In recent years, the growing complexity and scale of source code have rendered manual software vulnerability detection increasingly impractical. To address this challenge, automated approaches leveraging machine learning and code embeddings have gained substantial attention. This study investigates the optimal combination of code embedding techniques and machine learning classifiers for vulnerability detection in Python source code. We evaluate three embedding techniques, i.e., Word2Vec, CodeBERT, and GraphCodeBERT alongside two deep learning classifiers, i.e., Bidirectional Long Short-Term Memory (BiLSTM) networks and Convolutional Neural Networks (CNN). While CNN paired with GraphCodeBERT exhibits strong performance, the BiLSTM model using Word2Vec consistently achieves superior overall results. These findings suggest that, despite the advanced architectures of recent models like CodeBERT and GraphCodeBERT, classical embeddings such as Word2Vec, when used with sequence-based models like BiLSTM, can offer a slight yet consistent performance advantage. The study underscores the critical importance of selecting appropriate combinations of embeddings and classifiers to enhance the effectiveness of automated vulnerability detection systems, particularly for Python source code.

Software vulnerabilities are a fundamental reason for the prevalence of cyber attacks and their identification is a crucial yet challenging problem in cyber security. In this paper, we apply and compare different machine learning algorithms for source code vulnerability detection specifically for Python programming language. Our experimental evaluation demonstrates that our Bidirectional Long Short-Term Memory (BiLSTM) model achieves a remarkable performance (average Accuracy = 98.6%, average F-Score = 94.7%, average Precision = 96.2%, average Recall = 93.3%, average ROC = 99.3%), thereby, establishing a new benchmark for vulnerability detection in Python source code.

BGP Blackholing is an effective mitigation solution for networks to counter the frequent Distributed Denial of Service (DDoS) attacks. It enables to drop all network traffic that is directed towards a particular victim prefix under DDoS attack, ideally, as close to the source as possible. Despite its huge importance in the Internet, there is no tool available for the real-time visualization of BGP Blackholing activity. Visualization is one of the most powerful techniques for network operators to monitor network activity. From discovering successful network topology to expose anomalous behaviors in networks, easy-to-use visualizations are powerful weapons to capture important patterns on the Internet traffic[1, 5]. In this work, we propose a first real-time BGP Blackholing Visibility Tool (named as SmartX BGP-BVT) to detect and visualize community based BGP Blackholing on live BGP data. This tool will be helpful for network operators and researchers interested in BGP Blackholing service and DDoS mitigation in the Internet.

Accurate Internet traffic predictions can provide support to network operators for applications such as traffic engineering, bandwidth allocation, anomaly detection, etc. We apply and compare different forecasting techniques (traditional and machine learning-based techniques) on real BGP data that is collected from two well-known Internet exchange points (IXPs) to derive BGP future volume-based predictions. Our experimental evaluation shows that multivariate Bayesian Ridge outperforms all other forecasting techniques we consider. Through univariate LSTM, we are able to predict new BGP volume-based features. Furthermore, to study the impact of dataset size on BGP forecasting, we perform experiments on three BGP dataset sizes, i.e., Short (one-month), Medium (three-months), and Long (five-months) Periods. Our results show that the Short-Period BGP dataset seems to be sufficient for getting accurate predictions. We also present a use case study (forecast Google Leak anomaly) that supports our experimental evaluations. We provide our collected BGP datasets publically which will be helpful to perform further research experiments and analysis regarding BGP traffic predictions.

Direct access to the system's resources such as the GPU, persistent storage and networking has enabled in-browser crypto-mining. Thus, there has been a massive response by rogue actors who abuse browsers for mining without the user's consent. This trend has grown steadily for the last months until this practice, i.e., CryptoJacking, has been acknowledged as the number one security threat by several antivirus companies.

We introduce a new browser abuse scenario where an attacker uses local storage capabilities without the website's visitor knowledge to create a network of browsers for persistent storage and distribution of arbitrary data. We describe how security-aware users can use mechanisms such as the Content Security Policy (CSP), sandboxing, and third-party tracking protection, i.e., CSP & Company, to limit the network's effectiveness. From another point of view, we also show that the upcoming Suborigin standard can inadvertently thwart existing countermeasures, if it is adopted.

The term "phishing" describes a class of social engineering attacks on authentication systems, that aim to steal the victim's authentication credential, e.g., the username and password. The severity of phishing is recognized since the mid-1990's and a considerable amount of attention has been devoted to the topic. However, currently deployed or proposed countermeasures are either incomplete, cumbersome for the user, or incompatible with standard browser technology. In this paper, we show how modern JavaScript API's can be utilized to build PhishSafe, a robust authentication scheme, that is immune against phishing attacks, easily deployable using the current browser generation, and requires little change in the end-user's interaction with the application. We evaluate the implementation and find that it is applicable to web applications with low efforts and causes no tangible overhead.

Sanitizable signature schemes allow for altering signed data in a signer-controlled way by a semi-trusted third party. This is contrary to standard digital signature schemes, which do not permit any modifications by any party without invalidating the signature. Due to transparency, a strong privacy notion, outsiders cannot see if the signature for a message was created by the signer or by the semi-trusted party. Accountability allows the signer to prove to outsiders if a message was original or touched by the semi-trusted party. Currently, block-level accountability requires to drop transparency. We allow for accountability for sanitizable signatures with transparency on the block-level. Additionally, we generalize the concept of block-level properties to groups. This offers a even more fine-grained control and leads to more efficient schemes. We prove that group-level definitions imply both the block-level and message-level notions. We derive a provably secure construction, achieving our enhanced notions. A further modification of our construction achieves efficient group-level non-interactive public accountability. This construction only requires a constant amount of signature generations to achieve this property. Finally, we have implemented our constructions and the scheme introduced by Brzuska et al. at PKC '09 and provide a detailed performance analysis of our reference implementations.

Cross-site scripting (XSS) has emerged to one of the most prevalent type of security vulnerabilities. While the reason for the vulnerability primarily lies on the server-side, the actual exploitation is within the victim's Web browser on the client-side. Therefore, an operator of a Web application has only very limited evidence of XSS issues. In this paper, we propose a passive detection system to identify successful XSS attacks. Based on a prototypical implementation, we examine our approach's accuracy and verify its detection capabilities. We compiled a data-set of 500.000 individual HTTP request/response-pairs from 95 popular web applications for this, in combination with both real word and manually crafted XSS-exploits; our detection approach results in a total of zero false negatives for all tests, while maintaining an excellent false positive rate for more than 80% of the examined Web applications.

Binary decision diagrams (BDDs) are a well-known tool for representing Boolean functions. We show how BDDs can be extended to full first-order logic by integrating means for representing quantifiers. The resulting structures are called Shannon graphs. A calculus based on these Shannon graphs is set up, and its soundness and completeness proofs are outlined. A comparison of deduction with first-order BDDs and semantic tableaux shows that both calculi are closely related. From a practical perspective, however, BDDs have advantages over tableaux: they provide a more compact representation, since BDDs can be understood as a linear, graphical representation of a fully expanded tableaux. Furthermore, BDDs represent not only the models of a formula, but also its counter-models: this offers a very efficient way to represent lemmata during the proof search. The last part of the paper introduces a compilation-based approach to implementing deduction systems based on Shannon graphs. The idea is to compile the graphs into programs that carry out the proof search at run time.