Agent Skills: myfy Framework Patterns

myfy Framework Patterns

You are assisting a developer building an application with the myfy Python framework.

Core Principles

1. Opinionated, not rigid - Strong defaults, full configurability
2. Modular by design - Features are modules (web, data, tasks, frontend)
3. Type-safe DI - Constructor injection with compile-time validation
4. Async-native - Built on ASGI and AnyIO with contextvars
5. Zero-config defaults - Convention over configuration

Imports

Always import from the public API:

# Core
from myfy.core import Application, provider, SINGLETON, REQUEST, TASK
from myfy.core import BaseSettings, BaseModule

# Web
from myfy.web import WebModule, route, Query, abort, errors
from myfy.web import Authenticated, Anonymous, AuthModule
from myfy.web.ratelimit import RateLimitModule, rate_limit

# Data
from myfy.data import DataModule, AsyncSession

# Tasks
from myfy.tasks import TasksModule, task, TaskContext

# Frontend
from myfy.frontend import FrontendModule, render_template

# User
from myfy.user import UserModule

# CLI Commands
from myfy.commands import CliModule, cli

Application Structure

from myfy.core import Application
from myfy.web import WebModule, route
from myfy.data import DataModule

app = Application(
    settings_class=AppSettings,  # Optional custom settings
    auto_discover=False,          # Disable auto-discovery for explicit control
)

# Add modules in dependency order
app.add_module(DataModule())
app.add_module(WebModule())

Module Categories

| Module | Package | Purpose | |--------|---------|---------| | WebModule | myfy-web | HTTP routing, ASGI, FastAPI-like decorators | | DataModule | myfy-data | SQLAlchemy async, migrations, sessions | | FrontendModule | myfy-frontend | Jinja2, Tailwind 4, DaisyUI 5, Vite | | TasksModule | myfy-tasks | Background jobs, SQL-based task queue | | UserModule | myfy-user | Auth, OAuth, user management | | CliModule | myfy-commands | Custom CLI commands | | AuthModule | myfy-web.auth | Type-based authentication, protected routes | | RateLimitModule | myfy-web.ratelimit | Rate limiting per IP or user |

Key Conventions

1. File naming: app.py, main.py, or application.py for entry point
2. Settings: Extend BaseSettings from myfy.core
3. Routes: Use global route decorator from myfy.web
4. Providers: Use @provider(scope=...) decorator from myfy.core
5. Tasks: Use @task decorator for background jobs

Common Patterns

Creating a Route with DI

from myfy.web import route
from myfy.data import AsyncSession

@route.get("/users/{user_id}")
async def get_user(user_id: int, session: AsyncSession) -> dict:
    # session is auto-injected (REQUEST scope)
    result = await session.execute(select(User).where(User.id == user_id))
    return {"user": result.scalar_one_or_none()}

Creating a Provider

from myfy.core import provider, SINGLETON

@provider(scope=SINGLETON)
def email_service(settings: AppSettings) -> EmailService:
    return EmailService(api_key=settings.email_api_key)

Settings Class

from pydantic import Field
from pydantic_settings import SettingsConfigDict
from myfy.core import BaseSettings

class AppSettings(BaseSettings):
    app_name: str = Field(default="my-app")
    debug: bool = Field(default=False)
    database_url: str = Field(default="sqlite+aiosqlite:///./app.db")

    model_config = SettingsConfigDict(
        env_prefix="MYFY_",
        env_file=".env",
    )

Error Handling

from myfy.web import abort, errors

# Quick abort
abort(404, "User not found")

# Typed errors
raise errors.NotFound("User not found")
raise errors.BadRequest("Invalid email", field="email")

When Generating Code

Always:

• Use type hints for all function parameters and return types
• Use async/await for all handlers
• Import from myfy.* not internal modules
• Follow existing project structure
• Use Pydantic models for request/response bodies