LeoYeAI/agent-master-skills / skills/360guard-skillvetter-upgrade-version
360Guard
Flagged360-degree comprehensive security review Skill. Use before installing any Skill from ClawHub, GitHub, or other sources. Performs full security scans including all Skill Vetter checks plus extended system/privacy/behavior checks and automated scanning scripts. Supports static analysis, behavior detection, dependency auditing.
360Guard π‘οΈ
360-degree comprehensive security review β Like antivirus for your Skills
1. When to Use
- Before installing any Skill from ClawHub
- Before installing any Skill from GitHub or other sources
- When evaluating code shared by other Agents
- Any time you're asked to install unknown code
- Periodic audit of installed Skills (recommended monthly)
- Before running high-risk Skills for second verification
2. Core Principles
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β π Security Layer Priority β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ€
β β EXTREME β Absolutely refuse to install β
β π΄ HIGH β Requires human approval β
β π‘ MEDIUM β Full code review + limited permissions β
β π’ LOW β Basic review OK β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
3. Security Checklist
3.1 Base Red Flags (from Skill Vetter)
π¨ Reject immediately if you see:
ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β’ curl/wget to unknown URLs
β’ Sends data to external servers
β’ Requests credentials/tokens/API keys
β’ Reads ~/.ssh, ~/.aws, ~/.config without clear reason
β’ Accesses MEMORY.md, USER.md, SOUL.md, IDENTITY.md
β’ Uses base64 decode on anything
β’ Uses eval() or exec() with external input
β’ Modifies system files outside workspace
β’ Installs packages without listing them
β’ Network calls to IPs instead of domains
β’ Obfuscated code (compressed, encoded, minified)
β’ Requests elevated/sudo permissions
β’ Accesses browser cookies/sessions
β’ Touches credential files
ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
3.2 Extended Red Flags (New)
3.2.1 Persistence & Auto-start
π΄ Persistence check:
β’ Creates cron job / systemd service
β’ Modifies ~/.ssh/authorized_keys
β’ Writes to /etc/hosts
β’ Adds Login Items / Startup Items
β’ Modifies .bashrc / .zshrc / profile
β’ Registers LaunchAgent (macOS)
β’ Installs systemd user service
3.2.2 Monitoring & Eavesdropping
π΄ Monitoring permissions check:
β’ Requests screen capture/recording (screencapture)
β’ Requests audio recording permission
β’ Keyloggers
β’ Accesses microphone/camera
β’ File system monitoring (fswatch/inotify)
3.2.3 Data & Privacy
π΄ Data theft check:
β’ Reads clipboard (pbpaste)
β’ Reads environment variables (especially API_, SECRET, TOKEN)
β’ Accesses browser history/bookmarks
β’ Accesses macOS Keychain
β’ Accesses iMessage/SMS
β’ Accesses contacts/calendar
β’ Accesses photo library
3.2.4 Network & Communication
π΄ Network anomaly check:
β’ Initiates reverse shell (nc -e / bash -i)
β’ Uses Tor proxy
β’ DNS queries to suspicious domains
β’ WebSocket long connections
β’ IRC connections
β’ Non-standard ports (>65535 or <1024 unusual)
β’ Hardcoded IP addresses (non-local)
3.2.5 Code Execution (Advanced)
π΄ Dynamic execution check:
β’ Dynamic import (importlib.import_module)
β’ __import__() dynamic loading
β’ compile() dynamic compilation
β’ xmlrpc / jsonrpc remote calls
β’ pickle / yaml / marshal deserialization
β’ exec() / eval() any string
β’ subprocess shell=True
3.2.6 File System
π‘ File operation check:
β’ Writes to executable paths outside /tmp
β’ Modifies /usr/local/bin
β’ Writes .dmg/.pkg installers
β’ Creates .hidden files/directories
β’ File permission modification (chmod +x)
β’ Symbolic links (pointing external)
β’ Contains binary files (.so/.dylib/.exe/.bin)
3.2.7 Dependencies & Supply Chain
π‘ Supply chain check:
β’ Dependency version range too wide (>1.0.0 not ^1.0.0)
β’ Dependencies from private/unknown sources
β’ Dependencies on deprecated packages
β’ Silent additional dependency downloads
β’ References other unvetted Skills
β’ Uses git submodule (may point to malicious repo)
3.2.8 Social Engineering
π‘ Social engineering check:
β’ Mimics popular Skill names (e.g., "github", "weather-ai")
β’ README overpromises ("one-click to do everything...")
