# Jailbreaking

China’s Cybersecurity Watchdog Sounds Alarm Over Rogue AI Agent ‘Skills’

China’s CNCERT has warned of malicious AI agent plugins being used for model jailbreaking and illegal crypto-mining, highlighting new security risks in the nation’s rapidly expanding AI ecosystem. The regulator emphasized that such 'Skills' could lead to account bans, device performance loss, and legal liabilities for users.