing 5 024 source codes was constructed to solve the problems of insufficient open-source datasets and inconsistent definitions of vulnerability types in this field. Experimental results on this dataset demonstrate that while the comparable deep learning tool GNNSCVulDetector can only detect one type of vulnerability， CDFSentry detects five types with 12.68 percentage points higher accuracy. By leveraging deep learning to detect vulnerabilities in C/C++ smart contract source code， CDFSentry reduces reliance on experts while offering higher detection accuracy and broader coverage than similar tools. In addition， through continuous learning and training， its detection ability can be improved continuously.

Key words: smart contract, deep learning, function-body, vulnerability detection, integer overflow, transaction delay

摘要：

智能合约漏洞引发的安全事件频发，而现有检测工具对多语言支持不足，特别是缺乏对C/C++智能合约源码级别的漏洞检测能力。针对上述问题，提出基于深度学习的C/C++智能合约漏洞检测方法，并设计了函数体切片级检测工具CDFSentry。该工具从源码角度出发，借鉴深度学习在图像处理领域中的目标区域概念应用于智能合约漏洞检测。工具实现分为4步：一是提取漏洞函数体切片，获取完整函数体信息；二是对提取的切片进行标注；三是将切片编码为向量，转化为深度学习输入格式；四是完成向量标记与模型训练。此外，通过分析C/C++智能合约漏洞产生的原因，定义了5种类型的漏洞：整数溢出、权限控制、代币转移、内存管理和交易延迟，并构建了包含5 024个源码的数据集，解决了该领域开源数据集匮乏与漏洞类型不统一的问题。在该数据集上的实验结果表明，同类型的深度学习工具GNNSCVulDetector仅能检测1种漏洞，而CDFSentry可检测5种漏洞，且准确率提高了12.68个百分点。CDFSentry基于深度学习检测C/C++智能合约源码漏洞，可减少对专家依赖，且相比同类工具检测精度更高、范围更广；通过持续学习训练，它的检测能力可不断提升。

关键词: 智能合约, 深度学习, 函数体, 漏洞检测, 整数溢出, 交易延迟

CLC Number:

TP312

Yushu LI, Ying XING, Siqi LU, Heng PAN, Senchun CHAI, Xueming SI. Deep learning-based vulnerability detection tool for C/C++ smart contracts at function-body slice level[J]. Journal of Computer Applications, 2025, 45(11): 3493-3501.

李浴淑, 邢颖, 陆思奇, 潘恒, 柴森春, 斯雪明. 基于深度学习的函数体切片级C/C++智能合约漏洞检测工具[J]. 《计算机应用》唯一官方网站, 2025, 45(11): 3493-3501.

