Which state management does this template assume?
None in particular; it asks the agent to use what the project already uses. Name it in the notes, for example Riverpod or Bloc, so the agent knows.
A cross-platform Flutter app written in Dart. The rules keep widgets small and const where possible, keep state management consistent, and make flutter analyze and the test suite the definition of done.
# AGENTS.md
Project: flutter_app. A cross-platform Flutter app in Dart for iOS and Android.
This file is the single source of truth for every coding agent working in this repository. The tool-specific files (CLAUDE.md, GEMINI.md, .github/copilot-instructions.md and .cursor/rules/project.mdc) point here. Edit this file, not those.
## Stack
- Language: Dart
- Framework: Flutter
- Database: Firebase / Firestore
- Hosting: App Store and Google Play
- Tests: flutter test
- Package manager: pub (Dart)
## Commands
Run these from the repository root.
- Install dependencies: `flutter pub get`
- Run locally: `flutter run`
- Build: `flutter build apk --debug`
- Run all tests: `flutter test`
- Run one test file: `flutter test test/widgets/cart_test.dart`
- Lint: `flutter analyze`
- Format: `dart format .`
## Project layout
- `lib/`: Application code
- `lib/features/`: Feature folders: screens, widgets, state
- `lib/core/`: Shared services, theme and routing
- `test/`: Unit and widget tests
- `pubspec.yaml`: Dependencies and assets
## Code style
- Indent with 2 spaces. Never tabs.
- Match the style and patterns of the surrounding code, even where you would personally choose differently.
- Prefer small pure functions and immutable data; keep side effects at the edges (handlers, jobs, I/O modules).
- Never swallow an error. Handle it where you can act on it, or rethrow it with context about what failed.
- Every new UI control works with a keyboard, has a visible focus state and an accessible name.
## Working rules
- Before you say a task is done, run the test, lint and type-check commands above and make sure they pass. If you could not run them, say so.
- Never write secrets, API keys, tokens or passwords into code, tests, fixtures, logs or commits. Read them from environment variables, and never commit `.env` files.
- Do not add a new dependency without asking first. Prefer the standard library and what is already installed.
- Do not hand-edit generated files or lockfiles; change the source and regenerate.
- Change only what the task needs. Do not refactor, rename or reformat unrelated code.
## Product constraints
- Design and test every screen at phone width first, then scale up.
## Project-specific notes
- `flutter analyze` must report no issues, and code must be `dart format` clean.
- Use `const` constructors wherever possible.
- Split a widget when its `build` method grows beyond what fits on one screen; prefer small `StatelessWidget` classes over helper methods that return widgets.
- Use the state management approach the project already uses everywhere; do not introduce a second one.
- Never use `BuildContext` across an `await` without checking `context.mounted` first.
- Do not edit generated files (`*.g.dart`, `*.freezed.dart`); run `dart run build_runner build --delete-conflicting-outputs`.
- Add dependencies with `flutter pub add <pkg>` and ask before adding any.
## Project phase
This project is in active development. Follow the patterns that already exist; consistency beats novelty.
# CLAUDE.md
@AGENTS.md
## Claude Code
- The shared project instructions are imported from AGENTS.md on the line above. Edit AGENTS.md, not this file, so every agent stays in sync.
- Put Claude-only preferences below this line. Personal, uncommitted notes belong in CLAUDE.local.md.
# GEMINI.md
@./AGENTS.md
## Gemini CLI
- The shared project instructions are imported from AGENTS.md on the line above. Edit AGENTS.md, not this file.
- Alternative: delete this file and set "context": { "fileName": ["AGENTS.md"] } in .gemini/settings.json so Gemini CLI reads AGENTS.md directly.
# Copilot instructions
Project: flutter_app. A cross-platform Flutter app in Dart for iOS and Android.
The full instructions are in AGENTS.md at the repository root. Read it before you change code; if anything here disagrees with it, AGENTS.md wins. The essentials are repeated below for Copilot surfaces that only read this file.
