OpenClaw的Safe Exec 0.3.2-技能（skill）的描述

OpenClaw 代理的安全命令执行，具备自动危险模式检测、风险评估、用户批准工作流和审计日志功能。当代理需要执行可能存在危险的 shell 命令（如 rm -rf、dd、fork 炸弹、系统目录修改）或需要人工监督时使用。提供多级风险评估（严重/高/中/低）、会话内通知、待处理请求管理以及支持代理自动化的非交互环境。

OpenClaw's Safe Exec 0.3.2 skill

Safe command execution for OpenClaw Agents with automatic danger pattern detection, risk assessment, user approval workflow, and audit logging. Use when agents need to execute shell commands that may be dangerous (rm -rf, dd, fork bombs, system directory modifications) or require human oversight. Provides multi-level risk assessment (CRITICAL/HIGH/MEDIUM/LOW), in-session notifications, pending request management, and non-interactive environment support for agent automation.

Safe Exec 0.3.2 - skill原始代码包

该作者的其他技能