openpayd

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

• Potentially malicious external URL detected (high risk: 0.80). The skill runs the Membrane CLI against an external app URL (e.g., membrane connection ensure "https://www.openpayd.com/") at runtime and the returned connection object may include clientAction.agentInstructions which can supply instructions that directly control the AI agent's behavior, so https://www.openpayd.com/ is a runtime source of agent-controlling content.

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

• Direct money access detected (high risk: 1.00). The skill is an explicit integration with OpenPayd — a payments infrastructure platform — and exposes account, transaction, and transfer operations. It documents using Membrane to run pre-built actions or proxy raw API requests (including POST/PUT) against OpenPayd, which are clearly meant to send/receive/manage payments. This is a specific financial execution API (payment gateway/banking-like functionality), not a generic tool, so it grants direct financial execution capability.