safe-exec

Overview: SafeExec is a skill for OpenClaw Agents that provides secure command execution capabilities with automatic danger pattern detection, risk assessment, user approval workflow, and audit logging. It’s useful for executing shell commands that may be dangerous or require human oversight. Key Features: Automatic danger pattern detection, risk-based interception, in-session notifications, user approval workflow, complete audit logging, agent-friendly non-interactive mode support, and platform-agnostic design. How It Works: SafeExec monitors all shell command executions, intercepts potentially dangerous commands, and requests user approval through in-session terminal notifications. Use Cases: Automated workflows with human oversight via audit logs, executing shell commands that require human approval, and ensuring secure command execution in OpenClaw Agents.