A Security Analysis of the OpenClaw AI Agent Framework

Surada Suwansathit, Yuxuan Zhang, Guofei Gu

AI agent frameworks connecting large language model (LLM) reasoning to host execution surfaces -- shell, filesystem, containers, and messaging -- introduce security challenges stru…