crabukit 0.1.3

Security scanner for OpenClaw skills

🔒 Crabukit

A security scanner for OpenClaw skills.

Crabukit analyzes OpenClaw skills for security vulnerabilities, malicious code patterns, prompt injection attempts, and supply chain risks before installation.

🚀 Quick Start

# Install crabukit
pip install crabukit

# 🔒 Safely install a skill (downloads, scans, installs if safe)
crabukit install youtube-summarize

# Scan a skill before installing
crabukit scan ./my-skill/

# Scan an installed skill
crabukit scan /opt/homebrew/lib/node_modules/clawdbot/skills/suspicious-skill

# CI mode - fail on high severity
crabukit scan ./skill --fail-on=high

# JSON output for automation
crabukit scan ./skill --format=json

✨ Features

🔍 Comprehensive Detection

Category	Detections
Prompt Injection	Direct, indirect, encoded, typoglycemia attacks
Code Vulnerabilities	eval(), exec(), shell injection, path traversal
Secrets	AWS keys, GitHub tokens, OpenAI keys, JWTs, private keys
AI Malware	Self-modifying code, LLM API abuse (PROMPTFLUX patterns)
Supply Chain	Typosquatting, homoglyphs, hidden files
Tool Misuse	Dangerous tool combinations (Confused Deputy attacks)
Backdoors	Cron jobs, SSH keys, persistent execution

🛡️ Unique Protections

• Typoglycemia Detection: Catches scrambled-word attacks (ignroe → ignore)
• Tool Combination Analysis: Detects browser + exec download-and-execute chains
• Confused Deputy Protection: Prevents ReAct agent injection attacks
• AI Malware Patterns: Identifies PROMPTFLUX-style self-modifying code

📊 Example Output

🔒 Crabukit Security Report
━━━━━━━━━━━━━━━━━━━━━━━━━━━
Skill: malicious-skill
Files scanned: 3

  🔴 CRITICAL   13
  🟠 HIGH        5
  🟡 MEDIUM      6
  ⚪ INFO        1

Risk Level: CRITICAL
Score: 100/100

[CRITICAL] Dangerous tool combination: browser, exec
  Combination enables download-and-execute attack chains
  Fix: Remove unnecessary tools; implement confirmation

[CRITICAL] curl | bash pattern
  Downloads and executes remote code without verification
  Fix: Download to file, verify checksum, then execute

Recommendation: Do not install this skill.

🔌 Clawdex Integration

Crabukit automatically detects and uses Clawdex when installed:

# Install Clawdex for database-based protection
clawdhub install clawdex

Defense in depth:

• Layer 1: Clawdex checks 824+ known malicious skills (instant)
• Layer 2: Crabukit behavior analysis catches zero-days

Example with both scanners:

✓ External scanners: Clawdex

⚪ INFO
  → ✅ Clawdex: Verified safe
    Database reports 'skill-name' as BENIGN

🟡 MEDIUM
  → Destructive operation without warning
    (Crabukit behavior analysis)

🔧 Installation

Via pip (Recommended)

pip install crabukit

As OpenClaw Skill

clawdbot install crabukit

Development

git clone https://github.com/tnbradley/crabukit.git
cd crabukit
pip install -e ".[dev]"

🔒 Safe Install Wrapper (Recommended)

For the safest installation experience, use our wrapper script that combines Clawdex + Crabukit:

# Copy wrapper to your home directory
cp scripts/claw-safe-install.sh ~/.claw-safe-install.sh

# Add to your shell config
echo "source ~/.claw-safe-install.sh" >> ~/.zshrc

# Use it
claw-safe-install youtube-summarize
# or
csi youtube-summarize

Works with or without Clawdex installed. See scripts/README.md for details.

🧪 CI/CD Integration

GitHub Actions

name: Security Scan
on: [push, pull_request]

jobs:
  scan:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Install crabukit
        run: pip install crabukit
      - name: Scan skill
        run: crabukit scan ./my-skill --fail-on=high

Pre-commit Hook

repos:
  - repo: https://github.com/tnbradley/crabukit
    rev: v0.2.0
    hooks:
      - id: crabukit-scan
        args: ['--fail-on=medium']

📚 Research-Based Detection

Crabukit's detection rules are based on:

• OWASP Top 10 for LLM Applications (LLM01-LLM10)
• Lakera AI Q4 2025 Research - Agent attack patterns
• Google Threat Intelligence - PROMPTFLUX/PROMPTSTEAL malware
• WithSecure Research - ReAct Confused Deputy attacks
• arXiv:2410.01677 - Typoglycemia attacks on LLMs

See RESEARCH_SUMMARY.md for detailed references.

🎯 Use Cases

Before Installing a Skill

# Download and scan before installing
clawdbot download some-skill --to ./temp
crabukit scan ./temp/some-skill
# Review results, then decide to install

Auditing Installed Skills

# Scan all installed skills
for skill in /opt/homebrew/lib/node_modules/clawdbot/skills/*/; do
    crabukit scan "$skill" --fail-on=critical || echo "Issues in $skill"
done

CI/CD Security Gate

# Block PRs with critical/high issues
crabukit scan ./my-skill --fail-on=high
# Exit code 1 if issues found

📖 Documentation

• Detection Rules - Full list of security checks
• Contributing - How to contribute
• Security Policy - Reporting vulnerabilities

🤝 Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines.

🛡️ Security

For security issues, please use GitHub's private vulnerability reporting:
https://github.com/tnbradley/crabukit/security/advisories
Or see SECURITY.md for details.

📜 License

MIT License - see LICENSE file.

🙏 Acknowledgments

• OpenClaw community for the skill ecosystem
• OWASP GenAI Security Project
• Researchers at Lakera AI, WithSecure, and Google Threat Intelligence

---

_{Built with 🦀 by @tnbradley}