Openclaw Safe Guard

What This Skill Does

The Openclaw Safe Guard (openclaw-safe-guard) is a specialized security auditing agent for the OpenClaw ecosystem. Designed to prioritize user security, this skill performs static analysis on local or external OpenClaw skills before you execute them. It functions as a preventative layer, scanning code for malicious patterns, excessive permission requests, or vulnerable dependency chains. It provides a structured, easy-to-read risk assessment report in Chinese, enabling users to make informed decisions about the software they integrate into their workspace.

Installation

To install Openclaw Safe Guard, ensure your system meets the essential dependency requirements: curl, jq, git, grep, and find. Once dependencies are verified, run the installation command: clawhub install openclaw/skills/skills/ansengu11/openclaw-safe-guard. It is highly recommended to perform this installation in a non-root environment or a containerized sandbox to maintain optimal system isolation. Always review the source code on the GitHub repository provided by the author, ansengu11, before finalizing installation.

Use Cases

• Pre-Installation Auditing: Before installing a community-made skill, use Safe Guard to scan its repository for suspicious API calls or unauthorized file access patterns.
• System Health Check: Regularly scan your ~/.openclaw/skills directory to ensure that previously installed skills haven't introduced risks or outdated dependencies.
• Security Compliance: Use the tool to generate a standardized risk assessment score (Low, Medium, or High) to verify if a skill adheres to your personal or professional safety requirements.

Example Prompts

1. "扫描 Skill 安全，检查一下我刚下载的 AI 助手插件风险高吗？"
2. "安全审计：帮我看看 ~/.openclaw/skills 目录里安装的所有技能是否安全。"
3. "这个 Skill 安全吗？帮我分析一下 ansengu11 发布的这个插件。"

Tips & Limitations

• Static Analysis Only: This skill does not execute code; it performs static analysis. It is excellent at catching hardcoded sensitive keys or common shell injection patterns, but it cannot detect sophisticated runtime exploits.
• Privacy Awareness: Note that this skill requires read access to your ~/.openclaw directories. While it is designed to audit these paths, be aware of what sensitive data resides in these folders.
• Network Dependency: The tool requires active internet access to query the GitHub API for metadata, though it keeps the actual repository data in a temporary /tmp directory that is purged after scanning.