According to CoinMeta, OpenAI, Google, and Tencent Cloud have successively issued risk statements regarding the PHP vulnerability. These institutions pointed out that cybersecurity is entering a phase of "AI versus AI," with one company achieving a breakthrough in attacks targeting AI Agent, while another has launched the industry's first agent trust framework.