`module` `nemo_fabric.models`

Pydantic SDK models for NVIDIA NeMo Fabric configuration and requests.

The Rust core remains the source of truth for persisted schema snapshots. These models provide the Python SDK’s typed authoring surface and intentionally keep extension fields so consumers can carry adapter- or application-owned data without waiting for a schema release.

`class` `FabricBaseModel`

Base class for SDK-facing Pydantic models.

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `MetadataConfig`

Human-readable agent identity.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`name`	`str`	Yes	—	`MinLen(min_length=1)`	—
`description`	`str \| None`	No	`None`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `HarnessConfig`

Harness adapter selection plus adapter-owned settings.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`adapter_id`	`str`	Yes	—	`MinLen(min_length=1)`	—
`resolution`	`Literal['preinstalled', 'image_provided', 'pip_uv', 'npm', 'source', 'service', 'native_plugin'] \| None`	No	`None`	—	—
`settings`	`dict[str, Any]`	No	`dict()`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `InstructionConfig`

One portable instruction value.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`content`	`str`	Yes	—	`MinLen(min_length=1), _PydanticGeneralMetadata(pattern='\\S')`	—
`mode`	`Literal['replace']`	No	`'replace'`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `InstructionsConfig`

Harness-neutral agent instructions.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`system`	`InstructionConfig \| None`	No	`None`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `RuntimeConfig`

Invocation runtime contract.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`input_schema`	`str \| None`	No	`None`	—	—
`output_schema`	`str \| None`	No	`None`	—	—
`artifacts`	`str \| Path \| None`	No	`None`	—	—
`timeout_seconds`	`float \| None`	No	`None`	`Gt(gt=0), _PydanticGeneralMetadata(allow_inf_nan=False)`	—
`max_turns`	`int \| None`	No	`None`	`Gt(gt=0), Le(le=4294967295)`	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `EnvironmentConfig`

Execution environment configuration supplied by the consumer.

provider selects the environment implementation. workspace is the path visible to the harness, while artifacts is the provider-specific output location. settings configures the selected provider; connection describes how NeMo Fabric reaches an existing environment; and metadata carries consumer-owned values that NeMo Fabric does not interpret. ownership identifies who tears the environment down, and control_location identifies whether NeMo Fabric control code runs inside or outside it.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`provider`	`str`	No	`'local'`	`MinLen(min_length=1)`	Environment provider, such as local, docker, opensandbox, or k8s.
`workspace`	`str \| Path \| None`	No	`None`	—	Workspace path visible to the harness.
`artifacts`	`str \| Path \| None`	No	`None`	—	Environment-specific artifact path.
`env`	`dict[str, str]`	No	`dict()`	—	Environment variables visible to the harness and its tools. Values are serialized into configuration and run plans; prefer api_key_env-style environment-variable-name indirection for credentials.
`settings`	`dict[str, Any]`	No	`dict()`	—	Provider-specific configuration interpreted by the environment provider.
`metadata`	`dict[str, Any]`	No	`dict()`	—	Consumer-owned environment metadata passed through without NeMo Fabric semantics.
`connection`	`dict[str, Any]`	No	`dict()`	—	Connection data for an existing environment, such as URL, namespace, or credential reference.
`ownership`	`Literal['caller_owned', 'fabric_owned']`	No	`'caller_owned'`	—	Whether the caller or NeMo Fabric owns environment teardown.
`control_location`	`Literal['external_control', 'in_env_control']`	No	`'in_env_control'`	—	Whether NeMo Fabric control code runs outside or inside the environment.

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `ModelConfig`

Configuration for one model role.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`provider`	`str`	Yes	—	`MinLen(min_length=1)`	—
`model`	`str`	Yes	—	`MinLen(min_length=1)`	—
`api_key_env`	`str \| None`	No	`None`	—	—
`temperature`	`float \| None`	No	`None`	—	—
`base_url`	`str \| None`	No	`None`	`MinLen(min_length=1)`	—
`settings`	`dict[str, Any]`	No	`dict()`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `SkillConfig`

Skill capability configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`paths`	`list[str \| Path]`	No	`list()`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`method` `add_path`

1 def add_path(path: str | Path) -> Self

Add a skill path if absent.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `remove_path`

1 def remove_path(path: str | Path) -> Self

Remove a skill path if present.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `McpServerConfig`

