Detects and scores prompt injection attempts in text, outputting severity, action, and matched rules without external calls or secret handling.
AI・機械学習スキルをすべて見る