mobiagent-task-spec

Write precise mobile task specifications for MobiAgent execution.

44|1|Updated Mar 12, 2026
One-click install
npx skills add https://github.com/IPADS-SAI/MobiClaw --skill mobiagent-task-spec
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mobiagent-task-spec
Source: https://github.com/IPADS-SAI/MobiClaw/tree/main/mobiclaw/skills/mobiagent-task-spec
Command: npx skills add https://github.com/IPADS-SAI/MobiClaw --skill mobiagent-task-spec

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill turns vague mobile requests into precise, executable task descriptions for MobiAgent and steward workflows, reducing failed runs and ambiguous phone interactions.

Core Features & Use Cases

  • Task Writing for Mobile Automation: Converts broad goals into single-app, step-by-step instructions that use visible UI anchors.
  • Completion and Safety Criteria: Adds clear success conditions, fallback paths, and禁止 side effects such as sending messages or submitting actions unintentionally.
  • Evidence-Based Rewrites: Improves failed task descriptions using final screenshot evidence, VLM summaries, extracted screen text, and prior reasoning.
  • Use Case: A user wants to browse to a specific chat, confirm the target page, and stop before sending anything; this Skill rewrites that request into a robust MobiAgent-ready spec.

Quick Start

Write a single natural-language task description for MobiAgent that names the app, states the goal, lists visible anchors and fallback steps, and defines the exact success condition.

Frequently Asked Questions about mobiagent-task-spec

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

FAQPage Schema
How do I write precise mobile automation task instructions for MobiAgent?

To write precise mobile automation task instructions for MobiAgent, convert vague goals into single-app, step-by-step click and swipe instructions that use visible UI anchors and define exact success conditions to ensure robust execution.

How can I use screenshot evidence to fix failed mobile automation tasks?

You can use screenshot evidence to fix failed mobile automation tasks by applying the evidence-based rewrite process, which analyzes final screenshots, VLM summaries, and extracted screen text to refine ambiguous task descriptions into accurate specs.

How do I add safety constraints to mobile automation task specs?

Add safety constraints to mobile automation task specs by defining explicit fallback paths and prohibited side effects, such as preventing the agent from unintentionally sending messages or submitting actions before reaching the final verification page.

What is the best way to structure a natural language prompt for phone task automation?

The best way to structure a natural language prompt for phone task automation is to name the target app, state the goal, list visible anchors with fallback steps, and define the exact completion criteria to ensure reliable MobiAgent execution.

Why do my vague mobile automation requests fail during execution?

Vague mobile automation requests fail during execution because they lack specific UI anchors, step-by-step click instructions, and clear completion criteria, causing ambiguous phone interactions that the agent cannot reliably navigate.

Can I use MobiAgent task specs for multi-app automation workflows?

MobiAgent task specs are designed for single-app, step-by-step instructions that rely on visible UI anchors. They are not intended for complex multi-app automation workflows, ensuring focused and reliable execution within one application environment.