Open Autoglm Phone Agent

This gives Claude the ability to work with Open-AutoGLM, a framework that lets you control Android, HarmonyOS, and iOS devices through natural language by running a 9B parameter vision-language model. You point it at screenshots and say things like "open Meituan and search for nearby hot pot restaurants," and it generates structured ADB/HDC/WebDriverAgent commands to execute the task. The skill covers both self-hosted deployment with vLLM or SGLang and third-party APIs like BigModel. The documentation is thorough on the model serving requirements, which matters because the chain-of-thought output is fragile and breaks if you misconfigure the max_pixels or context length parameters. Useful if you're building phone automation workflows or testing mobile apps with AI-driven interactions instead of brittle UI selectors.

skills.sh installs: 1.3K
GitHub stars: 61
last updated: Jul 14, 2026

View on GitHub

Install

$ npx -y skills add aradotso/trending-skills --skill open-autoglm-phone-agent --agent claude-code

Installs into .claude/skills of the current project.

Files

Open-AutoGLM Phone Agent

Skill by ara.so — Daily 2026 Skills collection.

Open-AutoGLM is an open-source AI phone agent framework that enables natural language control of Android, HarmonyOS NEXT, and iOS devices. It uses the AutoGLM vision-language model (9B parameters) to perceive screen content and execute multi-step tasks like "open Meituan and search for nearby hot pot restaurants."

Architecture Overview

User Natural Language → AutoGLM VLM → Screen Perception → ADB/HDC/WebDriverAgent → Device Actions

Model: AutoGLM-Phone-9B (Chinese-optimized) or AutoGLM-Phone-9B-Multilingual
Device control: ADB (Android), HDC (HarmonyOS NEXT), WebDriverAgent (iOS)
Model serving: vLLM or SGLang (self-hosted) or BigModel/ModelScope API
Input: Screenshot + task description → Output: structured action commands