Security of Cross-Platform and Cross-Language Software Technologies: A Systematic Review

Abstract

Cross-platform frameworks and cross-language programming techniques have become the foundation for building modern mobile, desktop, IoT, and cloud applications, while large language models (LLMs) are increasingly used to generate and translate their code. These technologies promise “write once, run anywhere” development and rapid multi-language integration, but they also introduce new attack surfaces at abstraction layers, language boundaries, and AI-assisted tooling. This paper presents a comprehensive security-centric survey of cross-platform and cross-language technologies, with a particular focus on how emerging code LLMs reshape the threat landscape. We first provide a unified taxonomy of cross-platform approaches (native, Web, hybrid, interpreted, cross-compiled, and model-driven) and cross-language mechanisms (source-to-source translation, foreign-function interfaces, inter-process communication, and multi-language runtimes), highlighting their security assumptions and typical attack surfaces. We then systematically review security threats to cross-platform frameworks, cross-language interactions, and LLM-based code generation, including permission misuse, remote code execution, cross-language attacks, data and cryptographic weaknesses, supply-chain compromise, reverse engineering, and AI-induced vulnerabilities. Building on this analysis, we organize existing defenses into a “prevention–detection–remediation–audit” lifecycle, covering framework hardening, unified cross-language analysis, supply-chain governance, and LLM-aware safeguards. Finally, we distill best practices for engineering reliable cross-platform and cross-language applications and outline key open research directions.