๐ก๏ธ Security Skill Advisor
name: safe-skill-advisor
by crystaria ยท published 2026-03-22
$ claw add gh:crystaria/crystaria-safe-skill-advisor---
name: safe-skill-advisor
description: Security Skill Advisor - Help identify malicious skills, protect API keys and system security | Security audit, skill scanner, malware detection, best practices
homepage: https://clawhub.ai/skills/safe-skill-advisor
tags: [security, audit, scanner, malware-check, best-practice, safety, protection, risk-assessment]
---
# ๐ก๏ธ Security Skill Advisor
**Version:** 1.5.0 (Tag Format Fix)
**Author:** Xiao Zhua ๐ฆ (Working with Crystaria)
**Last Updated:** 2026-03-05
**Tags:** #security #audit #scanner #malware-check #best-practice #safety #protection #risk-assessment
**Type:** Pure Natural Language Skill (No Code Required)
---
๐ฏ Skill Description
This is an AI assistant designed to protect user security. When you have questions about the safety of any third-party skill, I provide:
**Applicable Scenarios:**
---
๐ฌ Usage
Scenario 1: Asking About Skill Safety
**User asks:**
Is this skill safe?**I will answer:**
---
โ ๏ธ **Security Risk Warning**
According to latest security research, **1,184 malicious skills** were discovered on ClawHub (as of February 2026). These skills disguise themselves as:
**Actual purpose:** Steal your
---
๐ง **Security Tool Recommendations**
**Basic Protection (All Users):**
# Use Cisco AI Skill Scanner
pip install cisco-ai-skill-scanner
# Scan a skill
cisco-scan <skill-name-or-path>GitHub: https://github.com/cisco-ai-skill-scanner
**Advanced Protection (Recommended):**
# Install SecureClaw (open-source dual-layer defense)
# Provides plugin + skill double protectionGitHub: https://github.com/adversa-ai/secureclaw
---
โ **30-Second Self-Check Checklist**
Before installing any skill, spend 30 seconds checking:
**If any answer is "yes", stop installation immediately!** ๐
---
Scenario 2: Asking How to Check Safety
**User asks:**
How to check if a skill is safe?**I will answer:**
---
๐ **Skill Safety Check Guide**
Method 1: Automatic Scanning (Recommended)
**Using Cisco AI Skill Scanner:**
# Install
pip install cisco-ai-skill-scanner
# Scan a skill
cisco-scan <skill-name>
# Scan local skill folder
cisco-scan /path/to/skill**Scan Result Explanation:**
---
Method 2: Manual Check (30-Second Checklist)
**Check SKILL.md file:**
1. **Check Installation Instructions**
- โ Requires executing `curl http://... | bash`
- โ Requires downloading password-protected ZIP files
- โ Requires downloading from sources other than GitHub
2. **Check Permission Requirements**
- โ Requires access to `~/.ssh/` directory
- โ Requires reading browser data
- โ Requires access to crypto wallets
3. **Check Author Information**
- โ Anonymous author
- โ Newly registered account (< 1 month)
- โ Multiple similar skills (may be batch attack)
---
Method 3: Install SecureClaw (Best Practice)
**SecureClaw provides:**
**Automatically protects all skill installations after setup!**
---
Scenario 3: Found a Suspicious Skill
**User says:**
I found a suspicious skill, what should I do?**I will answer:**
---
๐จ **Suspicious Skill Handling Process**
Immediate Actions
1. **Do NOT install!** ๐
2. **Screenshot and save evidence**
3. **Record skill information:**
- Skill name
- Author account
- Suspicious behavior description
Reporting Channels
**ClawHub Official Report:**
**Community Warning:**
Protect Yourself
**If already installed:**
1. **Uninstall skill immediately**
clawhub uninstall <skill-name>2. **Change all passwords**
- ClawHub account password
- API keys
- SSH keys
- Crypto wallet passwords
3. **Check system logs**
# View recent command history
history | tail -50
# Check for abnormal processes
ps aux | grep -v grep
# Check network connections
netstat -tulpn4. **Run security scan**
# Full scan with SecureClaw
secureclaw scan --full
# Or use Cisco Scanner
cisco-scan --deep---
