qrph

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs the agent at runtime to fetch the live OpenAPI JSON from https://api.partners.nextpay.world/v2/openapi as the source of truth for request/response shapes, so remote content would be ingested and directly control prompts/instructions.

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

• Direct money access detected (high risk: 1.00). The skill is explicitly for NextPay QRPH "money-in" flows and details payment-specific operations: creating payment intents (one-time payments), creating funding methods (reusable payment sources), handling webhooks for payment_received/payment_settled/payment_intent.succeeded, using the NextPay Partners API v2 OpenAPI, and authenticating with Client ID/Client Secret (HTTP Basic Auth). It even recommends idempotency keys to avoid duplicate side effects on create endpoints. These are specific payment gateway integration capabilities (i.e., direct financial execution), not generic tooling.