golden-dataset-validation

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

• Third-party content exposure detected (low risk: 0.10). The required pre-addition workflow (validate_before_add) reads the full content of a new, user-selected document to generate embeddings and check for duplicates, so outsider-authored text from arbitrary selected URLs can reach the agent, but only after the user chooses the document and the skill never forces delivery.