computer-use-skill

Remotely execute exact browser automation tasks by forwarding user commands directly to the Computer Use Agent (CUA) via its Python SDK.

Note: This skill operates in pass-through mode - tasks are transmitted directly to CUA without modification.

Architecture

CUA Integration Path

text

User Task → Python SDK → CUA Sandbox → Browser Automation → Results Stream

Key Components

• CUA SDK: Official Python SDK for browser automation
• Virtual Environment: Isolated environment with SDK installed
• Python Wrapper: Programmatic access for integration

File Structure

text

workspace/
├── tools/
│   └── execute_cua_task.py      # Task execution script
└── cua_venv/                    # Virtual environment with SDK

Quick Reference

Task Type	Example
Search	"打开google页面，搜索杭州天气"
Navigation	"访问github.com"
Forms	"在登录页面输入用户名和密码"
Screenshots	"访问产品页面并截图"
Scraping	"从电商页面提取价格信息"
Complex	"将购物车商品结算并截图确认"

Requirements

• No external credentials handled by user
• Virtual environment with CUA Python SDK installed

Usage Patterns

text

with run source to active cua_venv → run cd to directory tools → run python execute_cua_task.py like 'python execute_cua_task.py "在百度搜索人工智能"'

Pass-Through Protocol

Critical Rules

1. No task modification: Tasks are transmitted exactly as provided
2. No wrapping or packaging: Raw task text sent to CUA
3. No interpretation: CUA AI handles task understanding
4. No preprocessing: User text → CUA (direct path)

Workflow

text

User: "打开google页面，搜索杭州天气"
↓
Skill: Transmit "打开google页面，搜索杭州天气" (exact match)
↓
CUA: AI understands and executes browser automation
↓
Results: Stream of execution messages returned

What NOT to Do

• Don't wrap in additional instructions
• Don't process or modify user input
• Don't simplify or elaborate user task