Figures/Tables 12

References 29

[1]	BitShares. Delegated proof-of-stake consensus［EB/OL］. ［2024-06-07］. .
[2]	QUANTUM M. The proposal of PoS［EB/OL］. ［2024-07-10］..
[3]	BUTERIN V. A next-generation smart contract and decentralized application platform［EB/OL］. ［2024-10-01］..
[4]	DE CANDIA A. Increase of EOSIO transaction volumes［EB/OL］. ［2024-09-03］..
[5]	MOZILLA. Basic concepts for WASM［EB/OL］. ［2024-07-01］..
[6]	NONEAGE. Random number generation vulnerability［EB/OL］. ［2024-10-12］. .
[7]	Fake EOS transfer vulnerabilities in EOS smart contracts［EB/OL］. ［2024-11-02］. .
[8]	Forged transfer notification in EOS smart contracts［EB/OL］. ［2024-11-05］..
[9]	HE N， ZHANG R， WANG H， et al. EOSAFE： security analysis of EOSIO smart contracts［C］// Proceedings of the 30th USENIX Security Symposium. Berkeley： USENIX Association， 2021： 1271-1288.
[10]	QUAN L， WU L， WANG H. EVulHunter： detecting fake transfer vulnerabilities for EOSIO's smart contracts at WebAssembly level［EB/OL］. ［2024-10-26］..
[11]	LI W， WANG M， YU B， et al. Grey-box fuzzing based on execution feedback for EOSIO smart contracts［C］// Proceedings of the 29th Asia-Pacific Software Engineering Conference. Piscataway： IEEE， 2022： 1-10.
[12]	HUANG Y， JIANG B， CHAN W K. EOSFuzzer： fuzzing EOSIO smart contracts for vulnerability detection［C］// Proceedings of the 12th Asia-Pacific Symposium on Internetware. New York： ACM， 2020： 99-109.
[13]	JIN L， CAO Y， CHEN Y， et al. ExGen： cross-platform， automated exploit generation for smart contract vulnerabilities［J］. IEEE Transactions on Dependable and Secure Computing， 2020： 650-664.
[14]	ZHUANG Y， LIU Z， QIAN P， et al. Smart contract vulnerability detection using graph neural networks［C］// Proceedings of the 29th International Joint Conference on Artificial Intelligence. California： ijcai.org， 2020： 3283-3290.
[15]	KRUPP J， ROSSOW C. teEther： gnawing at Ethereum to automatically exploit smart contracts［C］// Proceedings of the 27th USENIX Security Symposium. Berkeley： USENIX Association， 2018： 1317-1333.
[16]	CHEN J， XIA X， LO D， et al. DefectChecker： automated smart contract defect detection by analyzing EVM bytecode［J］. IEEE Transactions on Software Engineering， 2022， 48（7）： 2189-2207.
[17]	JIANG B， LIU Y， CHAN W K. ContractFuzzer： fuzzing smart contracts for vulnerability detection［C］// Proceedings of the 33rd ACM/IEEE International Conference on Automated Software Engineering. New York： ACM， 2018： 259-269.
[18]	CHOI J， KIM D， KIM S， et al. SMARTIAN： enhancing smart contract fuzzing with static and dynamic data-flow analyses［C］// Proceedings of the 36th IEEE/ACM International Conference on Automated Software Engineering. Piscataway： IEEE， 2021： 227-239.
[19]	TANN W J W， HAN X J， GUPTA S S， et al. Towards safer smart contracts： a sequence learning approach to detecting security threats［EB/OL］. ［2024-11-05］..
[20]	ZHANG L， CHEN W， WANG W， et al. CBGRU： a detection method of smart contract vulnerability based on a hybrid model［J］. Sensors， 2022， 22（9）： No.3577.
[21]	LI Z， LU S， ZHANG R， et al. VulHunter： hunting vulnerable smart contracts at EVM bytecode-level via multiple instance learning［J］. IEEE Transactions on Software Engineering， 2023， 49（11）： 4886-4916.
[22]	LI Z， ZOU D， XU S， et al. SySeVR： a framework for using deep learning to detect software vulnerabilities［J］. IEEE Transactions on Dependable and Secure Computing， 2022， 19（4）： 2244-2258.
[23]	LI Z， ZOU D， XU S， et al. VulDeePecker： a deep learning-based system for vulnerability detection［EB/OL］. ［2024-11-05］..
[24]	钱鹏，刘振广，何钦铭，等.智能合约安全漏洞检测技术研究综述［J］.软件学报，2022，33（8）：3059-3085.
	QIAN P， LIU Z G， HE Q M， et al. Smart contract vulnerability detection technique： a survey［J］. Journal of Software， 2022， 33（8）： 3059-3085.
[25]	REN S， HE K， GIRSHICK R， et al. Faster R-CNN： towards real-time object detection with region proposal networks［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2017， 39（6）： 1137-1149.
[26]	SHRIVASTAVA A， GUPTA A， GIRSHICK R， et al. Training region-based object detectors with online hard example mining［C］// Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2016： 761-769.
[27]	HOCHREITER S， SCHMIDHUBER J. Long short-term memory［J］. Neural Computation， 1997， 9（8）： 1735-1780.
[28]	CHO K， VAN MERRIËNBOER B， BAHDANAU D， et al. On the properties of neural machine translation： encoder-decoder approaches［C］// Proceedings of the 8th Workshop on Syntax， Semantics and Structure in Statistical Translation. Stroudsburg： ACL， 2014： 103-111.
[29]	GRAVES A， SCHMIDHUBER J. Framewise phoneme classification with bidirectional LSTM and other neural network architectures［J］. Neural Networks， 2005， 18（5/6）： 602-610.

