Is the Agent Computer Use skill free?

Yes. The skill itself is free from kortix-ai/agent-computer-use. SkillProof publishes the install command and an independent test verdict at no cost.

What is the Agent Computer Use SkillProof Score?

8.0/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 5/10, docs & honesty 5/5.

How do I install Agent Computer Use?

Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.

Can I use Agent Computer Use with Cursor, Copilot, Gemini CLI, Codex or other AI tools?

The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.

Agent Computer Use

Drive any desktop app from the terminal via the agent-cu CLI (snapshot/click/type)

by kortix-ai · kortix-ai/agent-computer-use

Works with setup

Install — copy the command Download source ↗

Test report

Verdict: Works with setup
Score: 8.0/10
Tested: Jul 17, 2026
Environment: Claude Code 2.x (agent harness)
Upstream re-checked: Aug 10, 2026 · e8bd570

The SKILL.md is an unusually thorough desktop-automation playbook with verified Music/Calculator/Slack/Finder flows, a snapshot-identify-act-verify loop, and a detailed recovery ladder, but its delta over no skill could not be measured here: the agent-cu binary ships separately (cargo build or cargo install / npm) and requires a live desktop GUI, so the bundled wrapper reports 'agent-cu binary not found' in a headless run. Installing the skill folder alone gives you the instructions, not the tool. Worth flagging that the skill offers to auto-approve its own permission globally and unsupervised via `agent-cu setup --yes` (writes Bash(agent-cu *) to settings) — scoped to its own namespace and user-gated by default, but note it before installing.

Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works

Installs cleanly 5/5
Triggers reliably 5/5
Output vs. baseline 5/10
Docs & honesty 5/5

What Agent Computer Use does

A playbook for controlling macOS/Linux/Windows/Electron desktop apps through the agent-cu CLI using accessibility APIs: snapshot the screen, identify element refs, click/type/key/scroll/drag, then verify state. Triggers on desktop-automation tasks like 'open Music and play X', 'fill out this form', or 'send a Slack message'. The skill markdown is self-contained but the agent-cu binary is a separate cargo/npm install.

How to install Agent Computer Use

git clone https://github.com/kortix-ai/agent-computer-use
mkdir -p ~/.claude/skills
cd agent-computer-use && cp -r skills/agent-computer-use ~/.claude/skills/agent-computer-use

Skills live in ~/.claude/skills/ (global) or .claude/skills/ (per-project). Restart Claude Code after installing.

Commands — how to trigger Agent Computer Use

/agent-computer-use Drive any desktop app from the terminal via the agent-cu CLI (snapshot/click/type)

It also activates on plain-language prompts like these:

open Music and play my Espresso playlist for me
fill out this signup form in Safari using my saved info
send a Slack message to my team from the desktop app

Frequently asked questions

Is the Agent Computer Use skill free?: Yes. The skill itself is free from kortix-ai/agent-computer-use. SkillProof publishes the install command and an independent test verdict at no cost.
Does Agent Computer Use work with Claude Code?: We tested it with Claude Code 2.x (agent harness) on Jul 17, 2026. Verdict: Works with setup. The SKILL.md is an unusually thorough desktop-automation playbook with verified Music/Calculator/Slack/Finder flows, a snapshot-identify-act-verify loop, and a detailed recovery ladder, but its delta over no skill could not be measured here: the agent-cu binary ships separately (cargo build or cargo install / npm) and requires a live desktop GUI, so the bundled wrapper reports 'agent-cu binary not found' in a headless run. Installing the skill folder alone gives you the instructions, not the tool. Worth flagging that the skill offers to auto-approve its own permission globally and unsupervised via `agent-cu setup --yes` (writes Bash(agent-cu *) to settings) — scoped to its own namespace and user-gated by default, but note it before installing.
What is the Agent Computer Use SkillProof Score?: 8.0/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 5/10, docs & honesty 5/5.
How do I install Agent Computer Use?: Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
Can I use Agent Computer Use with Cursor, Copilot, Gemini CLI, Codex or other AI tools?: The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.