trade-sentiment

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

• Third-party content exposure detected (high risk: 0.95). The required workflow explicitly instructs runtime “targeted searches” for recent news and social media (e.g., Reddit/StockTwits/X) and to extract “Last 10-15 headlines” and “viral posts,” which are outsider-authored free text fetched from public web sources and then ingested into the agent’s LLM context for analysis.