GitHub - franzenzenhofer/big-arrow-on-the-screen: Let your AI agents paint big arrows, boxes and text on your Mac screen. One CLI, click-through, gone by itself. Skill for Claude Code and Codex. MIT.

big-arrow-on-the-screen is a macOS command-line tool that lets AI agents draw temporary arrows, boxes, rings, and text over applications to guide users visually.

Visit Website
GitHub - franzenzenhofer/big-arrow-on-the-screen: Let your AI agents paint big arrows, boxes and text on your Mac screen. One CLI, click-through, gone by itself. Skill for Claude Code and Codex. MIT.

Introduction

Overview

big-arrow-on-the-screen, also called bigarrow, is a macOS command-line tool and skill for Claude Code and Codex. It gives AI agents a visual way to point users toward a specific interface element instead of relying on a message hidden in a terminal.

The tool places a transparent overlay above other windows across displays and Spaces. Mouse clicks pass through the overlay, keyboard focus remains on the current application, and the annotation removes itself.

Key Features

  • Draws arrows with accompanying text over macOS applications.
  • Supports visual styles including rings and boxes, multiple arrow shapes, colors, borders, sizes, and directions.
  • Targets interface elements by name, application, and accessibility role.
  • Can raise an application window and select a named browser tab.
  • Runs as a single Swift binary without a daemon, menu-bar icon, account, or telemetry.
  • Does not require macOS permission merely to draw the overlay.

How It Works

Agents invoke bigarrow from the command line and identify the application and element that need attention. They can add explanatory text and customize where the arrow originates, how it looks, and whether a close button is shown.

The resulting overlay stays visually prominent while preserving interaction with the underlying application. This makes the annotation suitable for situations where a person must complete a step that an agent cannot perform directly.

Use Cases

bigarrow can indicate a permission toggle in System Settings, walk through controls in Keynote, distinguish the correct control in a print dialog, identify one browser tab among many, or clarify which Finder view button to use. These examples center on precise, temporary guidance within real desktop applications rather than automated clicking.

The project is MIT licensed and contains no built-in AI model; it serves as a visual communication layer for agents and command-line workflows.