meta-ads-connect

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow reads outsider-authored free text via the Meta Ads account data pulled from the user’s account (e.g., campaign/ad/creative fields) when running skills/meta-ads-connect/scripts/audit.py, where campaign/ad creative text or metadata authored by third parties could be included in the fetched datasets.

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

• Direct money access detected (high risk: 1.00). The skill explicitly includes mutation commands that change ad spend: the mutate.py CLI supports --update-adset-budget and --update-campaign-budget and also requires the Meta permission ads_management for applying changes. Those are specific APIs/functions to modify advertising budgets (direct ad spend management), which meets the definition of Direct Financial Execution.