Mano skill
Mano skill (mano-cua)
A CLI and skill for desktop control from a text task: screenshots, clicks, typing and scrolling, in the cloud or locally on Apple Silicon
- 47 ★ on the repository
- Last commit
Broad access
Why this level
- Controls the mouse and keyboard and opens and closes apps
- In cloud mode sends screenshots to an external server
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/mano-skill-mano-cuaDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Controls the mouse and keyboard and opens and closes apps
- In cloud mode sends screenshots to an external server
This is third-party code. Review the repository files before installing.
What it does
The mano-cua tool performs GUI tasks where an app has no API or CLI. At each step it takes a screenshot, sends it with the task to a model and executes the returned action: click, type, hotkey, scroll, drag or app launch. In cloud mode the server picks between the lightweight Mano model and Claude CUA. In local mode the Mano-P model runs on-device via MLX and no data leaves the machine. The skill teaches the agent to call this tool.
Who it is for. For people who need to automate desktop apps without an API, mainly on macOS.
Good fit when
- You need repetitive actions in an app that has no API
- You want to fill a form or create a record in a desktop app from a text description
- You need GUI automation without sending screenshots to the cloud on an Apple Silicon Mac
Not a fit when
- The app has an API, CLI or MCP server, which is more reliable
- Sensitive data is on screen that the model must not see
- You need several tasks running at once on one device
Example request
Open Notes and create a new note titled Meeting SummaryLimitations
In cloud mode screenshots go to the developer's server and complex steps are handled by a Claude model, so access from Russia without a VPN is not guaranteed. Local mode works only on macOS with Apple Silicon. Screen recording and accessibility permissions are required. One task runs at a time, only the primary display is used, and you must not touch the mouse or keyboard during a run. Windows and Linux support is incomplete.
How to disable. Stop a task with mano-cua stop, remove the tool with brew uninstall mano-cua and delete the skill folder from your agent's skills directory. Revoke the permissions in macOS privacy settings.
FAQ
How do I run without the cloud?
On an Apple Silicon Mac run mano-cua check, mano-cua install-sdk and mano-cua install-model, then start tasks with --local.
Does it ask before irreversible actions?
According to the README, sensitive or irreversible actions pause the task until the user explicitly confirms.
Discussion
No questions or reviews yet.
Related
SkillsA macOS CLI and MCP server: screenshots, accessibility-based UI inspection and app control via clicks and typing
An MCP server for macOS: the agent moves the mouse, types, manages windows and captures the screen when an app has no API
A Swift MCP server that lets an agent drive any macOS app: click, type and read the interface tree
Clawd Cursor: desktop control through MCP
Clawd Cursor
An MCP server that lets an agent drive windows and apps: it reads the screen as structure and acts by element name
Open Computer Use
open-computer-use
An open MCP server for desktop control: the agent reads app state and clicks, types and scrolls on macOS, Linux and Windows
A Rust CLI for macOS: an agent reads any app's interface through the accessibility tree and clicks, types and scrolls by element refs, with no screenshots
README in short
The README explains the loop: screenshot, server-side model selection, action execution. Install via Homebrew on macOS and Linux or a release archive on Windows. It covers run and stop with --url, --app, --minimize and --max-steps, local mode setup, required permissions and safety rules. MIT-0 licensed.
README badge
Original SKILL.md
---
name: mano-cua
description: Computer use for GUI automation tasks via VLA models. Use when the user describes a task in natural language that requires visual screen interaction and no API or CLI exists for the target app.
homepage: https://github.com/Mininglamp-AI/mano-skill
metadata: {"openclaw": {"emoji": "🖥️", "install": [{"id": "brew", "kind": "brew", "formula":"Mininglamp-AI/tap/mano-cua", "bins":["mano-cua"],"label": "Install mano-cua (brew)"}]}}
---
# mano-cua
Desktop GUI automation for tasks via VLA models. Use when the user describes a task in natural language that requires visual screen interaction and no API or CLI exists for the target app. Supports fully on-device local mode and cloud mode.
## Requirements
- A system with a **graphical desktop** (macOS / Windows / Linux)
- `mano-cua` binary installed (v1.1.0+ recommended for full feature support)
### Installation
**macOS / Linux (Homebrew):**
```bash
brew install Mininglamp-AI/tap/mano-cua
# Update to latest version
brew upgrade Mininglamp-AI/tap/mano-cua
```
**Windows:**
Download the latest `mano-cua-windows.zip` from [GitHub Releases](https://github.com/Mininglamp-AI/mano-skill/releases), extract it, and add the folder to your `PATH`.
## Usage
```bash
# Run a task
mano-cua run "your task description"
# Run with options (minimize UI panel and set max steps)
mano-cua run "task" --minimize --max-steps 10
# Open a URL in the browser before starting the task
mano-cua run "task" --url "https://example.com"
# Open an app before starting the task
mano-cua run "task" --app "Notes"
# Run in local mode (on-device inference, macOS Apple Silicon only)
mano-cua run "task" --local
# Stop the current running task
mano-cua stop
```