jina-web-fetcher

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

• Third-party content exposure detected (high risk: 1.00). SKILL.md explicitly instructs using r.jina.ai to fetch arbitrary public URLs (e.g., "curl -s 'https://r.jina.ai/http://目标URL'" and examples like Hacker News, Google search results, GitHub Trending), so the skill ingests untrusted, user-generated web content that the agent would read and could influence subsequent actions.