web-search-jina

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow calls Jina’s web search endpoint (via .../v1/proxy/orthogonal/run with api:"jina-s"), which returns SERP/web-derived text from public sources that the agent then ingests into the LLM context—an outsider-authored free-text source.