Security Settings

Clawdbot provides multi-layer security mechanisms to protect your AI assistant.

Security Audit

Run security audit regularly:

clawdbot security audit

Access Control

DM Policies

• pairing: Unknown users need pairing code (recommended)
• allowlist: Only allow listed users
• open: Allow all users

Group Policies

• Default requires @mention
• Allowlist configurable

Configuration Hardening

chmod 600 ~/.clawdbot/clawdbot.json
chmod 700 ~/.clawdbot

Sandboxing

Two sandbox approaches:

• host: Run directly on host
• docker: Docker container isolation

Notes

• There is no "perfectly secure" setup
• Use the latest AI models
• Audit configuration regularly