OpenClaw (formerly Clawdbot/Moltbot) is a powerful, self-hosted AI agent that runs on your local machine. By integrating it with Dart, you can give your OpenClaw bot a dedicated "identity" in your workspace. This allows it to autonomously create tasks, update statuses, and manage documentation using the Dart Python library and CLI.

Install OpenClaw

If you haven't already set up your autonomous agent, you'll need to install OpenClaw. The recommended method is via the official installation script or Docker.

Follow the official OpenClaw installation guide here.

Create a custom agent in Dart

First, give your bot a face and a name in your workspace. This ensures all actions (like creating tasks) are attributed to the bot, not you.

Go to Settings > Agents in your Dart workspace
Click Create custom agent
Name your agent (e.g., "OpenClaw Bot", "AutoPM") and upload an avatar
Save the agent

Get the agent's API Key

Now you need the credentials that will allow OpenClaw to log in as this agent.

In the Settings > Agents list, find your new agent
Click the (...) menu next to the agent's name
Select Create authentication token
Copy this token immediately. You will need to provide this to OpenClaw in the next step

Install the Dart library for OpenClaw

OpenClaw needs the Dart tools installed in its environment to execute commands.

Access the terminal where OpenClaw is running (or the container shell if using Docker) and install the library:

pip install dart-tools

Configure OpenClaw

Finally, set up OpenClaw to use your key. You can do this by setting an environment variable or a skill in OpenClaw.

Option A: Quick CLI setup (environment variable)

The easiest way to start is to export your API token as an environment variable in OpenClaw's configuration or .env file:

export DART_API_TOKEN="your_agent_token_here"

Now, OpenClaw can run authenticated commands directly via the CLI. Try sending this message to your bot:

"Run the dart CLI to create a new task titled 'Review Q3 Metrics' in the Marketing project."

Option B: Create a dedicated skill (recommended)

For a more robust setup, create a custom Skill file (e.g., dart-skill.md) in your OpenClaw skills directory. This prompts the agent on how to use the library effectively.

You must provide the API Token as an environment variable so the Skill can access it.

Open your OpenClaw config file: ~/.openclaw/openclaw.json
Add your token to the env section (if it does not exist create one)

{ 
"env": { "DART_API_TOKEN": "PASTE_YOUR_ACTUAL_TOKEN_HERE" }, 
"agents": { // ... your other settings } 
}

Example skill prompt:

--- name: Dart Project Manager 
emoji: 🎯
description: Manage Dart projects, tasks, and documentation using the Dart CLI. 
requires:   python: ">=3.10"   pip: ["dart-tools"] 
env  
  - DART_API_TOKEN
---

# Dart Skill  You have access to the `dart-tools` Python library and CLI. Use this to manage the user's project board.

## Capabilities 
- **Create Tasks:** `dart create "Title" --description "Details"` 
- **List Tasks:** `dart list --status "In Progress"`
- **Update Docs:** Use the `dart_tools` python library to push updates to documents.

## Rules 
1. Always run `dart list` before creating a duplicate task. 
2. If the user mentions "bugs", check for tasks labeled "bug" first.
3. If an API error occurs, report the specific error code to the user.

AI planning

Import tasks

API, Libraries, and CLI

Dart MCP

Dart in Terminal and Python integration