Exec Inspector

What This Skill Does

The Exec Inspector is a powerful observability tool designed for the OpenClaw ecosystem. It provides deep visibility into the execution history, patterns, and real-time activity of the OpenClaw Agent's exec tool. By tracking every shell command invoked by the AI, it transforms opaque execution history into actionable insights. The tool includes a robust command-line utility for querying historical logs and a background daemon service for monitoring command execution in real-time. Whether you are debugging an agent's reasoning process or auditing system interactions, Exec Inspector acts as your primary dashboard for agent activity.

Installation

To add this skill to your OpenClaw environment, use the following command in your terminal: clawhub install openclaw/skills/skills/zhengzk-bj/exec-inspector

Once installed, ensure that your shell environment is configured to recognize the script location. We highly recommend setting up the provided aliases in your .zshrc or .bashrc for immediate access to monitoring features. You may need to verify that your user account has appropriate permissions to execute scripts located in ~/.openclaw/scripts/.

Use Cases

• Debugging Agent Behavior: Understand why an agent took a specific action or why a specific command failed during an automated session.
• Efficiency Auditing: Identify frequently used commands that might be candidates for automation or optimization within your workflows.
• Real-Time Monitoring: Watch exactly what the agent is doing as it happens in a separate terminal window, perfect for development and testing phases.
• Security & Compliance: Maintain a clear audit trail of every command executed by an AI agent on your host system.

Example Prompts

1. "最近执行了哪些命令？请帮我列出最近 20 条，并帮我分析一下我今天的操作习惯。"
2. "查看今天的命令统计，我想知道哪个命令是我用得最多的。"
3. "我想查找之前关于 git 相关的操作记录，请搜索并显示所有的 git 执行历史。"

Tips & Limitations

• Proactive AI: Remember that the AI is trained to act, not just advise. Do not ask "how" to run the history tool; simply ask the AI to perform the check for you.
• Daemon Resource Usage: While the monitoring daemon is lightweight, remember to stop it using exec-monitor stop when you are finished to save system resources.
• Data Privacy: This tool stores execution history locally. If you are handling highly sensitive API keys or credentials within commands, be aware that these will be recorded in the local JSON export files generated by the tool.
• Context Awareness: When using the search function, be as specific as possible with the command keywords to ensure the best retrieval results.