| name | maaframework |
| description | Building automation pipelines with MaaFramework — JSON-based task pipelines with image/OCR recognition, multi-platform input control, and custom logic via Python/NodeJS agents |
| metadata | {"author":"kutius","version":"2026.4.16","source":"Generated from https://github.com/MaaXYZ/MaaFramework"} |
The skill is based on MaaFramework v5.x, generated at 2026-04-16.
MaaFramework (MaaFW) is a framework for automating tasks on mobile and desktop applications. It combines JSON-defined task pipelines with image recognition, OCR, neural network detection, and platform-specific input simulation. Supports Android (ADB), Windows (Win32), macOS, Linux (Wayland), gamepads, and more.
Core References
| Topic | Description | Reference |
|---|
| Architecture | Core concepts, terms, resource structure, language bindings | core-architecture |
| Pipeline Protocol | Node lifecycle, execution flow, v1/v2 formats, default properties | core-pipeline |
| Recognition Algorithms | TemplateMatch, OCR, FeatureMatch, ColorMatch, NeuralNetwork, And/Or | core-recognition |
| Action Types | Click, Swipe, Scroll, key input, Shell, Command, app control | core-actions |
Integration
| Topic | Description | Reference |
|---|
| Setup & Binding | Creating Tasker/Resource/Controller, Python/NodeJS examples, pipeline override | core-integration |
| Controller Methods | ADB/Win32/MacOS screencap & input methods, platform-specific config | core-controllers |
| Callback Protocol | Event message types, JSON structures, handling patterns | core-callbacks |
| Project Interface | interface.json — tasks, controllers, resources, options, presets, i18n | core-project-interface |
Advanced
