product-photography

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s required runtime workflow uses the user’s free-text “request” as the dominant constraint to draft an image prompt (see “The user's request always outranks…” and “Write the prompt using…”), meaning outsider-authored text is ingested by the LLM without needing to select a specific trusted item first.