host-computer-use-windows

Guides Windows UI Automation targeting for remote computer-use sessions.

19.1k|3.8k|Updated Jun 10, 2024
One-click install
npx skills add https://github.com/agent0ai/agent-zero --skill host-computer-use-windows
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: host-computer-use-windows
Source: https://github.com/agent0ai/agent-zero/tree/main/plugins/_a0_connector/skills/host-computer-use-windows
Command: npx skills add https://github.com/agent0ai/agent-zero --skill host-computer-use-windows

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automating a remote Windows desktop through pixel clicks is fragile and error-prone. This Skill provides backend-specific guidance for driving Windows applications through UI Automation structural targeting, so agents can reliably interact with buttons, text fields, dialogs, and windows instead of guessing screen coordinates.

Core Features & Use Cases

  • UIA Structural Targeting: Use uia_snapshot, uia_action, and element_action to locate and operate named controls by role, title, automation ID, or selector rather than pixel coordinates.
  • Background Dispatch: Prefer background element actions over foreground clicks, with a defined fallback order when background dispatch is unavailable.
  • Session Failure Handling: Detect Windows session problems such as locked screens, disconnected RDP, UAC prompts, or capture/UIA unavailability and stop to ask the user to re-arm instead of bypassing with unsafe fallbacks.
  • Use Case: An agent connected to a Windows host via A0 CLI needs to click OK in a dialog. It lists windows, gets the window state, invokes the button by element index in the background, then verifies the result from the attached screenshot.

Quick Start

Connect to the Windows host and use UI Automation targeting to click the OK button in the open dialog, then verify the result from the screenshot.

Frequently Asked Questions about host-computer-use-windows

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate Windows GUI applications remotely?

Use UI Automation structural targeting instead of pixel clicks. Take a uia_snapshot or get_window_state to find elements, then invoke actions like invoke, set_value, or focus by element index or a role-plus-title target, verifying results from the attached screenshot.

How to click a button in Windows using UI Automation?

Prefer the invoke operation over click when the snapshot lists it as an available action on the node. Target the button by role and title, automation ID, or element index, and use background dispatch when the backend supports it.

When should I use element_action instead of uia_action?

Prefer element_action when the backend advertises element-index-targeting, because it preserves the background-first dispatch contract. Use uia_action with semantic targets or selectors when working from a uia_snapshot tree.

Why does Windows desktop automation fail over Remote Desktop?

Capture and UI Automation depend on the interactive desktop session. Minimized or disconnected RDP sessions, locked screens, UAC prompts, and elevated apps can block access, returning errors like COMPUTER_USE_CAPTURE_UNAVAILABLE that require the user to re-arm the session.

How do multi-monitor coordinates work in Windows screen capture?

Windows captures can span a virtual desktop with multiple monitors and negative origins. Use the session's origin_x, origin_y, width, and height as the coordinate space, and interpret normalized coordinates relative to that virtual screen.