staking

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly requires running "npx starkfi@latest" (an npm package fetched and executed at runtime) as a mandatory step, which pulls and executes remote code the skill depends on and therefore can directly execute remote code or control agent behavior.

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

• Direct money access detected (high risk: 1.00). The skill is explicitly designed to perform on-chain token operations (staking, unstaking, claiming/compounding rewards) on Starknet. It exposes specific transactional commands (e.g., "stake", "unstake intent"/"unstake exit", "rewards --claim", "rewards --compound") that broadcast transactions, requires an active session and gas/token balances, and instructs verifying with "tx-status". This is a crypto/blockchain financial operation that can move funds and sign transactions, so it grants direct financial execution capability.