ClawKit Logo
ClawKitReliability Toolkit
Back to Registry
Official Verified browser automation Safety 3/5

Browser Use Local

Skill by fengjiajie

Why use this skill?

Automate web tasks, scrape dynamic sites, and run local AI agents with Browser Use Local. Perfect for complex workflows, debugging, and custom LLM integration.

skill-install — Terminal

Install via CLI (Recommended)

clawhub install openclaw/skills/skills/fengjiajie/browser-use-local
Or

What This Skill Does

The Browser Use Local skill, developed by fengjiajie, provides a robust, local-first browser automation environment within OpenClaw. It serves as a powerful alternative to standard browser tools by offering direct control over browser-use CLI and Python interfaces. Designed for high-reliability automation, it allows users to open pages, perform interactions (clicks, typing), take screenshots for visual debugging, extract HTML, and execute complex agent workflows using custom OpenAI-compatible LLMs like Moonshot/Kimi. It is particularly effective for navigating JavaScript-heavy sites where standard scrapers may fail, and it provides specific utilities for extracting login data and QR codes.

Installation

To install this skill, use the OpenClaw CLI command: clawhub install openclaw/skills/skills/fengjiajie/browser-use-local

Ensure that your environment has sufficient permissions for browser automation and that the virtual environment at /home/node/.openclaw/workspace/.venv-browser-use/ is correctly configured if you intend to run custom agent scripts.

Use Cases

  • Complex Web Automation: Handling multi-step workflows like logging into dashboards or filling out long forms that require state persistence using --session.
  • Data Extraction: Scraping data from dynamic, JS-rendered sites where standard 'state' queries return empty, by falling back to raw HTML parsing or lightweight DOM evaluation.
  • Debugging: Capturing screenshots during failed runs or headless operations to visualize why a flow stopped, especially when selectors are changing.
  • Custom LLM Agents: Executing autonomous agent tasks using private API keys and base_urls (e.g., Kimi/Moonshot), useful for enterprise-grade automation where cloud-managed browser keys are restricted.
  • Authentication Bypass: Automating the retrieval and processing of QR codes for login flows that involve scanning services.

Example Prompts

  1. "Open https://auth.example.com using session 'login-test', take a screenshot, and extract all URLs containing 'qr' from the current HTML."
  2. "Run the local Kimi agent script to navigate to my company dashboard and download the latest report using the browser-use-local skill."
  3. "Debug the current session; the page isn't loading elements correctly. Run a JavaScript evaluation to get the document.title and check the current URL."

Tips & Limitations

  • Persistence: Always use --session <name> for multi-step tasks to avoid re-authenticating or resetting the browser state.
  • HTML Fallback: If state returns 0 elements, don't assume the page is empty; use the get html command or custom JS eval to verify the DOM.
  • Kimi/Moonshot Tuning: When using the Python agent with Moonshot, ensure temperature is set to 1 and frequency_penalty to 0. If schema validation fails, enable remove_defaults_from_schema and remove_min_items_from_schema to prevent JSON schema errors.
  • Environment Stability: This tool operates within a containerized environment; if browser-use commands hang, ensure the browser binary is reachable and hardware acceleration requirements are met for the environment.

Metadata

Stars2387
Views0
Updated2026-03-09
View Author Profile
AI Skill Finder

Not sure this is the right skill?

Describe what you want to build — we'll match you to the best skill from 16,000+ options.

Find the right skill
Add to Configuration

Paste this into your clawhub.json to enable this plugin.

{
  "plugins": {
    "official-fengjiajie-browser-use-local": {
      "enabled": true,
      "auto_update": true
    }
  }
}

Tags(AI)

#browser-automation#web-scraping#automation#kimi#headless
Safety Score: 3/5

Flags: network-access, file-write, file-read, external-api, code-execution