agent-md-creator

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs the agent to perform online research using an external tool ("需要在线调研时使用 Tavily" / "use Tavily") and includes external sources/URLs (e.g., agents.md site and GitHub blog) that the agent may fetch and interpret, exposing it to untrusted public web content that can influence subsequent actions.