customer-support-verification

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md requires collecting the “customer request, Gmail thread/draft ids, and current draft body” and “duplicate-thread check” details, which at runtime will include outsider-authored customer email/thread text (Gmail/customer communications) fed into the agent’s LLM context for verification.