dvcrn Tribe Protocol
Trust lookup system for OpenClaw bots. Every non-owner interaction must be verified against the tribe database before responding.
Quick Start
# Initialize (first time only)
./scripts/tribe.sh init \
--bot-name Cheenu \
--bot-discord-id 000000000000000004 \
--human-name Nagarjun \
--human-discord-id 000000000000000002
# Look up an entity before responding
./scripts/tribe.sh lookup <discord_id>
# Add entities
./scripts/tribe.sh add --name Yajat --type human --discord-id 000000000000000001 --tier 3
# Manage trust
./scripts/tribe.sh set-tier <discord_id> 3 --reason "Promoted to tribe"
./scripts/tribe.sh set-status <discord_id> blocked --reason "Bad actor"
Trust Tiers
| Tier | Label | Access |
|---|---|---|
| 4 | Owner | Full trust, all data |
| 3 | Tribe | Collaborate freely, no private data |
| 2 | Acquaintance | Public info only |
| 1 | Stranger | Minimal engagement |
| 0 | Blocked | Ignore completely |
How It Works
- Message arrives from non-owner
- Bot reads TRIBE.md (generated at workspace root)
- Bot runs
tribe lookup <discord_id> - Script returns entity info + tier rules reminder
- Bot applies tier-appropriate behavior
Commands
tribe init— Initialize databasetribe lookup— Query entity by discord_id, name, tag, servertribe add— Add new entitytribe set-tier— Update trust tiertribe set-status— Update statustribe grant/tribe revoke— Channel accesstribe tag— Manage tagstribe roster— List all entitiestribe log— Audit trailtribe export— Dump to markdowntribe stats— Quick summary
Environment Variables
TRIBE_DB— Override database pathCLAWD_HOME— Base directory (default: ~/clawd)
Dependencies
sqlite3(pre-installed on macOS/most Linux)