dsh-mobile-gui-agent DSH plugin

dsh-mobile-gui-agent is an installable DeepSeek Harness plugin that adds a mobile_gui_agent entry to the Harness Web UI. It drives every task through an observe → decide → act → verify loop, enabling automated GUI interaction with Android devices via ADB. The plugin provides ADB discovery, screenshots, UIAutomator hierarchy capture, and actions like tap, swipe, text input, and navigation.

remote-execution skills-workflows

How to install the dsh-mobile-gui-agent DSH plugin

dsh plugin --profile web add github:kunjinkao-os/dsh-mobile-gui-agent#v0.2.0

Copying does not run this command. Review the repository and version before installing the dsh-mobile-gui-agent DSH plugin.

dsh-mobile-gui-agent DSH plugin data source

dsh-mobile-gui-agent DSH plugin snapshot date: Aug 16, 2026

discovered

What the dsh-mobile-gui-agent DSH plugin can do

ADB device discovery and wireless connection management
Screenshot and UIAutomator hierarchy capture with element IDs
Strict phone_observe and phone_act tools for agent loop
Stale-element protection, adaptive screen stabilization, and stuck detection
Harness approval for semantic controls that can send, publish, delete, purchase, etc.

Where the dsh-mobile-gui-agent DSH plugin fits

Open Android Settings and navigate to Wi-Fi page
Open browser and focus the address bar
Type text in a focused text field on a mobile app
Open WeChat, find File Transfer, and prepare to send a message

Who the dsh-mobile-gui-agent DSH plugin is for

Android developers and testers who need automated mobile GUI testing
Automation enthusiasts who want to control Android devices from Harness

dsh-mobile-gui-agent DSH plugin limitations

Only supports Android devices that are visible to adb and properly authorized (USB debugging enabled, RSA key accepted)
Requires a DeepSeek Harness Web profile; the mobile_gui_agent entry is not available in headless-only profiles

dsh-mobile-gui-agent DSH plugin: from the repository README

Quoted from the kunjinkao-os/dsh-mobile-gui-agent README, the upstream source of the dsh-mobile-gui-agent DSH plugin. Copyright remains with the original authors.

English | [中文](docs/README.zh.md) | [Changelog](CHANGELOG.md) `dsh-mobile-gui-agent` is an installable DeepSeek Harness plugin for controlling Android devices through ADB. It adds a **mobile_gui_agent** entry to the Harness Web UI and drives every task through an observe → decide → act → verify loop. The repository is a single publishable npm package. The bundle patch inserts one Cordis plugin row; that plugin composes the ADB Provider, the Phone Agent Consumer, the Phone tools, the Typert Remote adapter, and the browser client under one lifecycle. ## Quick start Connect and authorize an Android device, then install the pinned release into the Harness Web profile: ```bash adb devices -l dsh plugin --profile web add github:kunjinkao-os/dsh-mobile-gui-agent#v0.2.0 dsh --profile web --dump-config dsh --profile web ``` The device row must report `device`. The configuration dump must contain a `# == dsh-mobile-gui-agent` layer and one `dsh-mobile-gui-agent` row. For the clearest workflow, send one short ordinary message in the Harness conversation first, then select **mobile_gui_agent** and enter the real phone command in its **Task** field. Do not enter the phone command in the o

Read the full READMERepository license: MIT

dsh-mobile-gui-agent DSH plugin questions

How do I install dsh-mobile-gui-agent?

Use the command `dsh plugin --profile web add github:kunjinkao-os/dsh-mobile-gui-agent#v0.2.0` to install the pinned release. Make sure you have a DeepSeek Harness Web profile set up. You can also install from a local checkout or tarball.

Does it work with iOS or other mobile platforms?

No, this plugin only supports Android devices. It relies on ADB (Android Debug Bridge) for communication. iOS devices are not supported.

What are the prerequisites for using this plugin?

You need Android Platform Tools installed (adb must be in PATH), a device with Developer options and USB debugging enabled, and the device must be authorized via the RSA prompt. Additionally, you need a DeepSeek Harness Web profile.

Can I use it in a headless Harness environment?

No, the mobile_gui_agent entry is a browser client contribution and requires a Web profile. It won't appear in headless-only profiles.

How does Unicode input work?

Unicode input uses the external ADB Keyboard helper (com.android.adbkeyboard/.AdbIME). If the helper is already installed, the plugin discovers it and temporarily enables/selects it for Unicode operations, then restores the previous IME. If absent, you can configure an absolute path to a reviewed APK; the plugin never downloads an APK.