模型	漏洞种类	准确率（A）	精度（P）
LSTM	整数溢出	85.10	84.10
	权限控制	86.61	86.43
	代币转移	86.52	86.56
	内存管理	86.76	86.77
	交易延迟	86.29	86.30
BiLSTM	整数溢出	85.30	84.30
	权限控制	86.71	86.53
	代币转移	86.62	86.66
	内存管理	86.86	86.87
	交易延迟	86.39	86.40
GRU	整数溢出	84.80	83.80
	权限控制	86.21	86.03
	代币转移	86.12	86.03
	内存管理	86.36	86.37
	交易延迟	85.89	85.90
BGRU	整数溢出	85.30	84.30
	权限控制	86.71	86.53
	代币转移	86.62	86.66
	内存管理	86.86	86.87
	交易延迟	86.39	86.40

模型	漏洞种类	准确率（A）	精度（P）
LSTM	整数溢出	85.10	84.10
	权限控制	86.61	86.43
	代币转移	86.52	86.56
	内存管理	86.76	86.77
	交易延迟	86.29	86.30
BiLSTM	整数溢出	85.30	84.30
	权限控制	86.71	86.53
	代币转移	86.62	86.66
	内存管理	86.86	86.87
	交易延迟	86.39	86.40
GRU	整数溢出	84.80	83.80
	权限控制	86.21	86.03
	代币转移	86.12	86.03
	内存管理	86.36	86.37
	交易延迟	85.89	85.90
BGRU	整数溢出	85.30	84.30
	权限控制	86.71	86.53
	代币转移	86.62	86.66
	内存管理	86.86	86.87
	交易延迟	86.39	86.40

漏洞类型	源码数	函数体数		训练数据		测试数据		TP	FP	FN	TN	A/%
漏洞类型	源码数	Bad	Good	Bad	Good	Bad	Good	TP	FP	FN	TN	A/%
整数溢出	1 012	712	856	571	684	141	172	145	27	19	122	85.30
权限控制	1 002	651	782	521	626	130	156	135	21	17	113	86.71
代币转移	1 000	899	748	719	599	180	149	156	24	20	129	86.62
内存管理	1 000	802	669	642	532	160	137	139	21	18	119	86.86
交易延迟	1 010	751	626	604	501	147	125	108	17	20	127	87.29

漏洞类型	源码数	函数体数		训练数据		测试数据		TP	FP	FN	TN	A/%
漏洞类型	源码数	Bad	Good	Bad	Good	Bad	Good	TP	FP	FN	TN	A/%
整数溢出	1 012	712	856	571	684	141	172	145	27	19	122	85.30
权限控制	1 002	651	782	521	626	130	156	135	21	17	113	86.71
代币转移	1 000	899	748	719	599	180	149	156	24	20	129	86.62
内存管理	1 000	802	669	642	532	160	137	139	21	18	119	86.86
交易延迟	1 010	751	626	604	501	147	125	108	17	20	127	87.29

工具	漏洞类型	输入格式	检测数	漏洞数	FP	FN
EOSFuzzer	块信息依赖	源码	82	2	0	1
	伪造转让通知		82	4	0	0
	虚假EOS转账		82	2	1	0
GFuzzer	块信息依赖		82	3	0	0
	伪造转让通知		82	5	0	0
	虚假EOS转账		82	3	1	0

Deep learning-based vulnerability detection tool for C/C++ smart contracts at function-body slice level

基于深度学习的函数体切片级C/C++智能合约漏洞检测工具

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 12

References 29

Related Articles 15

Recommended Articles

Metrics

工具	范围
GNNSCVulDetector	无限循环
EOSAFE	虚假令牌、虚假接收、回滚、权限检查
EvulHunter	虚假令牌、虚假接收
EOSFuzzer	块信息依赖、虚假令牌、虚假接收
GFuuzer	块信息依赖、虚假令牌、虚假接收
ExGen	整数溢出
CDFSentry	整数溢出、权限控制、代币转移、内存管理、交易延迟

