@datplatform/openclaw-dat-trust

v1.0.0

Published

4 months ago

Decentralized identity and trust scoring for OpenClaw agents via the DAT platform

0High
0Medium
0Low

signacloud

openclaw openclaw-plugin dat trust identity did reputation ai-agent decentralized-identity signacloud erc-8004

DAT Trust Plugin for OpenClaw

Decentralized identity and trust scoring for OpenClaw agents via the DAT platform.

What It Does

Auto-registers your OpenClaw agent with DAT on startup (assigns a DID)
Reports trust signals automatically based on tool call outcomes
Provides tools to check trust scores and verify other agents
Background service keeps your trust score synced
Slash command /trust for quick status checks

Installation

From npm

npm install @datplatform/openclaw-dat-trust

From ClawHub

clawhub install dat-trust

From local path

openclaw plugins install ./skills/openclaw-dat-trust

Configuration

Add to your openclaw.json:

{
  "plugins": {
    "entries": {
      "dat-trust": {
        "enabled": true,
        "config": {
          "agentName": "my-agent",
          "agentDescription": "My OpenClaw agent with DAT trust",
          "network": "testnet"
        }
      }
    }
  }
}

Config Options

| Option | Type | Default | Description | |--------|------|---------|-------------| | datApiUrl | string | https://www.signacloud.io | DAT platform API URL | | agentName | string | auto-generated | Agent name on DAT | | agentDescription | string | auto-generated | Agent description | | orgDid | string | — | Organization DID to associate with | | network | mainnet | testnet | testnet | DAT network | | autoReport | boolean | true | Auto-report tool call outcomes | | scoreRefreshInterval | number | 60000 | Score refresh interval (ms) |

Tools

| Tool | Description | |------|-------------| | dat_trust_check | Check any agent's trust score by DID | | dat_my_trust | Check your own trust score and identity | | dat_verify_agent | Verify an agent meets a minimum trust threshold | | dat_report_event | Manually report a trust event |

Commands

| Command | Description | |---------|-------------| | /trust | Show your DID and trust score | | /trust did:dat:testnet:agent_xxx | Check another agent's score |

How Trust Scoring Works

DAT uses a 4-pillar trust model:

Reliability — Consistency of successful operations
Performance — Response times and throughput
Compliance — Adherence to policies and standards
Security — Safety of agent behavior

Your trust score (0-100) is computed from these pillars. Every tool call success adds to reliability; failures subtract. The plugin reports these automatically when autoReport is enabled.

Architecture

OpenClaw Agent
  └── dat-trust plugin
        ├── Service (background)
        │     ├── Startup: register → init reputation → self-verify
        │     └── Loop: refresh trust score every 60s
        ├── Tools (agent-callable)
        │     ├── dat_trust_check
        │     ├── dat_my_trust
        │     ├── dat_verify_agent
        │     └── dat_report_event
        ├── Hooks (automatic)
        │     ├── before_tool_call → track start time
        │     ├── after_tool_call → report success/failure
        │     └── message_received → audit trail
        └── Command
              └── /trust

License

MIT