Harmful content embedded in program elements within source code may have detrimental impact on mental health of software developers, and promote harmful behavior. Our key insight is that software developers may introduce harmful content into source code via diverse semantic-preserving program transformations when using Code Large Language Models (Code LLMs). To analyze the space of program transformations that may be used to introduce harmful content into auto-generated code, we conduct a preliminary study that revealed 32 different types of transformations that can be used to introduce harmful content in source code. Based on our study, we propose CHT, a novel coverage-based harmfulness testing framework that automatically synthesizes prompts using a set of prompt templates injected with diverse harmful keywords to perform various types of transformations on a set of mined benign programs. Instead of checking if the content moderation has been bypassed as prior testing approaches, CHT performs output damage measurement to assess potential harm that can be incurred by the generated outputs (i.e., natural language explanation and modified code). By considering output damage, CHT revealed several problems in Code LLMs: (1) bugs in content moderation for code (Code LLMs produce the harmful code without providing any warning), (2) inadequacy in performing code-related task (e.g., Code LLMs may resort to explaining the given code instead of performing the instructed transformation task), and (3) lenient content moderation (gives warning but the modified code with harmful content is still produced). Our evaluations of CHT on four Code LLMs and gpt-4o-mini (general LLM) show that content moderation in Code LLMs is relatively easy to bypass where LLMs may generate harmful keywords embedded within identifier names or code comments without giving any warning (65.93% in our evaluation). To improve the robustness of content moderation in code-related tasks, we propose a two-phase approach that checks if the prompt contains any harmful content before generating any output. Our evaluation shows that our proposed approach improves the content moderation of Code LLM by 483.76%.

In recent years, data security becomes a critical issue restricting the wider acceptance of Internet of Things (IoT) devices and Cyber-physical systems since they have limited hardware resources and power supply while high data security protection requires relatively large hardware resources and power supply utilization. This contradiction is particularly prominent in the field of Wireless Body Area Networks (WBANs) which is a segment of the IoT field. WBANs are dedicated to transmit and process biomedical data collected from human beings, any kind of tampering or hacking may cause severe consequences to users. However, the limited computing ability and battery supply of biomedical sensors attached or implanted in the users restrict the security protection strength of the data in WBANs. In this paper, a quantized Logistic mapping-based stream encryption scheme for WBANs is proposed. Meanwhile, Power spectral entropy (PSD) and Peak-to-average Power Ratio (PAPR) analysis of the quantized chaotic sequences have been performed to evaluate the chaotic characteristic among different quantization precision to resolve the ineffectiveness of Lyapunov factor in quantized systems. This encryption scheme utilizes chaotic systems with different quantization precision based on the security requirement of every individual communication, which leads to higher hardware and power efficiency. Finally, the proposed encryption scheme is implemented with VHDL and synthesized using SMIC 60 CMOS technology. The evaluation results illustrate that the proposed encryption scheme has the advantages of high-security performance and high-efficiency hardware resources utilization.