cult-film-curtis

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

• Third-party content exposure detected (high risk: 0.80). The skill's runtime handlers (fetchCultFilms, fetchMovieDetails, searchCultFilms) explicitly fetch data from the public TMDB API (https://api.themoviedb.org/...), which is community-sourced/untrusted content that the agent reads and uses to filter and select recommendations, so third-party text could materially influence its behavior.

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

• Direct money access detected (high risk: 1.00). The skill explicitly integrates a payments subsystem and is built to accept micropayments. It imports and uses @lucid-agents/payments with paymentsFromEnv(), defines pricing for entrypoints in USDC, and requires environment variables EVM_PRIVATE_KEY, PAYMENTS_RECEIVABLE_ADDRESS, FACILITATOR_URL and NETWORK. Those variables and the payments module indicate direct crypto payment handling (wallet/private-key usage and a facilitator endpoint) rather than a generic API caller. This is specifically designed to move/receive money (micropayments), so it grants direct financial execution capability.