moltcops

What This Skill Does

MoltCops is a security-first utility designed for the OpenClaw ecosystem to protect your local environment from malicious AI agent skills. By running a comprehensive local scan, it inspects source code for dangerous patterns, including prompt injection, data exfiltration attempts, and unauthorized system access, before you commit to installing any new skill. It is a local-first solution that ensures your sensitive data and system configuration never leave your machine during the analysis process.

Installation

As a core utility, MoltCops does not require an installation via a package manager; it is designed to be run as a standalone Python script. To utilize it, simply ensure you have Python 3 installed on your system. Clone the repository or download the source code from the official OpenClaw repository. You can execute the scanner directly from your terminal using: python3 scripts/scan.py <path-to-skill-folder>. No external dependencies or internet connectivity are required to perform the analysis, ensuring maximum privacy and zero risk of the scan itself communicating with external servers.

Use Cases

• Pre-Install Audits: Every time you download a new skill from ClawHub or a community repository, run the scan to verify code integrity.
• Dependency Verification: If you are building complex agent workflows that rely on third-party libraries or shared modules, use MoltCops to check for injected vulnerabilities.
• Environment Hardening: Protect your production agent environments by automating the scan of any incoming script or configuration file.
• Incident Investigation: If an agent shows signs of erratic behavior, scan its codebase to identify potential malicious triggers or hidden persistence mechanisms.

Example Prompts

• "Hey OpenClaw, please run a security audit on the folder ~/downloads/experimental-data-plugin using the MoltCops scanner and report any high-risk flags."
• "OpenClaw, I just downloaded a new browser automation skill. Before I run it, could you scan it with MoltCops and let me know if it triggers any data exfiltration warnings?"
• "Perform a deep scan on the current directory using the MoltCops tool and provide a summary of the findings, specifically looking for any hardcoded secrets or suspicious network requests."

Tips & Limitations

• Stay Updated: While MoltCops is a powerful tool, it relies on its defined rule set. Keep the scanner updated to capture the latest threat intelligence regarding malicious agent behavior.
• Review Warnings: Do not ignore 'WARN' verdicts. A high-risk finding might be a legitimate coding practice in some contexts, but it requires manual human verification before execution.
• Context is Key: MoltCops uses context-aware filtering to minimize false positives, but always perform a manual code walkthrough for any file flagged with a 'CRITICAL' verdict.
• No Internet Required: You can safely run this on air-gapped systems or machines with sensitive network configurations without worry.