tribe-protocol
# Tribe Protocol
Trust lookup system for OpenClaw bots. Every non-owner interaction must be verified against the tribe database before responding.
## Quick Start
```bash
# 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
1. Message arrives from non-owner
2. Bot reads TRIBE.md (generated at workspace root)
3. Bot runs `tribe lookup <discord_id>`
4. Script returns entity info + tier rules reminder
5. Bot applies tier-appropriate behavior
## Commands
- `tribe init` — Initialize database
- `tribe lookup` — Query entity by discord_id, name, tag, server
- `tribe add` — Add new entity
- `tribe set-tier` — Update trust tier
- `tribe set-status` — Update status
- `tribe grant` / `tribe revoke` — Channel access
- `tribe tag` — Manage tags
- `tribe roster` — List all entities
- `tribe log` — Audit trail
- `tribe export` — Dump to markdown
- `tribe stats` — Quick summary
## Environment Variables
- `TRIBE_DB` — Override database path
- `CLAWD_HOME` — Base directory (default: ~/clawd)
## Dependencies
- `sqlite3` (pre-installed on macOS/most Linux)
标签
skill
ai