opensea

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

• Third-party content exposure detected (medium risk: 0.30). At runtime, the opensea-api workflow ingests outsider-authored free text via get_collections/get_items/get_tokens auto-resolve query parameters (and also search_* queries) and can ingest untrusted user-generated metadata returned by OpenSea APIs (e.g., NFT names/descriptions) into the LLM context.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill explicitly references and routes to sub-skills that perform on-chain financial actions: buy/sell NFTs and sweeps, fulfill listings, ERC20 token swaps via a DEX aggregator, and configuring wallet signing (Privy/Turnkey/Fireblocks/Bankr). These are specific crypto/blockchain transaction and signing capabilities (i.e., directly moving assets or signing transactions), so it constitutes Direct Financial Execution authority.