I think I'm missing something since I can confirm that the

.env

is loaded when running tests:

sdk/singer_sdk/testing/factory.py

Lines 99 to 100 in</meltano/sdk/commit/d1ba40003a516c8d0c42b9fd81ca8a826de3d392|d1ba400>

| if "parse_env_config" not in kwargs: |

| ------------------------------------ |

| kwargs["parse_env_config"] = True |

and then

sdk/singer_sdk/plugin_base.py

Lines 191 to 193 in</meltano/sdk/commit/d1ba40003a516c8d0c42b9fd81ca8a826de3d392|d1ba400>

| if parse_env_config: |

| ---------------------------------------------------------- |

| self.logger.info("Parsing env var for settings config...") |

| config_dict.update(self._env_var_config) |

and

sdk/singer_sdk/plugin_base.py

Lines 271 to 284 in</meltano/sdk/commit/d1ba40003a516c8d0c42b9fd81ca8a826de3d392|d1ba400>

| @classproperty |

| ----------------------------------------------------------------------- |

| def _env_var_config(cls) -> dict[str, t.Any]: # noqa: N805 |

| """Return any config specified in environment variables. |

| |

| Variables must match the convention "<PLUGIN_NAME>_<SETTING_NAME>", |

| all uppercase with dashes converted to underscores. |

| |

| Returns: |

| Dictionary of configuration parsed from the environment. |

| """ |

| config_jsonschema = cls.config_jsonschema |

| cls.append_builtin_config(config_jsonschema) |

| |

| return parse_environment_config(config_jsonschema, cls._env_var_prefix) |

If you run pytest with the

-s

flag, you should be able to see something like

```

2025-01-21 110925,719 | INFO | singer_sdk.configuration._dict_config | Parsing 'account_id' config from env variable 'TAP_CHECKLY_ACCOUNT_ID'.

2025-01-21 110925,719 | INFO | singer_sdk.configuration._dict_config | Parsing 'token' config from env variable 'TAP_CHECKLY_TOKEN'.

```

even though the env vars are not set in the shell where this is running.