retail-product-search

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

• Third-party content exposure detected (medium risk: 0.30). In scripts/agent.py, the required runtime workflow ingests outsider-authored free text from each user chat turn as query: str into retrieve_docs() → search_collection() via search_text=query (Vertex AI Vector Search), so prompt-injection text is read without needing the agent to pre-select a specific item.