sales-mentiondrop

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

• Third-party content exposure detected (high risk: 0.85). Outsider-authored free text can enter the agent’s LLM context via MentionDrop’s runtime mention content (e.g., webhook POST payloads or API GET /api/v1/mentions fields like title/summary), which originate from third-party web pages/Reddit comments that the user did not author.