MCP server configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`transport`	`str`	Yes	—	`MinLen(min_length=1)`	—
`url`	`str`	Yes	—	`MinLen(min_length=1)`	—
`exposure`	`Literal['harness_native', 'fabric_managed']`	No	`'harness_native'`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `McpConfig`

MCP capability configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`servers`	`dict[str, McpServerConfig]`	No	`dict()`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`method` `add_server`

1 def add_server(
2     name: str,
3     *,
4     transport: str,
5     url: str,
6     exposure: Literal['harness_native', 'fabric_managed'] = 'harness_native',
7     extra_fields: Mapping[str, Any] | None = None,
8 ) -> Self

Add or replace a named MCP server.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `remove_server`

1 def remove_server(name: str) -> Self

Remove a named MCP server if present.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `RelayConfigPolicy`

NVIDIA NeMo Relay config validation policy.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`unknown_component`	`Literal['ignore', 'warn', 'error']`	No	`'warn'`	—	—
`unknown_field`	`Literal['ignore', 'warn', 'error']`	No	`'warn'`	—	—
`unsupported_value`	`Literal['ignore', 'warn', 'error']`	No	`'error'`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `RelayAtofFileSinkConfig`

NeMo Relay ATOF file sink configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`type`	`Literal['file']`	No	`'file'`	—	—
`output_directory`	`str \| Path \| None`	No	`None`	—	—
`filename`	`str \| None`	No	`None`	—	—
`mode`	`Literal['append', 'overwrite']`	No	`'append'`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `RelayAtofStreamSinkConfig`

NeMo Relay ATOF stream sink configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`type`	`Literal['stream']`	No	`'stream'`	—	—
`url`	`str`	Yes	—	—	—
`transport`	`Literal['http_post', 'websocket', 'ndjson']`	No	`'http_post'`	—	—
`headers`	`dict[str, str]`	No	`dict()`	—	—
`header_env`	`dict[str, str]`	No	`dict()`	—	—
`timeout_millis`	`int`	No	`3000`	—	—
`field_name_policy`	`Literal['preserve', 'replace_dots']`	No	`'preserve'`	—	—
`name`	`str \| None`	No	`None`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `RelayAtofConfig`

NeMo Relay ATOF export configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`enabled`	`bool`	No	`False`	—	—
`sinks`	`list[Annotated[RelayAtofFileSinkConfig \| RelayAtofStreamSinkConfig, Field(discriminator='type')] \| dict[str, Any]] \| None`	No	`None`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `RelayS3StorageConfig`

NeMo Relay ATIF S3 storage configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`type`	`Literal['s3']`	No	`'s3'`	—	—
`bucket`	`str`	No	`''`	—	—
`key_prefix`	`str \| None`	No	`None`	—	—
`access_key_id`	`str \| None`	No	`None`	—	—
`secret_access_key_var`	`str \| None`	No	`None`	—	—
`session_token_var`	`str \| None`	No	`None`	—	—
`region`	`str \| None`	No	`None`	—	—
`endpoint_url`	`str \| None`	No	`None`	—	—
`allow_http`	`bool \| None`	No	`None`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `RelayHttpStorageConfig`

NeMo Relay ATIF HTTP storage configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`type`	`Literal['http']`	No	`'http'`	—	—
`endpoint`	`str`	No	`''`	—	—
`headers`	`dict[str, str]`	No	`dict()`	—	—
`header_env`	`dict[str, str]`	No	`dict()`	—	—
`timeout_millis`	`int`	No	`3000`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `RelayAtifConfig`

NeMo Relay ATIF export configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`enabled`	`bool`	No	`False`	—	—
`agent_name`	`str`	No	`'NeMo Relay'`	—	—
`agent_version`	`str \| None`	No	`None`	—	—
`model_name`	`str`	No	`'unknown'`	—	—
`tool_definitions`	`list[dict[str, Any]] \| None`	No	`None`	—	—
`extra`	`dict[str, Any] \| None`	No	`None`	—	—
`output_directory`	`str \| Path \| None`	No	`None`	—	—
`filename_template`	`str`	No	`'nemo-relay-atif-{session_id}.json'`	—	—
`storage`	`list[Annotated[RelayS3StorageConfig \| RelayHttpStorageConfig, Field(discriminator='type')] \| dict[str, Any]] \| None`	No	`None`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `RelayOtlpConfig`

