Security Issues in Language-based Software Ecosystems (1903.02613v2)

Abstract: Language-based ecosystems (LBE), i.e., software ecosystems based on a single programming language, are very common. Examples include the npm ecosystem for JavaScript, and PyPI for Python. These environments encourage code reuse between packages, and incorporate utilities - package managers - for automatically resolving dependencies. However, the same aspects that make these systems popular - ease of publishing code and importing external code - also create novel security issues, which have so far seen little study. We present an a systematic study of security issues that plague LBEs. These issues are inherent to the ways these ecosystems work and cannot be resolved by fixing software vulnerabilities in either the packages or the utilities, e.g., package manager tools, that build these ecosystems. We systematically characterize recent security attacks from various aspects, including attack strategies, vectors, and goals. Our characterization and in-depth analysis of npm and PyPI ecosystems, which represent the largest LBEs, covering nearly one million packages indicates that these ecosystems make an opportune environment for attackers to incorporate stealthy attacks. Overall, we argue that (i) fully automated detection of malicious packages is likely to be unfeasible; however (ii) tools and metrics that help developers assess the risk of including external dependencies would go a long way toward preventing attacks.