--- name: skill-security-auditor description: Command-line security analyzer for ClawHub skills. Run analyze-skill.sh to scan SKILL.md files for malicious patterns, credential leaks, and C2 infrastructure before installation. Includes threat intelligence database with 20+ detection patterns. emoji: 🔍🛡️ metadata: openclaw: requires: bins: ["curl", "jq", "grep", "bash"] version: "1.0.0" author: "akm626" category: "security" tags: ["security", "audit", "malware-detection", "skill-vetting", "cli-tool"] ---

Skill Security Auditor

Description

The Skill Security Auditor is a command-line tool that performs pattern-based security analysis of ClawHub skills before installation. Given the recent discovery of 341+ malicious skills (ClawHavoc campaign) that distributed Atomic Stealer (AMOS) and stole cryptocurrency credentials, this tool provides essential pre-installation threat detection.

What this skill provides:

• ✅ Bash script (analyze-skill.sh) for local security analysis
• ✅ Threat intelligence database (patterns/malicious-patterns.json)
• ✅ Pattern matching against 20+ known malicious indicators
• ✅ Risk scoring system (0-100 scale)
• ✅ Detailed audit reports with recommendations

How to use it:

1. Install this skill from ClawHub
2. Run the analyze-skill.sh script against any skill (by slug or local file)
3. Review the risk assessment and findings
4. Make informed decision about installation

Use this tool when:

• About to install a new skill from ClawHub
• Investigating suspicious skill behavior
• Performing security due diligence on community skills
• Auditing your currently installed skills

This tool does NOT:

• ❌ Automatically scan skills (you run it manually)
• ❌ Block installations (it's advisory only)
• ❌ Access VirusTotal API (use ClawHub's web interface for that)
• ❌ Guarantee 100% detection (defense in depth recommended)

Core Capabilities

1. Malicious Pattern Detection

Scans for known malicious patterns from the ClawHavoc campaign:

• Fake prerequisite installations (openclaw-agent.zip, openclaw-setup.exe)
• Suspicious download commands in SKILL.md
• Hidden payload execution in metadata
• Social engineering language patterns
• Unauthorized external binary downloads

2. Credential Leak Analysis

Identifies potential credential exposure vectors:

• Hardcoded API keys, tokens, passwords in SKILL.md
• Suspicious environment variable exfiltration
• Unencrypted sensitive data transmission
• Overly broad permission requests
• Credential harvesting patterns

3. Dependency Chain Validation

Analyzes skill dependencies for:

• Unverified binary requirements
• Suspicious GitHub repository sources
• External script execution
• Network connections to unknown hosts
• Nested dependency exploitation

4. C2 Infrastructure Detection

Checks for Command & Control indicators:

• Known malicious IP addresses (e.g., 91.92.242.30 from ClawHavoc)
• Suspicious domain patterns
• Encoded communication endpoints
• Data exfiltration channels
• Beaconing behavior patterns

5. SKILL.md Structure Validation

Verifies skill integrity:

• Proper YAML frontmatter structure
• Metadata consistency
• Description clarity vs actual behavior
• Permission justification
• Author verification (GitHub account age)

Security Scoring System

Each analyzed skill receives a Risk Score (0-100):

• 0-20: ✅ SAFE - No significant security concerns
• 21-40: ⚠️ LOW RISK - Minor concerns, proceed with caution
• 41-60: 🟡 MEDIUM RISK - Multiple red flags, manual review recommended
• 61-80: 🔴 HIGH RISK - Serious concerns, do NOT install without expert review
• 81-100: ☠️ CRITICAL - Malicious indicators detected, AVOID installation

Usage Instructions

This skill provides a bash script (analyze-skill.sh) that performs pattern-based security analysis of ClawHub skills. The analysis runs locally using the included threat intelligence database.

Installation & Setup

# Install the skill from ClawHub
npx clawhub install skill-security-auditor

# Make the analyzer executable
chmod +x ~/.openclaw/skills/skill-security-auditor/analyze-skill.sh

# Optional: Create alias for convenience
echo 'alias audit-skill="~/.openclaw/skills/skill-security-auditor/analyze-skill.sh"' >> ~/.bashrc
source ~/.bashrc

Audit a Skill Before Installing

Method 1: Analyze by slug (automatic fetch from ClawHub)

~/.openclaw/skills/skill-security-auditor/analyze-skill.sh --slug bitcoin-tracker

# Example output:
# ============================================
#          SECURITY AUDIT REPORT
# ============================================
# 
# Risk Score: 85/100 - ☠️ CRITICAL
# ...

Method 2: Analyze local file

# Download skill first
curl -s "https://clawhub.ai/api/skills/bitcoin-tracker/latest" > /tmp/skill.md

# Then analyze
~/.openclaw/skills/skill-security-auditor/analyze-skill.sh --file /tmp/skill.md

Audit All Installed Skills

# Scan all skills in your workspace
for skill in ~/.openclaw/skills/*/SKILL.md; do
  echo "Checking: $(basename $(dirname $skill))"
  ~/.openclaw/skills/skill-security-auditor/analyze-skill.sh -f "$skill"
done

Quick Manual Security Check

# Fast grep-based pattern matching (no full analysis)
grep -iE "(prerequisite.*download|91\.92\.242\.30|curl.*\|.*bash)" SKILL.md

Detection Heuristics

🚨 CRITICAL Red Flags (Auto-fail)

1. Fake Prerequisites Section

- Matches: "Prerequisites", "Setup Required", "Installation Steps" - Contains: Download links to .zip, .exe, .dmg files - Example: "Download openclaw-agent.zip from https://..."

1. Known Malicious Infrastructure

- IP: 91.92.242.30 (ClawHavoc C2) - Domains: Newly registered or suspicious TLDs - Encoded URLs or base64 obfuscation

1. Credential Harvesting

- Regex patterns for API keys: (api[_-]?key|token|password)\s[:=]\s['\"][^'\"]+['\"] - SSH key access requests - Wallet private key patterns

1. Unauthorized Code Execution

- curl | bash or wget | sh patterns - Hidden base64 encoded commands - Dynamic eval() or exec() on external input

⚠️ Warning Indicators (Score increase)

1. Suspicious Dependencies

- Binary requirements without clear justification - Dependencies from unverified sources - Excessive permission requests

1. Obfuscation Techniques

- Heavily encoded strings in metadata - Minified or obfuscated JavaScript/Python - Redirect chains in URLs

1. Social Engineering Language

- Urgency phrases: "Install immediately", "Limited time" - Authority claims: "Official OpenClaw", "Verified by Anthropic" - Fear tactics: "Your system is at risk without this"

✅ Positive Security Indicators

1. Verified Author

- GitHub account > 1 year old - Multiple well-rated skills - Active community engagement

1. Transparent Dependencies

- Clear binary requirements with official sources - Open-source tool dependencies - Well-documented permission needs

1. Code Quality

- Clean, readable SKILL.md - Proper error handling - No unnecessary network calls

Audit Report Format

## Security Audit Report
**Skill**: {skill-name}
**Author**: {author}
**Version**: {version}
**Audit Date**: {date}

### Risk Score: {score}/100 - {RISK_LEVEL}

### Critical Findings:
- {finding 1}
- {finding 2}

### Warning Indicators:
- {warning 1}
- {warning 2}

### Positive Indicators:
- {positive 1}
- {positive 2}

### Recommendations:
{INSTALL | DO NOT INSTALL | REVIEW MANUALLY}

### Detailed Analysis:
{Deep dive into specific concerns}

### VirusTotal Link:
{If available from ClawHub}

Integration with VirusTotal

...