ClawKit Logo
ClawKitReliability Toolkit
Back to Registry
Official Verified utilities Safety 3/5

Android Agent

Skill by harshilmathur

Why use this skill?

Take control of your Android phone with the OpenClaw Android Agent. Let AI open apps, tap buttons, and complete tasks on your device via ADB.

skill-install — Terminal

Install via CLI (Recommended)

clawhub install openclaw/skills/skills/harshilmathur/android-agent
Or

What This Skill Does

The Android Agent skill is a powerful automation bridge that transforms any connected Android device into an extension of your OpenClaw ecosystem. By leveraging the DroidRun framework and GPT-4o vision capabilities, this agent can perceive the visual state of your phone screen, interpret complex user intent, and execute physical-like interactions such as tapping, scrolling, and text entry via ADB (Android Debug Bridge). Whether your phone is connected directly to your host machine or accessed remotely via a network node like a Raspberry Pi, the agent orchestrates the entire workflow—waking, unlocking, and navigating apps—to complete tasks exactly as a human user would.

Installation

To integrate the Android Agent into your environment, follow these steps:

  1. Prepare your Android device by navigating to Settings, tapping the Build Number seven times to unlock Developer Options, and ensuring 'USB Debugging' is enabled.
  2. Connect your device to your OpenClaw host or node machine via USB.
  3. Install the skill by running: clawhub install openclaw/skills/skills/harshilmathur/android-agent.
  4. Verify device connectivity with the adb devices command, ensuring you authorize the debugging prompt on the phone screen.
  5. Set your OPENAI_API_KEY as an environment variable to authorize the vision-based reasoning engine.

Use Cases

This skill is ideal for automating repetitive mobile-first tasks that lack dedicated desktop APIs. Common scenarios include managing home automation apps, interacting with messaging platforms like WhatsApp or Telegram, and streamlining shopping workflows on grocery or e-commerce apps. It is also an excellent tool for developers needing to perform regression testing on UI elements or remote administration of physical device farms.

Example Prompts

  • "Open WhatsApp and send a message to Mom saying I'm running ten minutes late."
  • "Open Google Maps and provide directions to the closest Starbucks."
  • "Search Amazon for a USB-C charging cable under twenty dollars and add it to my cart."

Tips & Limitations

For optimal performance, ensure your device remains unlocked or keep the screen timeout duration extended. The agent relies heavily on visual perception; therefore, ensure the screen is clean and brightness is set appropriately to allow the vision model to accurately interpret UI elements. Note that this skill requires physical or network access to an Android device and may be subject to latency depending on your network conditions if using Node mode.

Metadata

Stars2387
Views0
Updated2026-03-09
View Author Profile
AI Skill Finder

Not sure this is the right skill?

Describe what you want to build — we'll match you to the best skill from 16,000+ options.

Find the right skill
Add to Configuration

Paste this into your clawhub.json to enable this plugin.

{
  "plugins": {
    "official-harshilmathur-android-agent": {
      "enabled": true,
      "auto_update": true
    }
  }
}

Tags(AI)

#android#automation#adb#mobile-control#vision
Safety Score: 3/5

Flags: network-access, external-api, code-execution