skills.rest
omnimind-ai
omnimind-aiOFFICIAL
Explore all skills by this author →
WebsiteGitHub

omniflow-gui-runtime

Installs the pinned OmniFlow GUI runtime for Android VLM automation via MCP.

2.0k|140|Updated Mar 18, 2026
Search Tags:#mcp#android#vlm#runtime installation#gui automation#omniflow#on-device agent
One-click install
npx skills add https://github.com/omnimind-ai/OpenOmniBot --skill omniflow-gui-runtime
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: omniflow-gui-runtime
Source: https://github.com/omnimind-ai/OpenOmniBot/tree/main/plugins/omni-vlm-lite/runtime-skill/omniflow-gui-runtime
Command: npx skills add https://github.com/omnimind-ai/OpenOmniBot --skill omniflow-gui-runtime

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Running GUI automation with a vision-language model on Android requires a correctly versioned runtime, checkpoint, and transfer layer. This Skill installs the pinned OmniFlow GUI runtime so the Omni VLM plugin works without manual dependency resolution, virtual environments, or private installers.

Core Features & Use Cases

  • Pinned Runtime Installation: Installs a versioned OmniFlow runtime with canonical OmniTransfer v9, checkpoint, and bundled pure-Python dependencies from a GitHub Release asset.
  • Standard MCP Exposure: Exposes run_gui, run_function, list_functions, and register_function tools through OpenOmniBot's built-in Streamable HTTP MCP endpoint, usable by any standard MCP host.
  • Function Registry: Registers successful run_gui RunLogs as reusable Functions for repeatable GUI automation tasks.
  • Use Case: An MCP host connects to an Android phone running OpenOmniBot, calls run_gui to execute a VLM-driven GUI task, then registers the successful run as a Function for later replay.

Quick Start

Ask the assistant to install the OmniFlow GUI runtime and then run a new GUI task using the run_gui tool through the MCP endpoint.

Frequently Asked Questions about omniflow-gui-runtime

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run GUI automation tasks on Android with a VLM?▼

Use the run_gui tool exposed through OpenOmniBot's Streamable HTTP MCP endpoint after installing this runtime. Any standard MCP host can call run_gui to execute a new VLM-driven GUI task on the device.

What tools does the OmniFlow MCP runtime expose?▼

The runtime exposes run_gui for new tasks, run_function for registered Functions, list_functions for discovery, and register_function to save a successful run_gui RunLog as a reusable Function.

Does the OmniFlow runtime require a Python virtual environment?▼

No. The runtime has no nested ZIP, virtual environment, dependency resolver, or private installer. The versioned GitHub Release asset bundles OmniFlow, OmniTransfer v9, the checkpoint, and a small pure-Python dependency.

What Android devices are compatible with the OmniFlow GUI runtime?▼

The runtime requires Android arm64-v8a with the Omnibot Alpine runtime. OpenOmniBot owns Android permissions, observations, actions, model access, and the process bridge.

What configuration does the OmniFlow MCP runtime need?▼

Set OMNIFLOW_DEVICE_MCP_URL to the device MCP endpoint, OMNIFLOW_DEVICE_MCP_TOKEN copied from OpenOmniBot's MCP settings, and optionally OMNIFLOW_STORE_PATH for the local Function store.

Related Skills

Explore More Software Engineering Skills→
TrainLCDTrainLCD
71

argent-settings-permissions

Manage iOS simulator and Android device runtime permissions via platform permission stores.

Official
Advanced
HKUDSHKUDS
48.8k

cli-anything-macrocli

Define, inspect, and execute parameterized GUI macros through a JSON-output CLI.

Official
Intermediate
Proxy2021Proxy2021
5

remote_desktop

Execute input actions and capture screens on remote desktops.

Community
Intermediate
FilenCloudDiensteFilenCloudDienste
34

argent-android-emulator-setup

Boot an Android AVD and prepare device connectivity for UI automation.

Official
Intermediate