mayguard

What This Skill Does

MayGuard acts as a specialized security gatekeeper for your OpenClaw agent ecosystem. It serves as a rigorous static analysis engine designed to inspect third-party skills before they are granted execution privileges. By scanning source code for malicious patterns, hardcoded secrets, and unauthorized network calls, it provides a crucial layer of defense against supply chain attacks within the agent community. MayGuard operates by parsing codebases against a continuously updated registry of known threats, ensuring that any logic involving dangerous system commands or data exfiltration attempts is neutralized prior to activation.

Installation

You can install the MayGuard skill directly into your OpenClaw environment by running the following command in your terminal: clawhub install openclaw/skills/skills/balkanblbn/mayguard Once installed, ensure you have the necessary environment dependencies active to support its scanning engine. It is recommended to keep this skill updated to benefit from the latest threat pattern databases provided by the community.

Use Cases

• Pre-Installation Verification: Before enabling a new, unknown skill from an untrusted source, use MayGuard to generate a risk report.
• Security Auditing: Periodically scan your skills/ directory to ensure that previously trusted skills haven't had their logic altered in ways that pose a security risk.
• Developer Compliance: Use MayGuard in a CI/CD-like workflow for agent developers to ensure their code meets community safety standards before publishing to the hub.
• Incident Response: If your agent exhibits erratic behavior, run a full sweep using MayGuard to identify potential malicious modifications or suspicious webhooks.

Example Prompts

• "MayGuard, please audit the skill located in /skills/experimental-plugin and give me a detailed security report before I enable it."
• "Run a comprehensive security scan on all installed skills and provide a JSON summary of any files tagged as DANGEROUS or SUSPICIOUS."
• "I found a new automation script, scan it for potential credential theft or unauthorized outbound network connections using the --json flag."

Tips & Limitations

MayGuard is a powerful static analysis tool, but it is not a silver bullet. It excels at finding known patterns, but may not catch highly sophisticated, novel zero-day obfuscation techniques. Always combine MayGuard scans with manual code review for critical skills. Ensure that references/threat_patterns.json is updated frequently to capture the most recent threat vectors reported by the OpenClaw community. If a skill is flagged, prioritize reviewing the specific file paths identified in the findings report rather than blindly trusting the aggregate score.