pinecone-research

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

• Third-party content exposure detected (medium risk: 0.30). In the required runtime workflow, scripts/memory_manager.py ingests outsider-supplied free text via --text (store action) and later reads it back for recall via Pinecone query results (match["metadata"]["text"]), meaning an outsider can submit poison text that the agent workflow will retrieve without first selecting a specific pre-existing item.