Standards for generated Python application code, including project structure, dependency management, and Azure SDK patterns.

Domain: application

---

Check	Description
STAN-PY-001	Use Azure SDK with DefaultAzureCredential: Always use DefaultAzureCredential from azure-identity for authenticating to Azure services. This works with managed identity in Azure and developer credentials locally.
STAN-PY-002	Project Structure: Python projects must have a clear structure: src/ for application code, tests/ for tests, requirements.txt or pyproject.toml for dependencies. Use packages (directories with __init__.py) for logical grouping.
STAN-PY-003	Configuration via Environment Variables: Use environment variables for all configuration (loaded via os.environ or a settings class). Never hardcode service URLs, ports, or feature flags.
STAN-PY-004	Async for I/O-Bound Operations: Use async/await for HTTP handlers, database queries, and external API calls. Synchronous I/O blocks the event loop and reduces throughput.
STAN-PY-005	Type Annotations: Use type annotations on all function signatures and return types. This enables IDE support, documentation, and static analysis.

---

Use Azure SDK with DefaultAzureCredential: Always use DefaultAzureCredential from azure-identity for authenticating to Azure services. This works with managed identity in Azure and developer credentials locally.

Rationale: DefaultAzureCredential provides seamless authentication across local dev and managed identity in production.
Agents: python-developer

• from azure.identity import DefaultAzureCredential
• credential = DefaultAzureCredential()
• Never pass connection strings when managed identity is available

---

Project Structure: Python projects must have a clear structure: src/ for application code, tests/ for tests, requirements.txt or pyproject.toml for dependencies. Use packages (directories with init.py) for logical grouping.

Rationale: Pinned dependencies ensure reproducible builds and prevent unexpected breaking changes.
Agents: python-developer

• src/app/ — application package
• src/app/main.py — entry point
• src/app/routes/ — API route handlers
• src/app/services/ — business logic
• tests/ — test files

---

Configuration via Environment Variables: Use environment variables for all configuration (loaded via os.environ or a settings class). Never hardcode service URLs, ports, or feature flags.

Rationale: Parameterized configurations allow reuse across environments and prevent hardcoded values.
Agents: python-developer

• import os; port = int(os.environ.get('PORT', 8080))
• Use pydantic Settings class for type-safe config

---

Async for I/O-Bound Operations: Use async/await for HTTP handlers, database queries, and external API calls. Synchronous I/O blocks the event loop and reduces throughput.

Rationale: Following established standards ensures consistency, quality, and maintainability across the codebase.
Agents: python-developer

• async def get_items(db: AsyncSession) -> list[Item]:
• Use aiohttp or httpx for async HTTP clients

---

Type Annotations: Use type annotations on all function signatures and return types. This enables IDE support, documentation, and static analysis.

Rationale: Following established standards ensures consistency, quality, and maintainability across the codebase.
Agents: python-developer

• def get_user(user_id: str) -> User | None:
• async def create_order(order: OrderCreate) -> OrderResponse:

---