windows-rpa

name	windows-rpa
version	1.1.0
description	Windows 桌面 RPA 自动化控制。使用场景： (1) 操作桌面应用程序（非 Web 应用） (2) 模拟鼠标键盘操作 (3) 截图和图像识别 (4) 窗口管理和应用启动 (5) 剪贴板操作 (6) Shell 命令执行 NOT for: Web 应用操作（请使用 browser 工具）、Linux/Mac 系统
os	["win32","windows"]
security	{"permissions":["screen_capture","keyboard_input","mouse_input","clipboard_access","shell_execution"],"user_approval":"recommended","sandbox":"supported","sensitive_operations":["desktop_shell","desktop_clipboard_get","desktop_screenshot"]}
metadata	{"openclaw":{"emoji":"🖥️","os":["win32"],"requires":{"bins":["python"],"anyBins":["pyautogui","pywinauto"]},"install":[{"id":"pip-pyautogui","kind":"pip","package":"pyautogui pillow","bins":["pyautogui"],"label":"Install PyAutoGUI for mouse/keyboard automation"},{"id":"pip-pywinauto","kind":"pip","package":"pywinauto","bins":["pywinauto"],"label":"Install PyWinAuto for Windows UI Automation"},{"id":"pip-pyperclip","kind":"pip","package":"pyperclip","label":"Install Pyperclip for clipboard operations"}]}}
implementation	{"type":"python","entry":"scripts/rpa.py"}

Windows RPA 自动化

这个 skill 让 OpenClaw 能够直接操作 Windows 桌面，包括鼠标、键盘、窗口和应用程序。

核心能力

1. 鼠标操作

# 移动鼠标到坐标 (x, y)
desktop_mouse_move(x=500, y=300)

# 鼠标点击
desktop_mouse_click(x=500, y=300, button="left", clicks=1)

# 鼠标拖拽
desktop_mouse_drag(start_x=100, start_y=100, end_x=500, end_y=500)

# 鼠标滚轮
desktop_mouse_scroll(clicks=3)  # 正数向上，负数向下

# 获取鼠标位置
desktop_mouse_position()

2. 键盘操作

# 输入文本
desktop_keyboard_type(text="Hello World")

# 按键
desktop_keyboard_press(key="enter")
desktop_keyboard_press(key="tab")
desktop_keyboard_press(key="escape")

# 组合键
desktop_keyboard_hotkey(keys=["ctrl", "c"])  # 复制
desktop_keyboard_hotkey(keys=["ctrl", "v"])  # 粘贴
desktop_keyboard_hotkey(keys=["ctrl", "a"])  # 全选
desktop_keyboard_hotkey(keys=["alt", "f4"])  # 关闭窗口

3. 屏幕操作

# 截取全屏
desktop_screenshot()

# 截取指定区域
desktop_screenshot(region={"x": 0, "y": 0, "width": 800, "height": 600})

# 获取屏幕尺寸
desktop_screen_size()

# 图像定位（在屏幕上找图）
desktop_locate_on_screen(image_path="button.png", confidence=0.9)

4. 窗口管理

# 列出所有窗口
desktop_window_list()

# 激活窗口
desktop_window_activate(title_pattern="记事本")

# 查找窗口
desktop_find_window(title_contains="Chrome")

# 点击窗口控件
desktop_click_window(
    title_contains="记事本",
    control_type="Edit",
    action="set_text",
    control_name="Hello"
)

5. 应用程序启动

# 启动常用应用（支持别名）
desktop_launch_app(app_name="notepad")
desktop_launch_app(app_name="chrome")
desktop_launch_app(app_name="excel")
desktop_launch_app(app_name="word")

# 启动自定义程序
desktop_launch_app(app_name="C:\\Program Files\\MyApp\\app.exe")

# 带参数启动
desktop_launch_app(app_name="chrome", args="--incognito")

6. Shell 命令

# PowerShell 命令
desktop_shell(command="Get-Process | Select-Object -First 5")

# CMD 命令
desktop_shell(command="dir C:\\", shell_type="cmd")

7. 剪贴板

# 获取剪贴板内容
desktop_clipboard_get()

# 设置剪贴板内容
desktop_clipboard_set(text="要复制的内容")

8. 桌面状态

# 获取完整桌面状态（鼠标位置、活动窗口、截图等）
desktop_get_state(capture_screenshot=True)

# 检查环境
desktop_check()

使用示例

示例 1: 自动填写表单

# 1. 启动应用程序
desktop_launch_app(app_name="notepad")

# 2. 等待窗口出现
time.sleep(1)

# 3. 输入内容
desktop_keyboard_type(text="这是一段自动输入的文字")

# 4. 保存文件
desktop_keyboard_hotkey(keys=["ctrl", "s"])
desktop_keyboard_type(text="auto_saved.txt")
desktop_keyboard_press(key="enter")

示例 2: 图像识别点击

# 1. 先截图保存按钮图像
desktop_screenshot(save_path="screen.png")

# 2. 用户手动截取按钮区域保存为 button.png

# 3. 在屏幕上定位并点击
result = desktop_locate_on_screen(image_path="button.png", confidence=0.9)
# 如果找到，会返回中心坐标，然后点击