hive-nft-research

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

• Third-party content exposure detected (medium risk: 0.65). The workflow calls search_tools / invoke_api_endpoint to fetch NFT metadata, owner, floor, sales, rarity, and spam from external data provider marketplaces/APIs at runtime, which can include outsider-authored free text (e.g., scraped metadata fields, provider-returned descriptions) that may be ingested into the LLM context.