sn-search-image

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

• Third-party content exposure detected (high risk: 0.75). The skill calls Serper.dev at runtime and prints/returns the response fields (e.g., titles, page links, image URLs) from that public search API; those results are outsider-authored web content metadata that becomes LLM-readable text via the script’s stdout (formatted output / URL-only lines / raw JSON).