cua-driver

cua-driver is a skill for Claude Code, Codex from QwenLM/qwen-code. It costs 88 tokens per session (14,234 once invoked), scanned A, original, Apache-2.0.

Instructions for controlling a desktop application through its graphical interface on macOS, Windows, or Linux. The process reads the app's accessibility information before acting and checks the result from a fresh view.

In plain words
What is it for?
Use it to launch and operate native desktop apps, inspect their controls, choose menus or buttons, enter information, and confirm that actions produced the intended result.
Why use it?
It reduces mistakes caused by clicking the wrong control, using outdated screen information, or relying on platform-specific behavior without verification.

Skill for Claude CodeCodex ✓ vendor

Which agent this was written for is unclear — built for openclaw. Also seen: mentions subagents; mentions Claude Code; built for openclaw.

Good fit Use it to launch and operate native desktop apps, inspect their controls…

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/qwenlm/qwen-code/cua-driver
About the project

Qwen Code is an open-source AI coding agent that runs in a terminal and helps developers work with code through language models. It supports multiple model providers and can also be used through IDEs, desktop software, SDKs, and messaging bots.

QwenLM/qwen-code · 27,673 stars · on GitHub · qwenlm.github.io

Install

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.

Any agent
npx skills add QwenLM/qwen-code --skill cua-driver
Clone the repo
git clone --depth 1 https://github.com/QwenLM/qwen-code

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for cua-driver

README.md
[![agentmods](https://agentmods.dev/badge/skills/qwenlm/qwen-code/cua-driver.svg)](https://agentmods.dev/skills/qwenlm/qwen-code/cua-driver)
Your own site
<a href="https://agentmods.dev/skills/qwenlm/qwen-code/cua-driver"><img src="https://agentmods.dev/badge/skills/qwenlm/qwen-code/cua-driver.svg" alt="Measured on agentmods" height="20"></a>
Per session 88 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 14,234 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00088 $0.14234
Opus 5 $0.00044 $0.07117
Sonnet 5 $0.00018 $0.02847
Haiku 4.5 $0.00009 $0.01423

Measured 4d ago against content hash 808b67eb6ca7, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

cua-driver 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 4d 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.

packages/cua-driver/rust/Skills/cua-driver/SKILL.md · 1,082 lines

How it starts

The opening of the file, as written. The whole thing — 1,082 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Qwen Cua Driver

Orchestrates cross-platform app automation via qwen-cua-driver. Whenever a user asks to drive a native app, follow the loop in this skill rather than calling tools ad-hoc — the snapshot-before-action invariant is not optional and silently breaks if you skip it.

Platform-specific reading — read this first

This file is the cross-platform core: snapshot invariant, CLI vs MCP choice, tool surface naming, behavior matrix, canonical loop, pixel-click contract, common failure modes. The platform-specific material (forbidden-list, accessibility tree implementation, launch semantics, click dispatch) lives in companion files in this same directory:

  • macOS — read MACOS.md (no-foreground contract, forbidden open/osascript/cliclick invocations, AXMenuBar navigation, SkyLight pixel-click dispatch).
  • Windows — read WINDOWS.md (UIA tree vs AX, UWP / ApplicationFrameHost hosting, layered UIA+PostMessage click chain, Session 0 isolation, Windows-specific focus-steal vectors).
  • Linux — read LINUX.md (X11 background input via AT-SPI + XSendEvent and compositor-specific Wayland capabilities).

Cross-cutting topics also have their own files:

  • BROWSER.md — exact native-window binding, explicit browser preparation, typed Chromium/Electron page tools, input trust classes, and native fallbacks for browser chrome and unsupported engines.
  • RECORDING.md — session recording + replay_trajectory.

Use whichever combination matches the host. When in doubt, run qwen-cua-driver doctor — it reports the platform and the right entry point.

Start with the narrowest semantic route

Before opening or operating an application, name the desired postcondition and use the first applicable route below. Verify the result in the same domain before stopping or advancing:

  1. Caller-provided headless/background operation for a non-GUI outcome. Prefer an exact application API/SDK, service or database client, CLI, or filesystem operation over imitating a user. This includes batch-safe file moves, renames, copies, directory creation, archive extraction, data conversion, and process inspection. Read the resulting semantic state back; a zero exit status alone is not proof.
  2. Typed Cua operation for an application or window outcome. Use set_window_frame for exact geometry, invoke_menu for a known native application-menu path, typed browser tools for supported page content, and clipboard tools for clipboard state. Verify with list_windows, get_browser_state, or clipboard_read, respectively.
  3. Background accessibility action. Use a fresh AX/UIA/AT-SPI target.
  4. Background pixel action. Use the pixels from the same state snapshot.
  5. Foreground delivery. Retry only the action that evidence says could not land in the background.
  6. Desktop fallback. Select an exact desktop target for that call only. Later calls may return to an exact window target in the same session.

Read the full file on GitHub · 1,082 lines

Files

What ships with it

7 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

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.

  1. 4d ago Changed 808b67eb6ca7
  2. 7d ago First seen · 1,082 lines · 88 tokens per session scan A 812eae055040

Subscribe to this mod's changes

cua-driver is a skill published in the GitHub repository QwenLM/qwen-code (27,673 stars, last pushed today), licensed Apache-2.0. It adds 88 tokens to every session and 14,234 once invoked, about $0.0004 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 2026-08-30.

Related

Other skills, from other repositories

agent-manager

Run a fleet of AI coding agents as live tmux sessions with agent-manager. Use when a developer is running more than one coding agent, needs to see which one is working or blocked, wants to spawn another on an independent task, or wants to review an agent's diff without leaving the terminal.

YoanWai/agent-manager · 63 tokens

agent-manager-reference

Query agent-manager.dev's reference API and MCP server for documentation, the coding CLIs it manages, the tools its MCP server exposes, and the current release. Use instead of scraping the website's HTML.

YoanWai/agent-manager · 45 tokens

sls-dashboard-builder

A tool for creating and modifying importable JSON dashboards for Alibaba Cloud SLS, a service for searching and monitoring logs. It maps checked queries and analysis needs to dashboard charts.

alibaba/loongsuite-pilot · 72 tokens

loongsuite-pilot-insight

A reporting workflow for turning LoongSuite Pilot and AI coding-agent logs into structured reports about events, teams, data quality, development efficiency, and AI use. It defines the meaning of the log fields and the measurements used in dashboards.

alibaba/loongsuite-pilot · 91 tokens

loongsuite-pilot-ops

Skill "loongsuite-pilot-ops" from alibaba/loongsuite-pilot, covering loongsuite-pilot-ops, quick start, todo: add quick start commands and usage.

alibaba/loongsuite-pilot · 15 tokens

proactive-agent

Transform AI agents from task-followers into proactive partners that anticipate needs and continuously improve. Now with WAL Protocol, Working Buffer, Autonomous Crons, and battle-tested patterns. Part of the Hal Stack 🦞.

MerkyorLynn/Lynn · 46 tokens