web-browser
Allows to interact with web pages by performing actions such as clicking buttons, filling out forms, and navigating links. It works by remote controlling Google Chrome or Chromium browsers using the Chrome DevTools Protocol (CDP). When Claude needs to browse the web, it can use this skill to do so.
When & Why to Use This Skill
The Web Browser skill enables Claude to interact directly with live web pages by remote-controlling Google Chrome or Chromium via the Chrome DevTools Protocol (CDP). This powerful integration allows the agent to perform human-like actions such as navigating URLs, clicking buttons, filling out forms, and executing custom JavaScript. By bridging the gap between AI reasoning and active web browsing, it transforms Claude into a functional web agent capable of real-time data retrieval and complex site exploration.
Use Cases
- Automated Web Research: Navigate through multiple websites to gather real-time information, follow citations, and explore deep-link content.
- Web Application Testing: Simulate user workflows by clicking elements, submitting forms, and verifying page states to assist in QA and debugging.
- Dynamic Data Extraction: Execute custom JavaScript (eval) to scrape data from modern, reactive web applications that require client-side rendering.
- Visual Auditing and Monitoring: Capture screenshots of specific viewports to monitor UI changes, verify layouts, or document web-based evidence.
- Interactive Workflow Automation: Use the element picker and navigation tools to automate repetitive tasks like account setups, status checks, or multi-step web processes.
| name | web-browser |
|---|---|
| description | 'Allows to interact with web pages by performing actions such as clicking buttons, filling out forms, and navigating links. It works by remote controlling Google Chrome or Chromium browsers using the Chrome DevTools Protocol (CDP). When Claude needs to browse the web, it can use this skill to do so.' |
| isolatedContext | true |
| license | Stolen from Armin who stole it from someone else |
Web Browser Skill
Minimal CDP tools for collaborative site exploration.
Start Chrome
```bash ./tools/start.js # Fresh profile ./tools/start.js --profile # Copy your profile (cookies, logins) ```
Start Chrome on :9222 with remote debugging.
Navigate
```bash ./tools/nav.js https://example.com ./tools/nav.js https://example.com --new ```
Navigate current tab or open new tab.
Evaluate JavaScript
```bash ./tools/eval.js 'document.title' ./tools/eval.js 'document.querySelectorAll("a").length' ./tools/eval.js 'JSON.stringify(Array.from(document.querySelectorAll("a")).map(a => ({ text: a.textContent.trim(), href: a.href })).filter(link => !link.href.startsWith("https://")))' ```
Execute JavaScript in active tab (async context). Be careful with string escaping, best to use single quotes.
Screenshot
```bash ./tools/screenshot.js ```
Screenshot current viewport, returns temp file path
Pick Elements
```bash ./tools/pick.js "Click the submit button" ```
Interactive element picker. Click to select, Cmd/Ctrl+Click for multi-select, Enter to finish.