Step by step documentation¶

Step 1 Install¶

We're ust going to follow https://docs.openclaw.ai/install.

I skipped everything

Configure gateway¶

We're running this in a VM and want to open up to the lan.

You're gateway should look something like

"gateway": {
    "port": 18789,
    "mode": "local",
    "bind": "lan",
    "auth": {
      "mode": "token",
      "token": "[redacted]"
    },
    "trustedProxies": ["192.168.5.8"],
    "controlUi": {
      "allowedOrigins": ["https://claw.domain.com"]
    },
    "tailscale": {
      "mode": "off",
      "resetOnExit": false
    },
    "nodes": {
      "denyCommands": [
        "camera.snap",
        "camera.clip",
        "screen.record",
        "contacts.add",
        "calendar.add",
        "reminders.add",
        "sms.send"
      ]
    }
  },

Go ahead and navigate to your website, paste in your token.

Then run openclaw devices list. Find the device id for approval and run openclaw devices approve <id>.

Browser¶

Let's go ahead and get the browser working.

I installed playwright with

npm install playwright

Then you need a browser engine

npx playwright install chromium
npx playwright install-deps chromium

then configure the browser block

"browser": {
    "enabled": true,
    "ssrfPolicy": {
      "dangerouslyAllowPrivateNetwork": true
    },
    "remoteCdpTimeoutMs": 1500, 
    "remoteCdpHandshakeTimeoutMs": 3000,
    "defaultProfile": "openclaw",
    "color": "#FF4500",
    "headless": true,
    "noSandbox": true,
    "attachOnly": false,
    "executablePath": "/home/thomas/.cache/ms-playwright/chromium-1208/chrome-linux64/chrome",
    "profiles": {
      "openclaw": { "cdpPort": 18800, "color": "#FF4500" },
      "work": { "cdpPort": 18801, "color": "#0066CC" },
      "user": {
        "driver": "existing-session",
        "attachOnly": true,
        "color": "#00AA00"
      },
      "remote": { "cdpUrl": "http://10.0.0.42:9222", "color": "#00AA00" }
    }
  },

Test out everything works

openclaw browser --browser-profile openclaw start
openclaw browser --browser-profile openclaw open https://example.com
openclaw browser --browser-profile openclaw screenshot --full-page

Models¶

Give it some models

Sandbox¶

Built openclaw-sandbox-common:bookworm-slim.
To use it, set agents.defaults.sandbox.docker.image to "openclaw-sandbox-common:bookworm-slim" and restart.
If you want a clean re-create, remove old sandbox containers:
  docker rm -f $(docker ps -aq --filter label=openclaw.sandbox=1)

If you run into an error rebuilding:

docker builder prune -f

Browser¶

I had trouble getting the browser to work inside of a sandbox. But I should look at this more closely (https://docs.openclaw.ai/gateway/sandboxing#what-gets-sandboxed). Sounds like I could also just use the hosts browser.

Instead I simply baked it into the Dockerfile.sandbox-common

# Install Playwright with Chromium and all system dependencies baked in
ENV PLAYWRIGHT_BROWSERS_PATH=/opt/ms-playwright
ENV NODE_PATH=/usr/local/lib/node_modules
RUN npm install -g playwright \
  && playwright install chromium --with-deps \
  && chmod -R o+rx /opt/ms-playwright

And then updated his TOOLS.md

Web Browser¶

Use: Playwright via exec — Chromium and Playwright are installed in your sandbox container
When: Whenever asked to screenshot a website, navigate through it, or interact with web UIs
How: Write a Node.js script and run it with exec. Always use --no-sandbox and --disable-dev-shm-usage launch args (required in containers).

Example — take a screenshot:

const { chromium } = require('playwright');
(async () => {
  const browser = await chromium.launch({
    headless: true,
    args: ['--no-sandbox', '--disable-dev-shm-usage']
  });
  const page = await browser.newPage();
  await page.goto('https://example.com');
  await page.screenshot({ path: '/workspace/screenshot.png', fullPage: true });
  await browser.close();
  console.log('done');
})();

```

Heatbeat¶

Read https://docs.openclaw.ai/gateway/heartbeat

openclaw system heartbeat

Make sure it's enabled

openclaw system heartbeat enable

To manually trigger a heartbeat

openclaw system event --text "Perform a heartbeat" --mode now

mcporter¶

Install mcporter

npm install -g mcporter

After installing you can see example files in ~/.npm-global/lib/node_modules/mcporter/config/mcporter.json

For global config you can add a config file in /home/<user>/.mcporter/mcporter.json or navigate to an agents workspace and create a ./config/mcporter

From the workspace if you run the following, it will show you the active configs mcporter will use when running from this directory.

mcporter config list

This will show you the project config and system config.

To list mcp servers

mcp list

To see available tools for a server

mcp youtrack tools

name a tool to see its api docs

mcp youtrack tools projects

Execute a tool

mcporter call youtrack.issues action="query" query="project: BS" fields="id,summary,state,assignee"

Copilot CLI¶

Refer to https://github.com/github/copilot-cli for installation and basic usage

Info on custom agents: https://docs.github.com/en/copilot/how-tos/copilot-cli/customize-copilot/create-custom-agents-for-cli

Interesting looking repo for remote control: https://github.com/devartifex/copilot-unleashed?tab=readme-ov-file

Step by step documentation¶

Step 1 Install¶

Configure gateway¶

Browser¶

Models¶

Sandbox¶

Browser¶

Web Browser¶

Heatbeat¶

mcporter¶

Copilot CLI¶

Comments