tasks

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

• Third-party content exposure detected (high risk: 0.80). The skill's shared Map Phase explicitly allows and instructs agents to fetch and read arbitrary web pages using WebFetch or a Chrome-navigation flow (see "_shared/coordination/map-phase.md" → "網頁抓取策略" and related WebFetch/Chrome steps), which exposes the agent to untrusted public third‑party content that can be read and used in its analysis and task-generation workflow.