pydantic-settings

Add a typed field with a sensible default to the Settings class in core/config.py, add the matching UPPER_SNAKE_CASE line to env.jinja, then import the settings singleton wherever the value is needed. BaseSettings maps the field to its env var automatically.

A snake_case field on a BaseSettings subclass maps to its UPPER_SNAKE_CASE environment variable by default, so database_url reads DATABASE_URL with no explicit alias. Real environment variables take priority, then the .env file, then the field's Python default.

A field declared without a default has no corresponding environment variable or .env entry, so pydantic validation fails when Settings() is constructed at import time. The error names the offending field; add the env var or give the field a default.

Use a single module-level settings = Settings() instance imported everywhere. Building it is pure and side-effect-free, so eager construction at import time is safe, unlike database engines or broker connections that must be deferred to a lifespan hook.

In this project the generator mirrors whatever env.jinja renders into both .env and the checked-in .env.example at generation time. After generation nothing syncs them automatically, so add new variables to env.jinja rather than hand-editing one file.