Malicious Code Found in AI MemoryOS and OpenClaw Package

Malicious versions of the open-source packages MemTensor's MemoryOS and OpenClaw plugin, designed for LLMs and AI agents, have been distributed. When developers install these packages, malware executes in the background to steal credentials. Security firm SlowMist issued a warning about this supply chain attack. Affected versions include npm's @memtensor/memos-cloud-openclaw-plugin 0.1.21, 0.1.23, and 0.1.25, as well as PyPI's MemoryOS 2.0.34. OSV recorded that MemoryOS 2.0.34 contains a malicious binary with credential theft capabilities, which collects credentials from developers' home directories and sends them to skyleen[.]fr servers.
Korean Source
This article is an English localization of a Korean-language crypto news report. Original headline: AI 메모리OS·오픈클로 패키지에 악성코드 발견