lynx-strategy

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

• Third-party content exposure detected (medium risk: 0.65). At runtime, the producer calls the outsider-authored free-text field ai_reasoning via audit_query in fetch_closed_trades(), then parses it with parse_score_from_reasoning() and uses the extracted score to update MIN_SCORE (indirect prompt-injection risk through untrusted text from other users’/system-generated trade records).