A local-first automation skill for controlling Android devices over a network without requiring USB connections, ADB, or root access.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install agent-android
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 agent-android using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
Agent Android is a powerful addition to the library of Openclaw Skills designed for remote device interaction and robotic process automation (RPA). It allows an AI agent to drive an Android device through a lightweight HTTP service hosted directly on the phone. This skill operates on a local-first principle, ensuring that UI reads, taps, inputs, and screenshots remain within your local area network (LAN) for maximum privacy and performance. By bypassing the need for traditional developer tools like ADB, it simplifies the workflow for non-technical setups while providing deep technical access to the device UI tree.
To get started with this skill from the Openclaw Skills collection, ensure the CLI is available on your path. Install it using the uv package manager:
uv tool install aivane-agent-android
uv tool update-shell
You must also have the AIVane app running on your Android device to host the local API service at a URL like http://<device-ip>:8080.
The skill organizes data primarily through the Android UI hierarchy and metadata. Information is structured as follows:
| Data Type | Description |
|---|---|
| UI Tree | A hierarchical representation of all visible screen elements and their attributes. |
| RefID | A temporary session-based integer assigned to elements for quick REPL interaction. |
| Resource ID | The native Android developer-defined ID for specific UI components. |
| XPath | Dynamically generated path strings used to uniquely identify elements in complex layouts. |
| Attributes | Metadata for elements including text, class, bounds, and clickable status. |
Loading
A hybrid memory management system providing vector search and SQLite-based FTS for persistent AI agent context.

A high-traceability memory continuity system for AI agents that uses structured Markdown files as a persistent source of truth.

A sophisticated dual-layer memory management system providing AI agents with persistent, hybrid search capabilities.

An AI agent skill that generates high-contrast dark posters and tutorial covers featuring authentic handwritten marker typography and minimalist design.

An AI-powered consultant designed to build, optimize, and present winning business models for the China International College Students' Innovation Competition.

A sophisticated cognitive architecture tool designed to transform fragmented information into a structured, actionable external brain.








































