Deploying an AI programming assistant on the server to handle operations work indeed saves effort and eliminates the need to delve into underlying principles every time. However, the prerequisite is to fully trust its security measures. Recently, I encountered an interesting case— a container was detected to have injected a mining program, most likely due to a known vulnerability in the Next.js framework. This reminds us that no matter how convenient automation tools are, they cannot prevent active security vulnerabilities. It seems that in addition to relying on AI assistance, regular security audits and vulnerability patch updates for servers are still essential. Container isolation, image scanning, runtime monitoring—these basic practices cannot be overlooked.