IAPS warns on Kimi Claw: security and governance risks of Chinese-hosted 'always-on' AI agents

IAPS·low signal

The Institute for AI Policy and Strategy published a research note on Kimi Claw — Moonshot AI's managed OpenClaw offering (native on Kimi.com with ~5,000 community skills and 40GB cloud storage) — flagging the risks of always-on autonomous agents running on Chinese-hosted infrastructure with persistent data access. It frames the data-sovereignty and supply-chain exposure of routing agent execution and skills through a foreign cloud. Relevant to the broader thread of government AI blacklisting and agent supply-chain designation as policy levers.