NeMo Relay OTLP export configuration for OpenTelemetry/OpenInference.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`enabled`	`bool`	No	`False`	—	—
`transport`	`Literal['http_binary', 'grpc']`	No	`'http_binary'`	—	—
`endpoint`	`str \| None`	No	`None`	—	—
`headers`	`dict[str, str]`	No	`dict()`	—	—
`resource_attributes`	`dict[str, str]`	No	`dict()`	—	—
`service_name`	`str`	No	`'nemo-relay'`	—	—
`service_namespace`	`str \| None`	No	`None`	—	—
`service_version`	`str \| None`	No	`None`	—	—
`instrumentation_scope`	`str \| None`	No	`None`	—	—
`timeout_millis`	`int`	No	`3000`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `RelayObservabilityConfig`

NeMo Relay observability component configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`version`	`int`	No	`2`	—	—
`atof`	`RelayAtofConfig \| dict[str, Any] \| None`	No	`None`	—	—
`atif`	`RelayAtifConfig \| dict[str, Any] \| None`	No	`None`	—	—
`opentelemetry`	`RelayOtlpConfig \| dict[str, Any] \| None`	No	`None`	—	—
`openinference`	`RelayOtlpConfig \| dict[str, Any] \| None`	No	`None`	—	—
`policy`	`RelayConfigPolicy \| dict[str, Any] \| None`	No	`None`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `RelayComponentConfig`

Generic NeMo Relay plugin component configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`kind`	`str`	Yes	—	`MinLen(min_length=1)`	—
`enabled`	`bool`	No	`True`	—	—
`config`	`dict[str, Any]`	No	`dict()`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `RelayConfig`

First-class NeMo Relay integration configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`project`	`str \| None`	No	`None`	—	—
`output_dir`	`str \| Path \| None`	No	`None`	—	—
`observability`	`RelayObservabilityConfig \| dict[str, Any] \| None`	No	`None`	—	—
`components`	`list[RelayComponentConfig \| dict[str, Any]]`	No	`list()`	—	—
`policy`	`RelayConfigPolicy \| dict[str, Any] \| None`	No	`None`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `TelemetryProviderConfig`

Provider-specific telemetry configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`config`	`dict[str, Any] \| None`	No	`None`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `TelemetryConfig`

Telemetry configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`providers`	`dict[Literal['relay', 'native'], TelemetryProviderConfig \| dict[str, Any]]`	No	`dict()`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`method` `enable_native`

1 def enable_native(*, config: Mapping[str, Any] | None = None) -> Self

Let the selected adapter handle telemetry natively.

`method` `enable_relay`

1 def enable_relay() -> Self

Enable NeMo Relay telemetry for subsequently started runtimes.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `remove_provider`

1 def remove_provider(provider: Literal['relay', 'native']) -> Self

Remove a configured telemetry provider.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `ToolsConfig`

Harness-neutral tool capability configuration.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`enabled`	`list[str] \| None`	No	`None`	—	Adapter-native tools to expose. None preserves the harness default; an empty list exposes no tools.
`blocked`	`list[str]`	No	`list()`	—	Adapter-native tool names to deny.

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached JSON-compatible mapping for Rust/core calls.

`class` `FabricConfig`

SDK-facing typed NeMo Fabric agent configuration.

NeMo Fabric-owned fields apply uniformly. Adapter-translated fields are checked against the selected descriptor; refer to the normalized configuration compatibility table.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`schema_version`	`str`	No	`'fabric.agent/v1alpha1'`	—	—
`metadata`	`MetadataConfig`	Yes	—	—	—
`harness`	`HarnessConfig`	Yes	—	—	—
`runtime`	`RuntimeConfig`	No	`RuntimeConfig()`	—	—
`environment`	`EnvironmentConfig \| None`	No	`None`	—	—
`models`	`dict[str, ModelConfig]`	No	`dict()`	—	—
`instructions`	`InstructionsConfig \| None`	No	`None`	—	—
`mcp`	`McpConfig \| None`	No	`None`	—	—
`skills`	`SkillConfig \| None`	No	`None`	—	—
`telemetry`	`TelemetryConfig \| None`	No	`None`	—	—
`relay`	`RelayConfig \| dict[str, Any] \| None`	No	`None`	—	—
`tools`	`ToolsConfig \| None`	No	`None`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`method` `add_mcp_server`

1 def add_mcp_server(
2     name: str,
3     *,
4     transport: str,
5     url: str,
6     exposure: Literal['harness_native', 'fabric_managed'] = 'harness_native',
7     extra_fields: Mapping[str, Any] | None = None,
8 ) -> Self

Add or replace a named MCP server and return this config.