## Stack
- Language: Dart
- Framework: Flutter
- Database: Firebase / Firestore
- Hosting: App Store and Google Play
- Tests: flutter test
- Package manager: pub (Dart)
## Commands
- Install dependencies: `flutter pub get`
- Run locally: `flutter run`
- Build: `flutter build apk --debug`
- Run all tests: `flutter test`
- Run one test file: `flutter test test/widgets/cart_test.dart`
- Lint: `flutter analyze`
- Format: `dart format .`
## Working rules
- Before you say a task is done, run the test, lint and type-check commands above and make sure they pass. If you could not run them, say so.
- Never write secrets, API keys, tokens or passwords into code, tests, fixtures, logs or commits. Read them from environment variables, and never commit `.env` files.
- Do not add a new dependency without asking first. Prefer the standard library and what is already installed.
- Do not hand-edit generated files or lockfiles; change the source and regenerate.
- Change only what the task needs. Do not refactor, rename or reformat unrelated code.
---
description: Project instructions for every task (mirrors AGENTS.md)
globs:
alwaysApply: true
---
# Project rules
The project instructions live in @AGENTS.md at the repository root, which Cursor loads automatically. They are not repeated here, so they cost context only once. Edit AGENTS.md, not this rule.
Before you change code, check the Commands and Working rules sections of AGENTS.md, and run its test and lint commands before you say a task is done.
Put Cursor-only rules below, or add scoped rules as more .mdc files with `globs` (for example `src/components/**/*.tsx`).
These are the exact files the generator writes for this preset. Change any command or rule in the generator and download your own set.
The preset describes a a cross-platform Flutter app in Dart for iOS and Android. Its AGENTS.md is 49 non-empty lines with 7 exact commands, 5 key directories, and 7 stack-specific rules on top of the shared working rules.
flutter pub get, flutter run, flutter build apk --debug, flutter test, flutter test test/widgets/cart_test.dart, flutter analyze, dart format ..lib/, lib/features/, lib/core/, test/, pubspec.yaml.Agents extract parts of a build method into helper functions returning widgets. Separate widget classes rebuild independently and can be const, which matters for performance and readability.
Using BuildContext after an async gap can reference a widget that has been disposed, causing crashes. The analyzer flags it and agents often suppress the lint.
Flutter has many options. Agents pick whichever their examples used, and a codebase can end up with three. Naming the rule keeps the app consistent.
Code generation output is overwritten on every build_runner run, so manual fixes do not survive.
flutter analyze must report no issues, and code must be dart format clean.const constructors wherever possible.build method grows beyond what fits on one screen; prefer small StatelessWidget classes over helper methods that return widgets.BuildContext across an await without checking context.mounted first.*.g.dart, *.freezed.dart); run dart run build_runner build --delete-conflicting-outputs.flutter pub add <pkg> and ask before adding any.Want different commands or rules? Load this preset into the generator, change anything, and download your own set.
AGENTS.md is read directly by Codex, Cursor, Windsurf, Cline, Zed and Jules, and by GitHub Copilot's cloud agent, code review and CLI. Claude Code reads it through the generated CLAUDE.md, and Gemini CLI through GEMINI.md. See the compatibility table for details, last verified 2026-11-13.
All presets are on the stacks page.
None in particular; it asks the agent to use what the project already uses. Name it in the notes, for example Riverpod or Bloc, so the agent knows.
flutter analyze performs Dart static analysis, including type checking, so it covers both lint and types.
Yes, the rules are platform-neutral. Change the build command and hosting choice in the generator.
Yes. Copy them into any project, commercial or not, without attribution. They are a starting point: adjust the commands and rules to match your repository.
You do not have to take our word for it. Developer tools show every request a page makes, so you can watch this page work and confirm that nothing you type is sent. The check takes under a minute.
F12, or Cmd+Option+I on a Mac, and select the Network tab.Only the feedback form can add a line to that log, and it sends only the vote and comment you type into it. 2prompt has no server code at all: no API, no model and no database of your files. The production site also sends a Content-Security-Policy header that tells your browser to refuse connections to anywhere except this site, the fonts and the anonymous analytics endpoint.