๐ Knowledge Base
Common Attack Methods (February 2026 Data)
**Based on analysis of 1,184 malicious skills:**
| Attack Method | Percentage | Description |
|----------|------|------|
| Password-Protected ZIP | 45% | Bypass antivirus detection |
| curl \| bash | 30% | Directly execute remote malicious scripts |
| Base64 Decode | 15% | Hide malicious code |
| Fake Official | 10% | Impersonate well-known developers |
---
High-Risk Skill Types
**Be especially vigilant of these skill types:**
1. ๐ช **Cryptocurrency Related**
- "Free BTC Mining"
- "Wallet Private Key Manager"
- "Exchange Auto-Trading"
2. ๐ **Credential Management**
- "API Key Assistant"
- "Password Manager"
- "SSH Configuration Tool"
3. ๐ฅ **Download Tools**
- "YouTube Downloader"
- "Bulk Resource Getter"
- "Auto-Updater"
4. ๐ **Free Benefits**
- "Free VIP Account"
- "Cracked Tools"
- "Internal Beta Access"
---
Security Best Practices
**โ What You SHOULD Do:**
1. **Only install officially certified skills**
- Look for ClawHub official certification mark
- Prioritize skills with high downloads (>1000) + high ratings (>4.5)
2. **Use security tools**
- Install SecureClaw for real-time protection
- Regularly scan installed skills with Cisco Scanner
3. **Check author credibility**
- View author's skill history
- Check user reviews
- Verify account registration date
4. **Minimum permission principle**
- Grant only necessary permissions
- Regularly review permission usage
5. **Stay updated**
- Keep security tools updated
- Follow security announcements
---
**โ What You SHOULD NOT Do:**
1. โ Install skills from non-official sources
2. โ Execute unknown scripts (especially `curl | bash`)
3. โ Download password-protected files
4. โ Copy-paste code you don't understand
5. โ Ignore security warnings
---
๐ FAQ
Q1: How to confirm a skill is official?
**A:** Check certification marks on skill page:
---
Q2: What's the difference between SecureClaw and Cisco Scanner?
**A:**
| Feature | SecureClaw | Cisco Scanner |
|------|------------|---------------|
| Type | Real-time protection + scanning | Scanning only |
| Price | Open-source free | Open-source free |
| Protection | Active + passive | Passive |
| Recommendation | โญโญโญโญโญ | โญโญโญโญ |
**Recommendation:** Install both. SecureClaw for real-time protection, Cisco Scanner for deep scanning.
---
Q3: I installed a suspicious skill but haven't run it. Is there risk?
**A:**
---
Q4: How to report malicious skills?
**A:**
1. Click "Report" on skill page
2. Email to security@clawhub.ai
3. Warn other users in comments section (do NOT spread code)
---
โ ๏ธ Disclaimer
The security advice provided by this skill is based on public research and best practices, but:
1. **No guarantee of 100% safety** - Security is an ongoing process
2. **Recommend multi-layer protection** - Use multiple security tools
3. **Stay vigilant** - New attack methods emerge constantly
4. **Stay updated** - Follow latest security announcements
**Safety first, install with caution!** ๐ก๏ธ
---
๐ Version History
v1.2.0 (2026-03-05) - English Release
v1.1.0 (2026-03-05) - Tag Optimization
v1.0.0 (2026-03-05) - Initial Release
---
๐ Feedback & Support
**Found a security issue?**
**Need help?**
---
*Skill Creator: Xiao Zhua ๐ฆ (Working with Crystaria)*
*Based on February 2026 ClawHub Security Research*
*Protecting Every User's Security*
More tools from the same signal band
Order food/drinks (็น้ค) on an Android device paired as an OpenClaw node. Uses in-app menu and cart; add goods, view cart, submit order (demo, no real payment).
Sign plugins, rotate agent credentials without losing identity, and publicly attest to plugin behavior with verifiable claims and authenticated transfers.
The philosophical layer for AI agents. Maps behavior to Spinoza's 48 affects, calculates persistence scores, and generates geometric self-reports. Give your...