ad-creative

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

• Third-party content exposure detected (medium risk: 0.30). Source ad-creative reads performance “data” provided by the user (including free-text pasted CSV/API output) in Mode 2 (“When the user provides performance data (CSV, paste, or API output)”), which the agent then analyzes at runtime.

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

• Direct money access detected (high risk: 1.00). The skill includes explicit integration commands for ad platform campaign management — e.g., "google-ads campaigns create" — which is a specific API-level capability to create/manage ad campaigns (and can be used to set/update budgets/ad spend). That qualifies as direct financial execution related to managing ad spend.