appfunctions

Generate Kotlin AppFunctions from Android app workflows and validate via ADB.

174|8|Updated Jan 9, 2024
One-click install
npx skills add https://github.com/klxiaoniu/QQVersionList --skill appfunctions-klxiaoniu
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: appfunctions
Source: https://github.com/klxiaoniu/QQVersionList/tree/main/.agents/skills/appfunctions
Command: npx skills add https://github.com/klxiaoniu/QQVersionList --skill appfunctions-klxiaoniu

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

AppFunctions reduces the friction of letting AI agents perform real user workflows on-device without relying on brittle UI automation or asking the user to manually open the app.

Core Features & Use Cases

  • Workflow Discovery & Recommendation: Analyze an Android app’s codebase to recommend high-value AppFunctions that translate user journeys into agent-executable actions.
  • Implementation & Configuration Generation: Generate Kotlin AppFunction implementations plus required metadata wiring so the system can discover them.
  • KDoc Refinement for Agent Reliability: Optimize KDoc so LLMs and Model Context Protocol consumers interpret parameters, constraints, and dependencies correctly.
  • Testing & Debugging via ADB: Use ADB commands to list registered functions, execute them safely with JSON parameters, and verify service availability.

Quick Start

Tell the AI to analyze an Android codebase and generate AppFunction implementations for the top user workflows, then refine the KDoc and provide ADB commands to test them on a connected device.

Frequently Asked Questions about appfunctions

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

FAQPage Schema
How do I let AI agents execute Android app workflows without UI automation?

You can let AI agents execute Android app workflows by exposing them as on-device AppFunctions. This allows agents to discover and run specific actions without opening the app UI or relying on brittle automation.

How to generate Kotlin AppFunction implementations from an Android codebase?

To generate Kotlin AppFunction implementations, analyze your Android codebase to identify high-value workflows. Then generate the Kotlin code with correct AppFunctionContext parameter ordering and produce the required app_metadata configuration.

What is the best way to test AppFunctions on a connected Android device?

The best way to test AppFunctions is via ADB commands. You can use ADB tooling to list registered functions, execute them safely with JSON parameters, and verify service availability on a connected device.

Why do AI agents need optimized KDoc for AppFunctions?

AI agents need optimized KDoc to correctly interpret parameters, constraints, and dependencies. Refining KDoc ensures that LLMs and Model Context Protocol consumers can reliably understand and execute the exposed AppFunctions.

Can I use AppFunctions for constrained multi-step execution workflows?

Yes, you can use AppFunctions for constrained multi-step execution workflows. The system applies to agentic automation scenarios such as notes, media playback, and shortcut or voice-triggered actions.

Does generating AppFunctions require scanning Android app manifests?

Yes, generating AppFunctions requires scanning Android app entry points and manifests. This process identifies user journeys and translates them into agent-executable actions with the correct metadata wiring.