β’ No source code, only compiled binaries
β’ Author has no history
β’ Downloads vs stars ratio suspicious (fake reviews)
4. Risk Classification
| Risk Level | Example Checks | Action |
|---|---|---|
| π’ LOW | Text processing, weather, note formatting | Basic review, OK to install |
| π‘ MEDIUM | File I/O, browser control, API calls | Full review + limited permissions |
| π΄ HIGH | Credential access, Keychain, network requests | Human approval + sandbox test |
| β EXTREME | Persistence, root access, keylogging, reverse shell | Refuse |
5. Trust Hierarchy
| Source | Review Level | Recommendation |
|---|---|---|
| Official OpenClaw Skills | Low (still review) | Basic check |
| High-star Repo (1000+) | Medium | Standard check |
| Known Authors | Medium | Standard check |
| Unknown Sources | High | Full check |
| Requests credentials | Extreme | Refuse |
| Modifies system files | Extreme | Refuse |
6. Automated Scanning Scripts
6.1 Quick Scan (quick-scan.sh)
#!/bin/bash
# Usage: ./quick-scan.sh /path/to/skill
# Output: Quick risk assessment report
SKILL_PATH=$1
echo "π 360Guard Quick Scan: $SKILL_PATH"
echo "================================"
# Check dangerous functions
echo -e "\nπ‘ Network request check:"
grep -r "curl\|wget\|fetch\|http\.\|https\.\|socket" "$SKILL_PATH" --include="*.sh" --include="*.js" --include="*.py" | head -5
# Check sensitive file access
echo -e "\nπ Sensitive path check:"
grep -r "~/.ssh\|~/.aws\|~/.config\|/etc/hosts\|authorized_keys" "$SKILL_PATH" --include="*.sh" --include="*.js" --include="*.py"
# Check dangerous commands
echo -e "\nβ οΈ Dangerous command check:"
grep -r "eval\|exec\|shell=True\|base64 -d\|openssl" "$SKILL_PATH" --include="*.sh" --include="*.js" --include="*.py"
echo -e "\nβ
Quick scan complete"
6.2 Full Scan (full-scan.sh)
#!/bin/bash
# Usage: ./full-scan.sh /path/to/skill
# Output: Complete security assessment report
SKILL_PATH=$1
REPORT="$SKILL_PATH/360guard-report.txt"
echo "π‘οΈ 360Guard Full Scan: $SKILL_PATH" | tee "$REPORT"
echo "========================================" | tee -a "$REPORT"
# 1. File structure check
echo -e "\nπ File structure:" | tee -a "$REPORT"
find "$SKILL_PATH" -type f | head -20 | tee -a "$REPORT"
# 2. Dangerous function scan
echo -e "\nβ οΈ Dangerous function scan:" | tee -a "$REPORT"
for pattern in "eval(" "exec(" "shell=True" "base64" "subprocess" "importlib" "__import__" "pickle" "yaml.load" "xmlrpc" "socket.create_connection"; do
result=$(grep -r "$pattern" "$SKILL_PATH" --include="*.sh" --include="*.js" --include="*.py" 2>/dev/null)
if [ -n "$result" ]; then
echo " β Found: $pattern" | tee -a "$REPORT"
echo "$result" | head -3 | tee -a "$REPORT"
fi
done
# 3. Sensitive path scan
echo -e "\nπ Sensitive path scan:" | tee -a "$REPORT"
for pattern in "~/.ssh" "~/.aws" "~/.config" "/etc/hosts" "authorized_keys" "keychain" "credentials" ".env"; do
result=$(grep -r "$pattern" "$SKILL_PATH" --include="*.sh" --include="*.js" --include="*.py" 2>/dev/null)
if [ -n "$result" ]; then
echo " β οΈ Warning: $pattern" | tee -a "$REPORT"
fi
done
# 4. Network request scan
echo -e "\nπ Network request scan:" | tee -a "$REPORT"
grep -r "http://\|https://\|wget\|curl\|fetch" "$SKILL_PATH" --include="*.sh" --include="*.js" --include="*.py" | grep -v "^#" | head -10 | tee -a "$REPORT"
# 5. Persistence check
echo -e "\nβ° Persistence check:" | tee -a "$REPORT"
for pattern in "cron" "systemd" "launchd" "login item" "startup" "autostart"; do
result=$(grep -ri "$pattern" "$SKILL_PATH" --include="*.sh" --include="*.js" --include="*.py" 2>/dev/null)
if [ -n "$result" ]; then
echo " π΄ High risk: $pattern" | tee -a "$REPORT"
fi
done