漏洞类型	输入格式	检测数	TP	FP	TN	P/%	R/%	A/%
虚假转账	字节码	184	75	26	83	74.26	100	85.87
虚假通知		195	141	0	54	100.00	100	100.00
混合		379	216	26	137	89.26	100	93.14

[1]	Hongjun ZHANG, Gaojun PAN, Hao YE, Yubin LU, Yiheng MIAO. Multi-source heterogeneous data analysis method combining deep learning and tensor decomposition [J]. Journal of Computer Applications, 2025, 45(9): 2838-2847.
[2]	Jin LI, Liqun LIU. SAR and visible image fusion based on residual Swin Transformer [J]. Journal of Computer Applications, 2025, 45(9): 2949-2956.
[3]	Bing YIN, Zhenhua LING, Yin LIN, Changfeng XI, Ying LIU. Emotion recognition method compatible with missing modal reasoning [J]. Journal of Computer Applications, 2025, 45(9): 2764-2772.
[4]	Wei SHE, Tianxiang MA, Haige FENG, Zhao TIAN, Wei LIU. Blockchain covert communication method based on contract call concealment [J]. Journal of Computer Applications, 2025, 45(9): 2865-2872.
[5]	Weigang LI, Jiale SHAO, Zhiqiang TIAN. Point cloud classification and segmentation network based on dual attention mechanism and multi-scale fusion [J]. Journal of Computer Applications, 2025, 45(9): 3003-3010.
[6]	Zhixiong XU, Bo LI, Xiaoyong BIAN, Qiren HU. Adversarial sample embedded attention U-Net for 3D medical image segmentation [J]. Journal of Computer Applications, 2025, 45(9): 3011-3016.
[7]	Panfeng JING, Yudong LIANG, Chaowei LI, Junru GUO, Jinyu GUO. Semi-supervised image dehazing algorithm based on teacher-student learning [J]. Journal of Computer Applications, 2025, 45(9): 2975-2983.
[8]	Haiyang PENG, Weixing JI, Fawang LIU. Blockchain-based data notarization model for autonomous driving simulation testing [J]. Journal of Computer Applications, 2025, 45(8): 2421-2427.
[9]	Peng PENG, Ziting CAI, Wenling LIU, Caihua CHEN, Wei ZENG, Baolai HUANG. Speech emotion recognition method based on hybrid Siamese network with CNN and bidirectional GRU [J]. Journal of Computer Applications, 2025, 45(8): 2515-2521.
[10]	Shuo ZHANG, Guokai SUN, Yuan ZHUANG, Xiaoyu FENG, Jingzhi WANG. Dynamic detection method of eclipse attacks for blockchain node analysis [J]. Journal of Computer Applications, 2025, 45(8): 2428-2436.
[11]	Lina GE, Mingyu WANG, Lei TIAN. Review of research on efficiency of federated learning [J]. Journal of Computer Applications, 2025, 45(8): 2387-2398.
[12]	Yanhua LIAO, Yuanxia YAN, Wenlin PAN. Multi-target detection algorithm for traffic intersection images based on YOLOv9 [J]. Journal of Computer Applications, 2025, 45(8): 2555-2565.
[13]	Jinxian SUO, Liping ZHANG, Sheng YAN, Dongqi WANG, Yawen ZHANG. Review of interpretable deep knowledge tracing methods [J]. Journal of Computer Applications, 2025, 45(7): 2043-2055.
[14]	Zhenzhou WANG, Fangfang GUO, Jingfang SU, He SU, Jianchao WANG. Robustness optimization method of visual model for intelligent inspection [J]. Journal of Computer Applications, 2025, 45(7): 2361-2368.
[15]	Qiaoling QI, Xiaoxiao WANG, Qianqian ZHANG, Peng WANG, Yongfeng DONG. Label noise adaptive learning algorithm based on meta-learning [J]. Journal of Computer Applications, 2025, 45(7): 2113-2122.