Openclaw Config Pack — Security skill for Claude Code

openclaw-config-pack

Hardened configuration pack for [OpenClaw](https://openclaw.ai) personal AI assistant.

Includes everything needed to deploy OpenClaw with full security configuration: 8 Telegram topic templates, 6 cognitive skills, hardened Docker/systemd deployment, and a 9-layer defense architecture.

---

9-Layer Security Architecture

Layer 1  │ Cloudflare WAF/DDoS        │ Network perimeter
Layer 2  │ UFW Firewall               │ Host firewall
Layer 3  │ Loopback-only binding      │ Gateway isolation
Layer 4  │ Telegram allowlist         │ User authentication
Layer 5  │ Input filter regex         │ Pattern-based rejection
Layer 6  │ CIF + prompt-guard         │ Cognitive injection defense
Layer 7  │ SOUL.md CRITICAL rules     │ Agent identity constraints
Layer 8  │ Docker sandbox             │ Execution isolation
Layer 9  │ Systemd hardening          │ Process isolation

---

Quick Start

1. Clone

git clone https://github.com/itsnex1s/openclaw-config-pack.git
cd openclaw-config-pack

2. Run installer

chmod +x install.sh
./install.sh

The installer will:

• Generate a gateway password and token
• Copy workspace files (SOUL.md, SECURITY.md, skills)
• Create .env from template
• Set file permissions (chmod 600/700)
• Optionally install the systemd service

3. Edit configuration

nano ~/.openclaw/config/openclaw.json

Replace all `YOUR_*` placeholders:

• YOUR_TELEGRAM_ID — your Telegram user ID
• YOUR_GROUP_ID — your Telegram group ID
• Topic IDs — update to match your group's forum topics

4. Add credentials

nano ~/.openclaw/credentials/.env

Fill in your API keys and tokens.

5. Start

# With Docker:
cd ~/.openclaw && docker compose up -d

# Or with systemd:
sudo systemctl start openclaw

6. Voice transcription (optional)

To enable automatic voice message transcription, see [docs/VOICE-TRANSCRIPTION.md](docs/VOICE-TRANSCRIPTION.md). Quick version:

`