eval_results.json

{
  "average_total": 0.518,
  "clause_completeness": 1.0,
  "playbook_compliance": 0.114,
  "missing_key_terms": 0.92,
  "invented_legal_terms": 0.32,
  "business_usefulness": 0.133,
  "internal_consistency": 0.24,
  "risk_flag_accuracy": 0.633,
  "citation_support": 0.304
}