safe-exec

Intercept dangerous shell commands from OpenClaw agents with risk assessment and approval workflows.

Updated Apr 3, 2026

Search Tags:#automation #security #audit #shell #risk-assessment #openclaw #safe-exec

Ready to run

npx skills add https://github.com/Ghenghis/openclaw-claude-code-integration --skill safe-exec-ghenghis

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: safe-exec
Source: https://github.com/Ghenghis/openclaw-claude-code-integration/tree/main/skills/safe-command-execution
Command: npx skills add https://github.com/Ghenghis/openclaw-claude-code-integration --skill safe-exec-ghenghis

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

SafeExec intercepts potentially dangerous shell commands issued by OpenClaw agents and provides automatic risk detection, approval workflows, and audit logging to prevent system compromise.

Core Features & Use Cases

Automatic danger pattern detection and interception of risky commands.
Multi-level risk assessment (CRITICAL/HIGH/MEDIUM/LOW) with in-session prompts and audit trails.
Agent-mode support with non-interactive operation and comprehensive logging for compliance and debugging.
Use cases include automated task pipelines, secure agent orchestration, and audits of command activity.

Quick Start

Enable SafeExec to automatically intercept dangerous shell commands and require approvals for risky actions.

Frequently Asked Questions about safe-exec

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I prevent dangerous shell commands from running in automated agent workflows?▼

You can prevent dangerous shell commands in automated workflows by intercepting them to identify risky patterns before execution. This approach applies multi-level risk assessment and requires user approval for high-risk actions.

What is multi-level risk assessment for shell command execution?▼

Multi-level risk assessment for shell command execution categorizes commands into CRITICAL, HIGH, MEDIUM, or LOW risk levels. It identifies dangerous patterns before execution to trigger appropriate safety protocols like user approval.

Does safe command execution work with non-interactive agent orchestration?▼

Yes, safe command execution supports non-interactive agent orchestration through an agent mode. This mode enables automated operation while maintaining comprehensive audit logging for compliance and debugging purposes.

How do I audit shell command activity for security compliance?▼

You can audit shell command activity by enabling comprehensive audit logging that records intercepted commands and their risk assessments. This generates an audit trail of all command executions and approval workflows for compliance review.

Can I require manual approval for high-risk shell commands in automated pipelines?▼

Yes, you can require manual approval for high-risk shell commands in automated pipelines through optional user approval workflows. The system manages pending requests and sends in-session notifications when dangerous patterns are detected.