qianwenai-observe

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The workflow reads application log files and slow query logs via cloud assistant commands, which can contain untrusted user-submitted free text or external inputs.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill contains instructions to download the Aliyun CLI installer via wget https://aliyuncli.alicdn.com/aliyun-cli-linux-latest-amd64.tgz, which constitutes an operational dependency retrieved from a remote URL during setup.