product-marketing-context-global

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

• Third-party content exposure detected (high risk: 1.00). The variant files (e.g., variants/01-us.md Section 9 "Voice of Customer" and the equivalent sections in the EU/SEA/LATAM variants) explicitly instruct pulling verbatim quotes from public, user-generated sources like G2, Trustpilot, App Store/Google Play reviews, Reddit, TikTok comments, Shopee/Lazada reviews and other forums — content that is fetched/ingested and then read by downstream global skills as authoritative context, which could enable indirect prompt injection.