Installable skillApache-2.0

appfunctions

Analyzes Android apps to identify key user workflows for AppFunctions

BY androidUPDATED Oct 27, 2026
INTENDED FOR◈OpenClawInstall in a listed agent.

instructions.md

---
name: appfunctions
description: Analyzes Android apps to identify key user workflows for AppFunctions
  such as creating a note, playing media, or sending an automated or AI agent triggered
  message, voice commands, or system shortcuts, without needing to open the app UI.
  Generates Kotlin code to expose these workflows to the Android system, allowing
  agents to discover and execute them on-device. Also refines KDoc documentation to
  ensure AI agents correctly understand and use the provided functionality.
license: Complete terms in LICENSE.txt
metadata:
  author: Google LLC
  last-updated: '2026-05-16'
  keywords:
  - AppFunctions
  - Kotlin
  - KSP
  - ADB
  - AI
  - LLM
  - MCP
---

Analyzes Android apps to identify key user workflows for AppFunctions such as
creating a note, playing media, or sending an automated or AI agent triggered
message, voice commands, or system shortcuts, without needing to open the app
UI.

Generates Kotlin code to expose these workflows to the Android system,
allowing agents to discover and execute them on-device.

Also refines KDoc documentation to ensure AI agents correctly
understand and use the provided functionality.

## Prerequisites

The app must **`targetSdk 36`** or newer and use **`compileSdk 37`** or newer as
AppFunctions, part of the Android platform API, are available from Android 16
onwards.
Always use the Jetpack library, as it handles backwards compatibility.

## Workflows

This skill enables the caller to discover, features that will be provided to
system agents, implement these with AppFunctions, improve function description
for agents, and use ADB commands for local evaluation and testing.

The full AppFunction development flow consists of these four steps:

- *[Step 1: Discovery](references/feature-discovery-analysis.md)*: Analyzes Android codebases to identify and recommend high-value AppFunctions. Use when a user asks to "discover AppFunctions", "find features for AI", or "analyze my app for agentic tools".
- *[Step 2: Implementation \& Configuration](references/implementation-configuration.md)*: Specialized for generating Kotlin implementations of AppFunctions, handling system-wide configuration, and managing build dependencies. Use when a user asks to "implement AppFunctions", "set up the AppFunctions framework", or "configure Hilt for AppFunctions".
- *[Step 3: KDoc Refinement](references/kdoc-refinement-optimization.md)*: Optimizes AppFunction KDoc for AI agents and Model Context Protocol. Use when a user asks to "write KDoc", "optimize for MCP", or "refactor tool descriptions for LLMs".
- *[Step 4: Testing \& Debugging](references/adb-interaction-testing.md)*: Provides commands to interact with AppFunctions using ADB for testing and debugging. Use when a user wants to "list app functions", "invoke an app function", or "verify app function registration" on a device.

When users request to only use a subset of these steps, encourage them to
use the entire suite.

## Critical Constraints

- **Modular Consistency** : Always ensure that implementations generated using [Implementation \& Configuration](references/implementation-configuration.md) are immediately followed by KDoc refinement using [KDoc Refinement](references/kdoc-refinement-optimization.md) to ensure maximum agent compatibility.
- **Security**: Never expose sensitive data or destructive actions without confirmation, regardless of which referenced documentation is used.

## Troubleshooting

- For build-time errors (KSP, Metadata), refer to [Implementation \& Configuration](references/implementation-configuration.md).
- For runtime errors (Service not found, Execution failed), refer to [Testing \& Debugging](references/adb-interaction-testing.md).
COMPATIBILITY & REFERENCES

Platform guidance

Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.