appfunctions

Generate Kotlin AppFunction implementations with KSP and ADB testing.

Updated May 22, 2026
One-click install
npx skills add https://github.com/ekawijayasusilo/kmp_template --skill appfunctions-ekawijayasusilo
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: appfunctions
Source: https://github.com/ekawijayasusilo/kmp_template/tree/main/.claude/skills/appfunctions
Command: npx skills add https://github.com/ekawijayasusilo/kmp_template --skill appfunctions-ekawijayasusilo

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

AppFunctions helps you make Android app workflows usable by AI and system agents without forcing them to navigate the app UI, which reduces friction and improves hands-free automation.

Core Features & Use Cases

  • Workflow Discovery & Recommendation: Identify high-value, user-meaningful features and translate them into atomic AppFunctions suitable for agent use.
  • Implementation & Configuration: Generate Kotlin AppFunction implementations with correct KSP setup, Android metadata wiring, and safe parameter conventions.
  • KDoc Refinement for AI/MCP: Refine function and app-level descriptions so agents can reliably interpret constraints, sequencing, and inputs.
  • ADB Testing & Debugging: Validate registration and execution locally using device commands, including listing and invoking functions with JSON parameters.

Use case example: After implementing note creation and editing functions, an agent can trigger “create a note” or “edit a note” via system-level execution on-device instead of requiring a user to open the app and tap through menus.

Quick Start

Use the appfunctions skill to generate and refine AppFunctions for your Android app, then verify them by listing and invoking the registered functions via ADB.

Frequently Asked Questions about appfunctions

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

FAQPage Schema
How do I expose Android app workflows to AI agents without opening the UI?

You expose Android app workflows by implementing atomic Kotlin AppFunctions with KSP integration, allowing system agents to discover and execute tasks like note creation or media playback on-device without UI navigation.

What are the Android API level requirements for implementing AppFunctions?

Implementing AppFunctions requires targeting Android SDK level 36 or higher and compiling against SDK level 37 or higher, alongside Jetpack compatibility and KSP integration for metadata wiring.

How do I test and invoke AppFunctions locally using ADB?

You test AppFunctions locally by using ADB commands to list registered functions and invoke them directly with JSON parameters, validating registration and execution without needing the app UI.

How should I format KDoc for Android AppFunctions to ensure AI agents interpret constraints correctly?

Refine function and app-level KDoc descriptions to explicitly state constraints, sequencing, and safe parameter types so AI agents can reliably interpret and follow metadata-driven rules during execution.

Why must @AppFunctionContext be the first parameter in Android agent workflows?

@AppFunctionContext must be the first parameter to enforce safe type and serialization rules, ensuring system agents reliably select and run the function on-device without execution errors.