alchemy-agentic-gateway

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill instructs agents to extract and replay sensitive auth/payment tokens (e.g., SIWE tokens, Payment-Signature, Authorization: Payment credentials) into request headers and shows CLI usage that can pass private keys, which requires including secret values verbatim in generated commands/requests and thus poses an exfiltration risk.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly instructs running npx/npm to install and execute remote packages required for runtime (e.g., @alchemy/x402 and mppx, fetched from the npm registry such as https://registry.npmjs.org/@alchemy/x402), which causes remote code to be executed and is required for the x402/MPP workflows.

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

• Direct money access detected (high risk: 1.00). The skill explicitly implements payment protocols and tools to move funds. It documents wallet-based USDC payments (x402 using EIP-3009 / Solana x402), an MPP path that supports on-chain USDC via Tempo and Stripe card payments, and client libraries/CLI commands that perform payments (e.g., npx @alchemy/x402 pay, mppx account fund, and mppx which "handles auth and payment automatically"). It also references creating/funding wallets, extracting Payment-Signature/Authorization headers, and using Stripe as a payment method. These are specific, dedicated financial execution capabilities (crypto transfers and card payments), not generic tooling.