monad

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent to fetch runtime content from raw GitHub URLs (e.g. https://raw.githubusercontent.com/therealharpaljadeja/monskills/main/concepts/references/async-execution.md), which would be retrieved at runtime and injected as skill instructions, so a remote URL directly controls agent prompts.

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

• Direct money access detected (high risk: 1.00). The skill set explicitly includes crypto/blockchain transaction capabilities: the "Wallet" skill describes agent wallet management, Safe multisig creation, deploying smart contracts and performing on-chain actions via Safe multisig, and proposing transactions to a Safe Transaction Service with EIP-712 signatures. The "Wallet Integration" and "Addresses" skills further provide wallet/auth integrations (embedded MPC wallets, external wallet connect) and canonical onchain addresses. These are specific tools for managing wallets, signing and submitting blockchain transactions — i.e., direct financial execution.