fittings

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

• Third-party content exposure detected (medium risk: 0.30). In scripts/call.mjs, the runtime path /catalog/search?q=... and /catalog/{id} ingests outsider-provided free text via the --search query parameter and returns free-form service descriptions that the agent prints, before any specific item is selected.

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

• Direct money access detected (high risk: 1.00). The skill explicitly implements paid calls settled in USDC over the x402 protocol and requires a wallet private key (FITTINGS_PRIVATE_KEY) so the agent can sign off-chain USDC transfer requests. It describes the payment handshake (402 Payment Required), signing the returned challenge, and instructs exporting a spending wallet private key. These are direct crypto wallet/payment operations (signing and authorizing transfers), i.e. direct financial execution authority.