A robust automation skill for controlling Android devices through the Android Debug Bridge using a precision-based observe-locate-act-verify loop.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install adb-phone-control
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
Copy this prompt to OpenClaw to install it automatically.
Help me install adb-phone-control using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
The ADB Phone Control skill enables AI agents to interact directly with Android devices through a series of structured commands. By utilizing Openclaw Skills, this tool moves beyond simple coordinate guessing by analyzing the underlying UI hierarchy to find exact pixel bounds for elements. This ensures that actions like tapping, swiping, and typing are executed with high reliability across different screen resolutions and device models.
This skill is particularly valuable for developers and QA engineers who need to automate mobile workflows without manually identifying coordinates for every interaction. It bridges the gap between high-level intent and low-level device communication, providing a developer-friendly interface for mobile device orchestration.
To use this skill, ensure that adb and python3 are installed and available in your system PATH. Connect your Android device via USB with Debugging enabled, or connect over the network. Verify the connection with:
adb devices
Before running commands, source the provided helper script to load the necessary automation functions:
source "./adb-helpers.sh"
You can optionally set the ADB_OUTPUT_DIR environment variable to define where screenshots and UI dumps are stored.
The skill organizes its operational data and metadata as follows:
| Data Type | Format | Default Location | Description |
|---|---|---|---|
| UI Hierarchy | XML | /tmp/ui_dump.xml |
Full view tree including IDs, text, and bounds |
| Screen Capture | PNG | /tmp/adb_screen.png |
Visual snapshot for human-readable verification |
| Device Metadata | Text | Console Output | Screen resolution and package information |
| Logic Flow | Bash Helpers | adb-helpers.sh |
Core functions for parsing and acting on data |
Loading
A robust architecture that transforms AI agents from passive followers into persistent, self-improving, and proactive partners that anticipate needs.

An expert scheduling tool designed to create and manage reliable cron jobs and reminders specifically optimized for Feishu and multi-agent environments.

A comprehensive real-time monitoring and technical analysis system for Hong Kong and A-share stock markets featuring automated price alerts.

An AI-powered travel assistant that generates surprise itineraries to unvisited destinations based on your location, budget, and desire for relaxed exploration.

A comprehensive security toolkit for AI agents to identify code vulnerabilities, insecure dependencies, and hardcoded secrets.

A bidirectional synchronization tool that transforms human-written Obsidian notes into a machine-queryable ontology graph for advanced knowledge analysis.








































