Safe Share

🛡️ Pre-Publish Security Gate

A security-first content sanitization system that scans, classifies, and redacts sensitive information before public sharing.

English | 简体中文

✨ Features

• 🔍 Smart Detection - Rule-based engine with 40+ patterns for sensitive information
• 🎯 Risk Classification - Three-tier system (P0/P1/P2) for graded response
• 🎨 Pixelation Masking - Mosaic-style masking for effective redaction
• 📊 Detailed Reports - Markdown and JSON outputs for audit trails
• 🚀 Fast Processing - ~1000 lines/second for text scanning
• 🛡️ Privacy First - All processing done locally, no external requests

🎯 Use Cases

Perfect for:

• 📸 Screenshots - Code, terminal, chat logs before sharing
• 📄 Documents - Logs, configs, documentation before publishing
• 💬 Messages - Forum posts, emails, chat messages
• 🔐 Security Reviews - Pre-publication security audits

🚀 Quick Start

Installation

# Clone the repository
git clone https://github.com/ijerrywong/safe-share.git
cd safe-share

# Install dependencies
pip install -r requirements.txt

# Install tesseract OCR (required for image scanning)
# macOS
brew install tesseract

# Ubuntu/Debian
sudo apt-get install tesseract-ocr

# Windows
# Download from https://github.com/UB-Mannheim/tesseract/wiki

Usage

Option 1: Command Line Tool

```bash
# Scan a text file
python3 main.py --file config.txt

# Scan an image
python3 main.py --file screenshot.png

# Verbose mode
python3 main.py --file screenshot.png --verbose

# Different security modes
python3 main.py --file screenshot.png --mode strict     # Maximum protection
python3 main.py --file screenshot.png --mode balanced   # Balance
python3 main.py --file screenshot.png --mode relaxed    # Minimal

Option 2: As OpenClaw Skill

To use as an OpenClaw skill, copy the skill file to your skills directory:

# Find your OpenClaw skills directory (usually ~/.openclaw/skills/)
# Create the directory if it doesn't exist
mkdir -p ~/.openclaw/skills/

# Copy the skill file
cp SKILL.md ~/.openclaw/skills/safe-share.md

# Or for Chinese version
cp SKILL_CN.md ~/.openclaw/skills/safe-share-cn.md

Then you can use it directly in OpenClaw:

Scan this file for sensitive information before sharing

Or:

Use safe-share to check this screenshot

Output Files

After scanning, Safe Share generates:

outputs/
├── redacted_<filename>      # Masked content (image or text)
├── risk_report.md            # Human-readable audit report
└── findings.json             # Machine-readable detection data

📋 What Gets Detected

🚨 P0 - Critical (Always Masked)

• API keys & tokens
• Database passwords
• Personal data (phone, email, SSN)
• Private keys & certificates

⚠️ P1 - High Risk (Masked by Default)

• File paths (.env, .ssh, secrets/)
• Local services (127.0.0.1:5000)
• Database usernames
• Infrastructure URLs

ℹ️ P2 - Medium Risk (Warning Only)

• Project names
• Device information
• Business context

See RULES.md for complete pattern list.

🎨 Masking Styles

Style	Description	Use Case
pixelate	Mosaic/blur blocks	Default, recommended
black	Solid black bars	High contrast needed
red	Solid red bars	Warning emphasis
blur	Gaussian blur	Softer redaction

📖 Documentation

• Installation Guide
• Configuration
• Pattern Rules
• API Reference
• Contributing

🔧 Development

# Create virtual environment
python3 -m venv venv
source venv/bin/activate

# Install development dependencies
pip install -r requirements-dev.txt

# Run tests
pytest tests/

# Run linting
black utils/ main.py

🤝 Contributing

Contributions are welcome! Please read CONTRIBUTING.md for details.

📊 Performance

Input Type	Speed	Memory
Text (1000 lines)	~1s	<50MB
Image (1920x1080)	~6s	<100MB

⚠️ Disclaimer

Safe Share uses automated pattern matching. While it catches many common patterns, it may have:

• False positives (safe content flagged as sensitive)
• False negatives (sensitive content not detected)

Always review the output manually before sharing content.

📜 License

MIT License - see LICENSE for details.

🙏 Acknowledgments

• Built with Pillow for image processing
• Uses tesseract-ocr for OCR
• Pattern matching powered by PyYAML

📮 联系方式

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Email: ijerrywong@gmail.com

---

⭐ Star us on GitHub!

Made with ❤️ by JerryWong