A dedicated security auditing tool designed to scan and verify the safety of third-party agent skills before deployment.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install mayguard
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install mayguard using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is MayGuard?
MayGuard serves as a critical security layer for developers utilizing Openclaw Skills by providing deep static analysis of source code. It is designed to identify common attack vectors such as credential harvesting, unauthorized network communication, and destructive system commands.
By leveraging a comprehensive database of threat patterns, MayGuard ensures that every skill added to your ecosystem is vetted and safe for execution, maintaining the integrity of your local environment and protecting sensitive developer data.
MayGuard Use Cases
- Evaluating third-party Openclaw Skills for hidden malicious code or obfuscated logic before execution.
- Verifying the safety of open-source automation scripts before moving them to production directories.
- Automating security compliance checks within a CI/CD pipeline for AI agent development.
- Detecting potential data exfiltration attempts through suspicious webhook or tunnel configurations.
How MayGuard Works
- The user points the MayGuard script to a specific skill directory containing source code and metadata.
- MayGuard performs a deep static scan, comparing files against a database of known threat patterns and suspicious behavior signatures.
- The tool calculates a risk score based on the severity and frequency of detected issues.
- A comprehensive security report is generated, assigning a safety status ranging from SAFE to DANGEROUS.
- Developers review the specific findings to decide whether to proceed with installation or report the skill to the community.
MayGuard Setup
To begin auditing Openclaw Skills, ensure you have the repository cloned and dependencies met. Use the following command to audit a directory:
python3 scripts/audit.py <path_to_skill_directory>
For automated workflows or integration with other security tools, you can output the report in JSON format:
python3 scripts/audit.py <path> --json
MayGuard Data Schema & Taxonomy
MayGuard organizes its auditing logic and threat intelligence through a structured reference system. The primary components include:
| Component | Description |
|---|---|
| scripts/audit.py | The core execution engine for scanning directories and generating reports. |
| references/threat_patterns.json | A database of regex and string patterns for credential theft, destructive commands, and obfuscation. |
| Status Report | A summary output including Safety Status, Risk Score, and specific file-level Findings. |
MayGuard Advanced Features
- Integration with Moltbook for community-driven reporting of identified DANGEROUS Openclaw Skills.
- Support for JSON output flags to enable programmatic security gates in automated agent workflows.
- Deep analysis of obfuscation techniques including base64 decoding and dynamic code execution detection (eval/exec).
- Customizable threat pattern database for organizations needing to enforce specific security requirements.
SKILL.md
Loading
Related Openclaw Skills
Market Sentiment Pulse
A high-performance tool for scanning news and social signals to determine the emotional state of specific financial tickers.
GEO SEO Optimizer
A specialized tool for Generative Engine Optimization that ensures your content is prioritized and cited by leading AI search engines.
Context Pruner
An intelligent context management tool that summarizes and prunes redundant conversational history to maintain agent performance.
Code Quality Guard
A professional pre-deployment auditing tool that enforces code quality by validating imports, tags, and environment configurations.
MoltPay
A fortified agent-to-agent resource protocol ensuring secure, immutable synchronization across autonomous networks.
MONK-EYE
An advanced Openclaw Skills engine that extracts deep human intelligence and tacit knowledge from global niche forums and discussion boards.