Add Agent to OpenClaw#

Workflow#

Step 1: Collect Agent Info#

Required info from user:

• Bot token: Telegram bot token (format: 123456789:ABCdef...)
• Agent ID: Unique identifier (e.g., translator, customer_support)
• Purpose/Name: Display name (optional, defaults to agent ID)
• Allow users: Telegram user IDs that can access this bot (default: same as existing agents)

Step 2: Update openclaw.json#

Edit /home/xgao/.openclaw/openclaw.json:

A. Add agent to agents.list:

json

{
  "id": "<agent_id>",
  "name": "<name>",
  "workspace": "/home/xgao/clawd-workspace/<agent_id>"
}

B. Add binding:

json

{
  "agentId": "<agent_id>",
  "match": {
    "channel": "telegram",
    "accountId": "<agent_id>"
  }
}

C. Add telegram account:

json

"<agent_id>": {
  "enabled": true,
  "dmPolicy": "allowlist",
  "botToken": "<token>",
  "allowFrom": [<user_ids>],
  "groupPolicy": "allowlist",
  "streaming": "partial"
}

Step 3: Create Workspace Directory#

bash

mkdir -p /home/xgao/clawd-workspace/<agent_id>

Step 4: Ask About Memory Isolation#

Ask user: "是否需要独立的memory？"

• Yes → Each agent already has independent workspace, memory is automatically isolated
• No → Share workspace (not recommended)

Step 5: Restart OpenClaw#

Inform user to restart: openclaw restart

Minimal Config Example#

For a simple translator bot:

json

// agents.list
{ "id": "translator", "name": "translator", "workspace": "/home/xgao/clawd-workspace/translator" }

// bindings
{ "agentId": "translator", "match": { "channel": "telegram", "accountId": "translator" } }

// channels.telegram.accounts
"translator": { "enabled": true, "dmPolicy": "allowlist", "botToken": "8223784004:AAE-...", "allowFrom": ["8538882690"], "groupPolicy": "allowlist", "streaming": "partial" }