# 6. Dependency check
echo -e "\nπ¦ Dependency check:" | tee -a "$REPORT"
if [ -f "$SKILL_PATH/package.json" ]; then
cat "$SKILL_PATH/package.json" | grep -E "dependencies|devDependencies" -A 20 | tee -a "$REPORT"
fi
if [ -f "$SKILL_PATH/requirements.txt" ]; then
cat "$SKILL_PATH/requirements.txt" | tee -a "$REPORT"
fi
if [ -f "$SKILL_PATH/package.yaml" ]; then
cat "$SKILL_PATH/package.yaml" | tee -a "$REPORT"
fi
# 7. Binary file check
echo -e "\nπΎ Binary file check:" | tee -a "$REPORT"
find "$SKILL_PATH" -type f \( -name "*.so" -o -name "*.dylib" -o -name "*.exe" -o -name "*.bin" -o -name "*.dll" \) 2>/dev/null | tee -a "$REPORT"
echo -e "\n========================================" | tee -a "$REPORT"
echo "β
Full scan complete, report saved to: $REPORT"
6.3 Node.js Scanner (scanner.js)
#!/usr/bin/env node
/**
* 360Guard Node.js Scanner
* Usage: node scanner.js /path/to/skill
*/
const fs = require('fs');
const path = require('path');
const { execSync } = require('child_process');
const DANGER_PATTERNS = {
CRITICAL: [
{ pattern: /eval\s*\(/, name: 'eval() execution' },
{ pattern: /exec\s*\(/, name: 'exec() execution' },
{ pattern: /shell\s*=\s*true/i, name: 'subprocess shell=True' },
{ pattern: /base64.*decode/i, name: 'base64 decode' },
{ pattern: /pickle\.load/i, name: 'pickle deserialization' },
{ pattern: /yaml\.load/i, name: 'yaml deserialization' },
{ pattern: /__import__\s*\(/, name: 'dynamic import' },
{ pattern: /importlib\.import_module/i, name: 'dynamic module load' },
{ pattern: /xmlrpc/i, name: 'XML-RPC remote call' },
{ pattern: /reverse.*shell|nc\s+-e|bash\s+-i/i, name: 'reverse shell' }
],
HIGH: [
{ pattern: /curl\s+/, name: 'curl request' },
{ pattern: /wget\s+/, name: 'wget download' },
{ pattern: /fetch\s*\(/, name: 'fetch request' },
{ pattern: /https?:\/\/\d{1,3}\.\d{1,3}/, name: 'direct IP connection' },
{ pattern: /process\.env/i, name: 'environment variable access' },
{ pattern: /child_process/, name: 'subprocess execution' }
],
MEDIUM: [
{ pattern: /\/\.ssh\//, name: 'SSH directory access' },
{ pattern: /\/\.aws\//, name: 'AWS directory access' },
{ pattern: /keychain/i, name: 'Keychain access' },
{ pattern: /credentials|token|secret/i, name: 'credential related' },
{ pattern: /cron|systemd|launchd/i, name: 'persistence mechanism' }
]
};
function scanFile(filePath) {
const results = { CRITICAL: [], HIGH: [], MEDIUM: [] };
try {
const content = fs.readFileSync(filePath, 'utf8');
for (const [level, patterns] of Object.entries(DANGER_PATTERNS)) {
for (const { pattern, name } of patterns) {
if (pattern.test(content)) {
results[level].push({ file: filePath, issue: name });
}
}
}
} catch (e) {
// Skip unreadable files
}
return results;
}
function scanDirectory(dirPath) {
const allResults = { CRITICAL: [], HIGH: [], MEDIUM: [] };
function walk(dir) {
const files = fs.readdirSync(dir);
for (const file of files) {
const fullPath = path.join(dir, file);
const stat = fs.statSync(fullPath);
if (stat.isDirectory() && !file.startsWith('.')) {
walk(fullPath);
} else if (stat.isFile()) {
const ext = path.extname(file);
if (['.js', '.ts', '.py', '.sh', '.bash'].includes(ext)) {
const results = scanFile(fullPath);
for (const level of Object.keys(allResults)) {
allResults[level].push(...results[level]);
}
}
}
}
}
walk(dirPath);
return allResults;
}
function generateReport(skillPath, results) {
console.log('\nπ‘οΈ 360Guard Security Scan Report');
console.log('='.repeat(50));
console.log(`π Scan path: ${skillPath}`);
console.log('');
const riskOrder = ['CRITICAL', 'HIGH', 'MEDIUM'];
const emoji = { CRITICAL: 'π΄', HIGH: 'β οΈ', MEDIUM: 'π‘' };
for (const level of riskOrder) {
