The Hong Kong Computer Emergency Response Team Coordination Centre (HKCERT) warned that AI agent platforms, which possess capabilities such as local operations, third-party plugin installations and external service integrations, pose greater risks than general chat-based AI tools. Organizations should assess version risks, supply chain risks and permission management arrangements when introducing such tools, and avoid executing high-risk operations suggested by agents without verification.Related NewsCiti Keeps BABA-W (09988.HK) Rating at Buy w/ TP $199As the popularity of OpenClaw continues to rise, associated security risks are increasingly surfacing, the HKCERT added. It is reported that malicious attackers have used fake GitHub code base and Bing AI search results to spread information-stealing malware and agent-type malware to users searching for OpenClaw Windows installers.
AASTOCKS Financial News