skill-security-audit

What This Skill Does

The skill-security-audit tool is a mission-critical utility for the OpenClaw ecosystem, designed to inspect, scan, and validate agent skills before they are executed or shared. In an environment where AI agents can interact with local files, environment variables, and network resources, this skill acts as a proactive security layer. It performs static analysis on skill directories to detect potential threats such as hardcoded API credentials (OpenAI, AWS, GitHub), execution of high-risk shell commands (e.g., rm -rf, sudo), attempts to access sensitive system paths (e.g., ~/.ssh), and unauthorized network communication. By integrating security into the workflow, this skill ensures that your agent infrastructure remains resilient against malicious or poorly written code.

Installation

To integrate this security suite into your environment, use the OpenClaw hub CLI: clawhub install openclaw/skills/skills/xiwan/jiajia-security-audit

Once installed, you can configure your environment for better accessibility by adding aliases to your shell configuration file (e.g., ~/.bashrc or ~/.zshrc): alias skill-audit='bash ~/clawd/skills/skill-security-audit/scripts/audit.sh' This allows you to initiate scans with a single command, streamlining the security pipeline for your development lifecycle.

Use Cases

• Pre-Installation Validation: Automatically scan any third-party skill from the marketplace before letting it run on your system.
• Continuous Deployment: Integrate the safe-publish.sh script into your CI/CD pipeline to ensure that no skill with a CRITICAL vulnerability is ever released to your team or the public.
• Regulatory Compliance: Perform periodic audits of existing skill libraries to detect "drift" or unauthorized changes that may have introduced security regressions.
• Vulnerability Patching: Use the detailed scan output to identify and remediate outdated dependencies or insecure configuration patterns in your custom-built skills.

Example Prompts

1. "Perform a security audit on the ./skills/my-new-plugin folder to ensure there are no sensitive keys exposed before I use it."
2. "Run a full system security scan on all installed skills and report any high-risk shell command usage."
3. "Check the current skill directory for security vulnerabilities and output the results in JSON format so I can process them in my logging system."

Tips & Limitations

• False Positives: If the auditor flags legitimate code (e.g., test files or mock data), use the # security-audit: ignore-next-line directive to suppress warnings.
• Environment Scope: This tool is best used as a gatekeeper; it cannot replace runtime behavioral monitoring or sandboxing. Always run skills with the principle of least privilege.
• Strictness: When in doubt, run with the --include-docs flag to ensure that credentials hidden in documentation or comment files are also swept.
• 不可绕过: Note that safe-publish.sh enforces a zero-tolerance policy for CRITICAL vulnerabilities, ensuring the integrity of the ecosystem.