nft

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly queries the public Hiro API for NFT data (see SKILL.md "Query operations ... use the public Hiro API" and the nft.ts use of getNftService) and the AGENT.md/Output Handling state that untrusted NFT metadata and history are read and used for display and decision logic (e.g., gating transfers), so arbitrary third-party/user-generated content can influence agent actions.

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

• Direct money access detected (high risk: 1.00). The skill explicitly includes a "transfer" subcommand that sends NFTs on the Stacks L2 network, requires an unlocked wallet, accepts a fee parameter, and returns a transaction id/explorer URL. This is a specific crypto/blockchain transaction capability (signing/sending on-chain transfers), which is direct financial execution authority.