upload-ad-sample

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

• Third-party content exposure detected (high risk: 0.75). The skill ingests outsider-authored free text when it reads a real ad build’s thread_json (e.g., threads/full-thread.json) and derives extracted_script by parsing the messages[].text fields, which are then embedded into the LLM context and sent as extracted_script in the registration payload.