browser

Files (1)

• SKILL.md

SKILL.md (first lines)

# Browser Skill — arifOS_bot

Triggers: "open browser", "screenshot", "take a screenshot", "scrape page", "get page content",
          "fill form", "navigate to", "render page", "save as pdf", "web automation",
          "click on", "browser automation", "full page", "visit"

Browser: **Browserless Chromium 145** at `http://headless_browser:3000`
Protocol: REST API (POST JSON) + OpenClaw native `browser` tool

---

## Method 1 — OpenClaw Native Browser Tool

Use for interactive navigation, clicking, form-filling.
OpenClaw's `browser` tool is wired to `BROWSERLESS_URL=http://headless_browser:3000`.

```
# In OpenClaw tool call:
browser: navigate to https://example.com
browser: click on "Submit" button
browser: fill input[name=email] with "arif@example.com"
browser: get page text
browser: take screenshot
```

---

## Method 2 — Direct Browserless REST API

Use for precise control, bulk operations, or when native tool isn't enough.

### Screenshot
```bash
# Full page screenshot → save PNG
curl -s -X POST http://headless_browser:3000/screenshot \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://YOURURL.com",
    "options": {
      "fullPage": true,
      "type": "png"
    }
  }' -o ~/.openclaw/workspace/logs/screenshot_$(date +%s).png

echo "Saved to workspace/logs/"
```

### Get Full Page Content (Markdown-friendly HTML)
```bash
curl -s -X POST http://headless_browser:3000/content \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://YOURURL.com",
    "rejectResourceTypes": ["image","font","stylesheet"],
    "waitForSelector": "body"
  }' | head -200
```

### Scrape Specific Elements
```bash
# Extract structured data from page
curl -s -X POST http://headless_browser:3000/scrape \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://YOURURL.com",
    "elements": [
      {"selector": "h1"},
      {"selector": ".price"},
      {"selector": "table tr"},
      {"selector": "article p"}
    ]
  }' | python3 -c "
import sys, json
data = json.load(sys.stdin)
for block in data['data']:
    print(f\"=== {block['selector']} ===\")
    for r in block['results'][:5]:
        print(r['text'])
"
```

### Save Page as PDF
```bash
curl -s -X POST http://headless_browser:3000/pdf \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://YOURURL.com",
    "options": {
      "printBackground": true,
      "format": "A4",
      "margin": {"top":"20px","bottom":"20px","left":"20px","right":"20px"}
    }
  }' -o ~/.openclaw/workspace/logs/page_$(date +%s).pdf
echo "PDF saved"
```

### Execute Custom JavaScript on Page
```bash
curl -s -X POST http://headless_browser:3000/function \
  -H "Content-Type: application/json" \
  -d '{
    "code": "module.exports = async ({ page }) => {
      await page.goto(\"https://YOURURL.com\");
      const title = await page.title();
      const links = await page.$$eval(\"a\", els => els.map(e => ({text: e.textContent.trim(), href: e.href})).filter(l => l.text).slice(0, 20));
      return { title, links };
    }"
  }' | python3 -m json.tool
```

---

## Common Use Cases

### Check if a site is up + get title
```bash
curl -s -X POST http://headless_browser:3000/scrape \
  -H "Content-Type: application/json" \
  -d '{"url":"https://TARGET.com","elements":[{"selector":"title"},{"selector":"h1"}]}' \

From ariffazil/arifOS, AGPL-3.0.

More from ariffazil/arifOS