sales-crowdreply

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

• Third-party content exposure detected (medium risk: 0.65). Outsider free text can enter the LLM context via the platform’s runtime “AI visibility tracking” and “social listening” features, which ingest and analyze third-party web/social content (e.g., Reddit/Quora/Facebook mentions and scraped pages) to generate prompts/insights that the agent then uses.