workflow-automation

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The Inngest "AI Orchestration" workflow in SKILL.md explicitly takes event.data.task (from external events/webhooks), passes it to llm.chat to generate a plan, and then executes subtasks—meaning untrusted third‑party/user-provided content can directly drive LLM output and subsequent actions.

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

• Direct money access detected (high risk: 1.00). The skill explicitly includes payment-processing code and references to payment gateways: examples call chargeCard/process-payment, show stripe.charges.create, and demonstrate Stripe paymentIntent usage with idempotency keys. It also discusses payment-specific retry handling (e.g., card_declined, InsufficientFunds). These are concrete, payment-specific integrations (Stripe), not generic API or browser automation, so the skill grants direct financial execution capability.