android-automation-agent

Automates Android UI tasks on a connected phone via natural-language commands and ADB.

21|4|Updated Apr 2, 2026
One-click install
npx skills add https://github.com/Mohd-Mursaleen/android-automation-agent --skill android-automation-agent
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: android-automation-agent
Source: https://github.com/Mohd-Mursaleen/android-automation-agent/tree/main
Command: npx skills add https://github.com/Mohd-Mursaleen/android-automation-agent --skill android-automation-agent

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires adb, python3, curl, bash, and includes scripts (resource) components.

What problem does it solve?

This Skill removes the need for manual tapping, scripting, and fragile UI selectors when controlling an Android device, letting you drive real phone workflows from plain language.

Core Features & Use Cases

  • Natural-language Android automation: Open apps, search, tap, scroll, type, and navigate through multi-step mobile flows.
  • Vision + accessibility-based control: Uses screenshots and the Android UI tree to target on-screen elements with precise coordinates.
  • Reliable task execution: Handles planning, execution, verification, retries, lock checking, and progress reporting for long-running tasks.
  • Use cases: Mobile app testing, Android RPA, shopping and food-ordering flows, account and inbox checks, and screen inspection.

Quick Start

Tell the agent exactly what you want done on the connected Android phone, such as opening an app, finding an item, or checking what is on screen.

Frequently Asked Questions about android-automation-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate Android UI testing using natural language instead of scripts?

Natural-language Android automation translates plain-language commands into UI actions using screenshots and the accessibility tree. It drives multi-step mobile workflows like app launching, tapping, and scrolling on a connected device without requiring manual scripting or fragile selectors.

Do I need ADB and Python to control an Android phone with natural language?

Yes, natural-language Android control requires ADB for device connectivity, Python 3.10+, and OpenRouter access. These dependencies enable screenshot capture, accessibility-tree parsing, and the execution of multi-step mobile workflows.

Can I use LLM vision to navigate multi-step mobile workflows like ordering food?

Yes, LLM vision combined with accessibility-tree parsing can navigate multi-step mobile workflows like food ordering and shopping. The agent plans, executes, verifies, and retries actions on a connected Android device to reliably complete long-running tasks.

What is the best way to run mobile app testing on Android without manual tapping?

Natural-language Android automation is ideal for mobile app testing without manual tapping. It uses LLM vision and the accessibility tree to drive real phone workflows, handling lock checking, retries, and progress reporting for reliable task execution.

Does Android automation work with Termux for executing plain-language commands?

Android automation with plain-language commands works within Termux using bash, Python 3.10+, and curl to interface with OpenRouter. It requires ADB connectivity to handle device wake-and-unlock sequences and execute UI actions on the connected phone.