Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add furkangonel/cowrangler --skill macos-computer-usegit clone --depth 1 https://github.com/furkangonel/cowranglerWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/furkangonel/cowrangler/macos-computer-use)<a href="https://agentmods.dev/skills/furkangonel/cowrangler/macos-computer-use"><img src="https://agentmods.dev/badge/skills/furkangonel/cowrangler/macos-computer-use.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00028 | $0.01159 |
| Opus 5 | $0.00014 | $0.00580 |
| Sonnet 5 | $0.00006 | $0.00232 |
| Haiku 4.5 | $0.00003 | $0.00116 |
Grade A, and why
macos-computer-use scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 110 lines — stays where its author put it; the contents beside it link to each section on GitHub.
macOS Computer Use (arka plan, model bağımsız)
computer_use aracın var ve Mac'i arka planda kontrol ediyor.
Yaptığın eylemler kullanıcının imlecini hareket ettirmiyor, klavye odağını çalmıyor, Space'i değiştirmiyor. Kullanıcı editörde yazarken sen başka bir Space'teki Safari'de tıklayabilirsin.
Her tool-capable modelle çalışır — Anthropic'e özgü schema yok.
Standart İş Akışı
Adım 1 — Önce capture.
computer_use(action="capture", mode="som", app="Safari")
Numaralı overlay'lerle ekran görüntüsü + AX ağacı döner:
#1 AXButton 'Geri' @ (12, 80, 28, 28) [Safari]
#2 AXTextField 'Adres ve Arama' @ (80, 80, 900, 32) [Safari]
#7 AXLink 'Giriş Yap' @ (900, 420, 80, 24) [Safari]
Adım 2 — Element index ile tıkla.
computer_use(action="click", element=7)
Koordinatlardan çok daha güvenilir. Her modelle çalışır.
Adım 3 — Doğrula. Durum değiştiren her eylemden sonra tekrar capture al:
computer_use(action="click", element=7, capture_after=True)
Capture Modları
mode |
Döner | En İyi Kullanım |
|---|---|---|
som (varsayılan) |
Ekran görüntüsü + numaralı overlay + AX index | Vision modeller; varsayılan tercih |
vision |
Sade ekran görüntüsü | SOM overlay görevi engellediğinde |
ax |
Yalnızca AX ağacı, görüntü yok | Metin tabanlı modeller |
Aksiyonlar
capture mode=som|vision|ax app=…
click element=N VEYA coordinate=[x, y]
double_click element=N VEYA coordinate=[x, y]
right_click element=N VEYA coordinate=[x, y]
scroll direction=up|down|left|right amount=3
type text="…"
key keys="cmd+s" | "return" | "escape"
wait seconds=0.5
list_apps
focus_app app="Safari" raise_window=false
Tüm aksiyonlar capture_after=True kabul eder.
Tüm element hedefli aksiyonlar modifiers=["cmd","shift"] kabul eder.
Arka Plan Kuralları
raise_window=Truekullanma — kullanıcı bunu açıkça istemedikçe.- Capture'ı uygulamaya kısıtla (
app="Safari") — daha az gürültü. - Space değiştirme — cua-driver, hangi Space'in görünür olduğundan bağımsız olarak her Space'teki elementleri kontrol eder.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 7d ago First seen · 110 lines · 28 tokens per session scan A 0e8aed5fabe9
macos-computer-use is a skill published in the GitHub repository furkangonel/cowrangler (2 stars, last pushed yesterday), licensed MIT. It adds 28 tokens to every session and 1,159 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
host-computer-use
Beta desktop control through the connected A0 CLI host. Use for the user's host/local computer screenshots, screen inspection, menus, native app UI, OS-level clicking, scrolling, typing, or checking computeruseremote status. Use instead of linux-desktop for host/local machine control. Do not use for ordinary browser…
macos-harness
Control a whole Mac from one persistent Python session with screenshots, PID-targeted input, an animated virtual pointer, targeted Apple Accessibility, Apple Events, Browser Harness CDP, and filesystem access. Use for native, Electron, browser, dialog, file, or cross-app tasks without moving the physical cursor or…
linux-desktop
Use only for Agent Zero's built-in Docker/Xpra Linux Desktop, XFCE apps, LibreOffice GUI apps, file manager, terminal, or visual workflows inside the Agent Zero runtime. Not for A0 CLI /computer-use or computeruseremote host control.
rust-desktop-applications
Build cross-platform desktop applications with Rust using Tauri framework and native GUI alternatives.
webview
Show a native macOS webview UI to the user and get structured input back. Use when an interactive macOS session needs human-in-the-loop for multi-field input, option selection from 5+ choices, approval with context, or content review — NOT for yes/no questions. Handles the full flow — spawning the webview, generating…
control-window
Use when the user wants Claude to control, drive, or automate another app's window on macOS — clicking, typing, pressing keys, scrolling, or reading what's on screen in a window that isn't the terminal. Triggers on requests like "control my Telegram window", "drive this app for me", "click the button in that window"…