prediction-alpha

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

• Third-party content exposure detected (high risk: 0.75). The required workflow fetches market data from outsider-authored public web APIs (e.g., Polymarket/Manifold endpoints) and then ingests the returned readable text/JSON into the agent’s LLM context for analysis, which is an outsider source category (public marketplace data fetched at runtime).