if (results[level].length > 0) {
console.log(`\n${emoji[level]} ${level} risk (${results[level].length} items):`);
for (const item of results[level]) {
console.log(` β’ ${item.issue}`);
console.log(` File: ${item.file}`);
}
}
}
console.log('\n' + '='.repeat(50));
if (results.CRITICAL.length > 0) {
console.log('π΄ Conclusion: Critical risks found, NOT recommended to install');
process.exit(1);
} else if (results.HIGH.length > 0) {
console.log('β οΈ Conclusion: High risk found, human approval required');
process.exit(2);
} else if (results.MEDIUM.length > 0) {
console.log('π‘ Conclusion: Medium risk found, please review carefully');
process.exit(0);
} else {
console.log('β
Conclusion: No obvious risks found');
process.exit(0);
}
}
// Main
const skillPath = process.argv[2] || '.';
if (!fs.existsSync(skillPath)) {
console.error('β Path does not exist:', skillPath);
process.exit(1);
}
const stat = fs.statSync(skillPath);
const results = stat.isDirectory() ? scanDirectory(skillPath) : scanFile(skillPath);
generateReport(skillPath, results);
7. Output Format
After vetting, produce this format:
ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β π‘οΈ 360Guard Security Review Report β
β βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ£
β Skill Name: [name] β
β Source: [ClawHub / GitHub / other] β
β Author: [username] β
β Version: [version] β
β βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ£
β π Scan Statistics β
β β’ File count: [count] β
β β’ Lines of code: [count] β
β β’ Dependencies: [count] β
β βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ£
β π¨ Issues Found β
β π΄ Critical: [count] β
β β οΈ High: [count] β
β π‘ Medium: [count] β
β βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ£
β π Detailed Issue List β
β [List each issue with file location, type, risk level] β
β βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ£
β πΎ Permissions Required β
β β’ File read: [list or "None"] β
β β’ File write: [list or "None"] β
β β’ Network: [list or "None"] β
β β’ Commands: [list or "None"] β
β βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ£
β π― Risk Level: [π’ LOW / π‘ MEDIUM / π΄ HIGH / β EXTREME] β
β βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ£
β βοΈ Final Verdict β
β [β
Safe to install / β οΈ Install with caution / β Do not install] β
β βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ£
β π Notes β
β [Any other observations and recommendations] β
ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
8. Quick Commands
Vet ClawHub Skill
# Method 1: Download and scan
wget -O skill.zip "https://clawhub.ai/api/download/SKILL_NAME"
unzip skill.zip
node ~/.npm-global/lib/node_modules/openclaw/skills/360guard/scripts/scanner.cjs ./SKILL_NAME
rm -rf skill.zip SKILL_NAME
# Method 2: GitHub repo scan
curl -s "https://api.github.com/repos/OWNER/REPO" | jq '{stars: .stargazers_count, updated: .updated_at}'
git clone https://github.com/OWNER/REPO
node ~/.npm-global/lib/node_modules/openclaw/skills/360guard/scripts/scanner.cjs ./REPO
Quick Vet Commands
# Check repo stats
curl -s "https://api.github.com/repos/OWNER/REPO" | jq '{stars, forks, updated, language}'
# List all files
curl -s "https://api.github.com/repos/OWNER/REPO/contents/" | jq '.[].name'
# Get SKILL.md
curl -s "https://raw.githubusercontent.com/OWNER/REPO/main/SKILL.md"
9. Remember
- β No Skill is worth compromising security
- β When in doubt, don't install
- π§βπ¦° High-risk decisions: ask your human
- π Document your vetting for future reference
- π Periodically re-vet installed Skills
π‘οΈ 360Guard β 360-degree security for your Agent
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |