clawhub-quarantine-installer_skill

Overview

This skill installs ClawHub skills into an isolated quarantine directory and runs a basic automated audit to surface potential security risks. It is designed to let you safely inspect third-party or flagged skills in a disposable environment before promoting them to production. The workflow produces a text audit report and preserves the quarantined files for manual review.

How this skill works

The installer script places the requested skill under $HOME/.openclaw/clawhub-quarantine/skills/ using npx clawhub install --force, then invokes an audit routine that scans the skill with ripgrep for risky patterns (command execution, network calls, secrets, environment access). A timestamped report is written to $HOME/.openclaw/clawhub-quarantine/reports/. You can then open the quarantined skill directory to perform deeper manual inspection and dependency checks.

When to use it

• When ClawHub flags a skill as suspicious or risky and you need to investigate safely.
• Before promoting third-party skills into your production OpenClaw workspace.
• When you want a quick automated scan for common dangerous code patterns.
• To archive and audit historical versions of a skill for incident response.
• When testing potentially malicious or untrusted skills inside a disposable VM or container.

Best practices

• Run the installer only in an isolated VM or container with no access to production secrets or networks.
• Treat the automated report as an aid—always perform manual code and dependency review before promotion.
• Verify external dependencies (npm packages) separately, checking their repositories and issue trackers.
• Keep Node.js and ripgrep up to date and limit network access from the quarantine environment.
• Remove or snapshot the quarantine environment after analysis to avoid lingering risks.

Example use cases

• Install a skill flagged by VirusTotal Code Insight to inspect for eval, secret keys, or outbound network calls.
• Audit archived versions of a skill to trace when a risky dependency was introduced.
• Validate whether a skill’s dependencies make system-level calls before allowing it into production.
• Generate a timestamped audit record for compliance or security review workflows.
• Use the quarantine as a safe workspace to reproduce reported problematic behavior.