Hong Kong Digital Policy Office Warns: Deploy OpenClaw With Security Risks in Mind and Take Protective Measures

The Hong Kong Digital Policy Office recently noted potential risks of the AI tool OpenClaw, including excessive permissions, data leaks, and system security issues. It advised relevant organizations and individuals to take sufficient safety measures when deploying and using the tool.

Specific measures include: strengthening network control and strictly isolating the runtime environment to reduce permission risks; enhancing credential management to avoid plaintext storage of keys; strictly reviewing plugin sources to ensure credibility and security; and promptly installing official patches and security updates.

AI security practitioners pointed out that these recommendations address core security challenges in the AI agent field and align with existing defense frameworks, reminding users to prioritize safety while enjoying automation capabilities.