Prompts are the primary mechanism for communicating with AI agents, and they directly influence the quality and reliability of AI-generated code. As AI-assisted programming becomes widely adopted, modern tools increasingly combine dynamic conversational prompts with static configuration-like prompt files. Despite the growing focus on prompt engineering, prior research has primarily focused on conversational prompts, while prompt files remain understudied. To address this gap, we conduct an empirical study of configuration prompt files in Cursor, a widely used AI-assisted code editor. We collect and analyze over 12,110 .cursorrules files from 11,427 GitHub repositories to characterize their distribution, evolution, and maintenance. Complementing this, we perform qualitative analysis on a random sample of 65 prompt files and develop a 65-code codebook capturing how developers express programming intent, project context, engineering practices, and security considerations. Our results show that .cursorrules files emerged rapidly from mid-2024. Their adoption is concentrated in small-scale, low-activity, single-maintainer repositories, suggesting toy projects rather than professional development. The content of prompt files is dominated by guidance on code quality and engineering practices, project structure and configuration, and maintainability, while security-related content appears less frequently. Our analysis shows that there is a continuity of themes and topics between the now-legacy .cursorrules files and the current standard .mdc files.

Side-channel analysis (SCA) poses a real-world threat by exploiting unintentional physical signals to extract secret information from secure devices. Evaluation labs also use the same techniques to certify device security. In recent years, deep learning has emerged as a prominent method for SCA, achieving state-of-the-art attack performance at the cost of interpretability. Understanding how neural networks extract secrets is crucial for security evaluators aiming to defend against such attacks, as only by understanding the attack can one propose better countermeasures. In this work, we apply mechanistic interpretability to neural networks trained for SCA, revealing $\textit{how}$ models exploit $\textit{what}$ leakage in side-channel traces. We focus on sudden jumps in performance to reverse engineer learned representations, ultimately recovering secret masks and moving the evaluation process from black-box to white-box. Our results show that mechanistic interpretability can scale to realistic SCA settings, even when relevant inputs are sparse, model accuracies are low, and side-channel protections prevent standard input interventions.

The deep learning-based profiling attacks have received significant attention for their potential against masking-protected devices. Currently, additional capabilities like exploiting only a segment of the side-channel traces or having knowledge of the specific countermeasure scheme have been granted to attackers during the profiling phase. In case either capability is removed, a practical profiling attack faces great difficulty and complexity. To address this challenge, we propose an efficient and scheme-agnostic Leakage Distillation-based Profiling Attack (LD-PA). By distilling univariate leakage from a reference, we can train an encoder that extracts multivariate leakage from raw traces and transforms it into an effective representation (transitional leakage). An indirect connection between multivariate leakage and the target variable is established by bridging through the transitional leakage, thereby facilitating the inference of leaked values. Remarkably, LD-PA achieves successful attacks on multiple public datasets using a simple multilayer perceptron (MLP) without necessitating an exhaustive hyperparameter search, while its performance is competitive with state-of-the-art methods. Simultaneously, we delve into the nature of transitional leakage, confirming the existence of combined leakage. This, in turn, validates that the guidance from univariate leakage references aids in the combination of multivariate leakage. Besides that, each component of the multivariate leakage is extracted and stacked in a highly aligned manner. Moreover, we explored several factors impacting LD-PA performance, covering scenarios with limited profiling traces, noisy references, alternative references, and hyperparameter tuning.