Detects and scores prompt injection attempts in text, outputting severity, action, and matched rules without external calls or secret handling.
查看全部AI 与机器学习技能