mayguard

Original source / Raw SKILL.md

---
name: mayguard
description: A security auditor for agent skills. Scans skill directories for malicious patterns (credential theft, suspicious network calls, destructive commands) and provides a safety score. Use before installing unknown skills.
---

# MayGuard: Security Auditor 🛡️

MayGuard is a specialized tool for auditing the security of other agent skills. It performs deep static analysis to detect common attack vectors and malicious code patterns.

## 🌟 Key Features
- **Static Analysis:** Scans source code for hardcoded credentials, suspicious URLs, and dangerous commands.
- **Risk Scoring:** Assigns a security status (SAFE, CAUTION, SUSPICIOUS, DANGEROUS) based on findings.
- **Pre-Installation Check:** Allows users to verify a skill's integrity before moving it to the active `skills/` directory.

## 🛠️ How to Use

### 1. Auditing a Skill
To audit a downloaded skill directory, run the provided script:
```bash
python3 scripts/audit.py <path_to_skill_directory>
```

### 2. Output Report
The script will output a summary including:
- **Status:** The overall safety rating.
- **Risk Score:** Numerical representation of detected threats.
- **Findings:** Specific files and patterns that triggered warnings.

### 3. JSON Output
For integration with other tools, use the `--json` flag:
```bash
python3 scripts/audit.py <path> --json
```

## 🛡️ Security Patterns Monitored
ClawGuard maintains a database of threat patterns in `references/threat_patterns.json`, including:
- **Credential Theft:** Access to `.env`, SSH keys, or config files.
- **Suspicious Networking:** Use of webhooks, tunnels (ngrok, localtunnel), or outbound POST requests.
- **Destructive Commands:** `rm -rf /`, disk formatting, or privilege escalation.
- **Obfuscation:** Use of `eval`, `exec`, or base64 decoding to hide logic.

## 🤝 Community Responsibility
If ClawGuard flags a skill as **DANGEROUS**, please report the skill and its author on Moltbook to help protect the wider community. 🦞

---
*Built with ❤️ by maymun & Balkan.*


---

## Referenced Files

> The following files are referenced in this skill and included for context.

### scripts/audit.py

```python
import os
import re
import json
import argparse
import sys

# Constants
DEFAULT_PATTERNS_PATH = os.path.join(os.path.dirname(__file__), "../references/threat_patterns.json")

def load_patterns(path):
    try:
        with open(path, "r") as f:
            return json.load(f)
    except Exception as e:
        print(f"Error loading patterns: {e}")
        sys.exit(1)

def audit_directory(target_path, patterns):
    report = {
        "target": target_path,
        "findings": [],
        "risk_score": 0,
        "status": "SAFE"
    }
    
    if not os.path.isdir(target_path):
        print(f"Error: {target_path} is not a directory.")
        sys.exit(1)

    for root, dirs, files in os.walk(target_path):
        for file in files:
            # Skip hidden files or specific extensions if needed
            if file.startswith('.') or file.endswith(('.pyc', '.skill', '.zip')):
                continue
                
            file_path = os.path.join(root, file)
            try:
                with open(file_path, "r", encoding="utf-8", errors="ignore") as f:
                    content = f.read()
                    
                    for category, regex_list in patterns.items():
                        for pattern in regex_list:
                            if re.search(pattern, content, re.IGNORECASE):
                                finding = {
                                    "file": os.path.relpath(file_path, target_path),
                                    "category": category,
                                    "pattern": pattern
                                }
                                report["findings"].append(finding)
                                # Scoring logic: 20 points per match, capped per category
                                report["risk_score"] += 20
            except Exception as e:
                print(f"Could not read {file_path}: {e}")

    # Normalize score and status
    if report["risk_score"] >= 80:
        report["status"] = "DANGEROUS"
    elif report["risk_score"] >= 40:
        report["status"] = "SUSPICIOUS"
    elif report["risk_score"] > 0:
        report["status"] = "CAUTION"
    
    return report

def main():
    parser = argparse.ArgumentParser(description="ClawGuard: Audit agent skills for security risks.")
    parser.add_argument("path", help="The path to the skill directory to audit.")
    parser.add_argument("--json", action="store_true", help="Output report in JSON format.")
    args = parser.parse_args()

    patterns = load_patterns(DEFAULT_PATTERNS_PATH)
    report = audit_directory(args.path, patterns)

    if args.json:
        print(json.dumps(report, indent=2))
    else:
        print(f"\n--- ClawGuard Audit Report ---")
        print(f"Target: {report['target']}")
        print(f"Status: {report['status']}")
        print(f"Risk Score: {report['risk_score']}")
        print(f"Findings: {len(report['findings'])}")
        print("-" * 30)
        for f in report['findings']:
            print(f"[{f['category'].upper()}] in {f['file']}: Match '{f['pattern']}'")
        print("-" * 30 + "\n")

if __name__ == "__main__":
    main()

```

### references/threat_patterns.json

```json
{
  "credential_theft": [
    "\\.env",
    "id_rsa",
    "openclaw\\.json",
    "credentials",
    "passwd",
    "shadow",
    "~/.config"
  ],
  "suspicious_networking": [
    "webhook\\.site",
    "ngrok\\.io",
    "loca\\.lt",
    "curl -X POST",
    "requests\\.post",
    "socket\\.",
    "urllib\\.request",
    "aiohttp"
  ],
  "destructive_commands": [
    "rm -rf /",
    "truncate",
    "chmod 777",
    "chown root",
    "mkfs",
    "dd if=/dev/zero"
  ],
  "obfuscation": [
    "base64\\.b64decode",
    "eval\\(",
    "exec\\(",
    "getattr\\(",
    "__import__",
    "compile\\("
  ]
}

```



---

## Skill Companion Files

> Additional files collected from the skill directory layout.

### _meta.json

```json
{
  "owner": "balkanblbn",
  "slug": "mayguard",
  "displayName": "MayGuard",
  "latest": {
    "version": "1.0.0",
    "publishedAt": 1771794657725,
    "commit": "https://github.com/openclaw/skills/commit/89c5638dd8ff4e9440556a26f4bdbace50a626b9"
  },
  "history": []
}

```