llmquant-prediction-markets

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

• Third-party content exposure detected (high risk: 0.75). The required runtime workflow ingests “related evidence such as … news, polling, sector prices, or issuer context” and “Retrieve … news, macro releases” from LLMQuant Data sources, which are outsider-authored free-text (e.g., fetched news/articles) that the agent reads into the LLM context.