self-hosted-funnel-launch

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

• Potentially malicious external URL detected (high risk: 0.90). The skill includes commands that fetch and execute remote runtime artifacts—curl -O https://raw.githubusercontent.com/autonnel/autonnel/master/docker-compose.yml (downloaded and run with docker compose up) and the container image ghcr.io/autonnel/autonnel:latest (pulled and executed via docker run)—so external content is fetched at runtime and executed.

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

• Direct money access detected (high risk: 1.00). The skill is explicitly built to take money: it documents checkout/order flows, requires real end-to-end purchases and refunds for testing, and lists Payments providers (Stripe, PayPal). It exposes an admin API (MCP) for agents with mutating/write-capable tools and references order-related tools (list_orders, deliver_order). Those elements indicate the skill includes specific payment gateway integrations and agent-accessible mutating APIs that can affect orders/payments, so it grants direct financial execution capability.