proprietary-data-generator

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

• Third-party content exposure detected (high risk: 0.90). This skill's workflow explicitly instructs the agent to perform web_search and scrape public sources (e.g., Step 1: web_search queries like "[niche] reddit" and "statistics 2025", and Data Aggregation: "web scraping targets", public databases, APIs), meaning the agent will ingest untrusted, user-generated third-party content and use it to drive decisions about survey design and subsequent actions.