computer-use

Inspect and control local desktop app windows through accessibility trees and safe UI actions.

366|90|Updated Aug 22, 2026
One-click install
npx skills add https://github.com/bam-bam-2/solo-skills --skill computer-use-bam-bam-2
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: computer-use
Source: https://github.com/bam-bam-2/solo-skills/tree/main/skills/computer-use
Command: npx skills add https://github.com/bam-bam-2/solo-skills --skill computer-use-bam-bam-2

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Automating or reading desktop applications normally requires brittle screen scraping or manual clicking. This Skill lets an agent inspect any local desktop app window through its accessibility tree, take screenshots, and perform safe UI actions like clicking, typing, scrolling, and dragging through Orca's computer-use CLI. ## Core Features & Use Cases - Window and app inspection: List running apps and windows, get app state, and read visible UI elements via accessibility trees. - Safe UI automation: Click controls, type text, press keys, scroll, drag, and set values in desktop apps, browser windows, webviews, and Orca's own UI. - Version-matched guidance: Loads the full command reference from the installed Orca binary via ORCA skills get computer-use, so instructions never drift from the binary actually executing commands. - Use Case: Ask the agent to read your unread Slack messages or check what is currently playing in Spotify, and it resolves the correct Orca executable, confirms the app is running, and reads the window state with JSON output. ## Quick Start Ask the agent to read the current state of a desktop app such as Slack or Spotify using Orca computer use.

Frequently Asked Questions about computer-use

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

FAQPage Schema
How do I control a desktop app from an AI agent?

Use Orca's computer-use CLI to list apps and windows, read the accessibility tree, then perform safe actions like click, type, scroll, and drag. Run ORCA skills get computer-use first to load the version-matched command guide for your installed binary.

How do I read what is visible in a desktop app window?

Read the app's accessibility tree through the Orca computer-use CLI, which exposes visible UI elements and app state as structured data. Prefer the --json flag so the agent can parse the output reliably.

Which Orca executable should I use on Linux?

On Linux outside an Orca-managed terminal, use orca-ide, never bare orca, since bare orca usually resolves to the GNOME Orca screen reader. In managed WSL sessions use ORCA_CLI_COMMAND, and in dev checkouts use orca-dev.

What if the Orca binary does not recognize skills get?

Only when the binary explicitly reports skills get as unknown, use the bounded read-only fallback: ORCA status --json, ORCA computer capabilities --json, and ORCA computer list-apps --json. Then update Orca to restore the full version-matched guide.

Can computer use automate browser windows and webviews?

Yes, the computer-use surface covers browser windows, webviews, and Orca's own UI in addition to native desktop apps. The same accessibility-tree inspection and safe UI actions apply across all of these window types.