`method` `add_skill_path`

1 def add_skill_path(path: str | Path) -> Self

Add a skill path and return this config.

`method` `block_tools`

1 def block_tools(*tools: str) -> Self

Block adapter-native tool names and return this config.

`method` `enable_relay`

1 def enable_relay(
2     *,
3     project: str | None = None,
4     output_dir: str | Path | None = None,
5     observability: RelayObservabilityConfig | Mapping[str, Any] | None = None,
6     components: Sequence[RelayComponentConfig | Mapping[str, Any]] | None = None,
7     policy: RelayConfigPolicy | Mapping[str, Any] | None = None,
8 ) -> Self

Enable NeMo Relay telemetry and return this config.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate the public agent config mapping shape.

`method` `remove_mcp_server`

1 def remove_mcp_server(name: str) -> Self

Remove a named MCP server and return this config.

`method` `remove_skill_path`

1 def remove_skill_path(path: str | Path) -> Self

Remove a skill path and return this config.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached mapping matching the Rust FabricConfig schema.

`class` `RunRequest`

One validated NeMo Fabric invocation request.

Fields

The model defines the following fields:

Field	Type	Required	Default	Constraints	Description
`input`	`Any`	No	`''`	—	—
`request_id`	`str`	No	`<generated>`	`MinLen(min_length=1)`	—
`context`	`dict[str, Any]`	No	`dict()`	—	—
`overrides`	`dict[str, Any] \| None`	No	`None`	—	—

`property` extra_fields

Return fields preserved by the extension point for this model.

`property` model_extra

Get extra fields set during validation.

Returns: A dictionary of extra fields, or None if config.extra is not set to "allow".

`property` model_fields_set

Returns the set of fields that have been explicitly set on this model instance.

Returns: A set of strings representing the fields that have been set, i.e. that were not filled from defaults.

`classmethod` `from_mapping`

1 def from_mapping(value: Mapping[str, Any]) -> Self

Validate a mapping using this Pydantic model.

`method` `to_mapping`

1 def to_mapping() -> dict[str, Any]

Return a detached request mapping for the Rust runtime.

This file was automatically generated via lazydocs.

module nemo_fabric.models

class FabricBaseModel

property extra_fields

property model_extra

property model_fields_set

classmethod from_mapping

method to_mapping

class MetadataConfig

Fields

property extra_fields

property model_extra

property model_fields_set

classmethod from_mapping

method to_mapping

class HarnessConfig

Fields

property extra_fields

property model_extra

property model_fields_set

classmethod from_mapping

method to_mapping

class InstructionConfig

Fields

property extra_fields

property model_extra

property model_fields_set

classmethod from_mapping

method to_mapping

class InstructionsConfig

Fields

property extra_fields

property model_extra

property model_fields_set

classmethod from_mapping

method to_mapping

class RuntimeConfig

Fields

property extra_fields

property model_extra

property model_fields_set

classmethod from_mapping

method to_mapping

class EnvironmentConfig

Fields

property extra_fields

property model_extra

property model_fields_set

classmethod from_mapping

method to_mapping

class ModelConfig

Fields

property extra_fields

property model_extra

property model_fields_set

classmethod from_mapping

method to_mapping

class SkillConfig

Fields

property extra_fields

property model_extra

property model_fields_set

method add_path

classmethod from_mapping

method remove_path

method to_mapping

class McpServerConfig

Fields

property extra_fields

property model_extra

property model_fields_set

classmethod from_mapping

method to_mapping

class McpConfig

Fields

property extra_fields

property model_extra

property model_fields_set

method add_server

classmethod from_mapping

method remove_server

`module` `nemo_fabric.models`

`class` `FabricBaseModel`

`property` extra_fields

`property` model_extra

`property` model_fields_set

`classmethod` `from_mapping`

`method` `to_mapping`

`class` `MetadataConfig`

`property` extra_fields

`property` model_extra

`property` model_fields_set

`classmethod` `from_mapping`

`method` `to_mapping`

`class` `HarnessConfig`

`property` extra_fields

`property` model_extra

`property` model_fields_set

`classmethod` `from_mapping`

`method` `to_mapping`

`class` `InstructionConfig`

`property` extra_fields

`property` model_extra

`property` model_fields_set

`classmethod` `from_mapping`

`method` `to_mapping`

`class` `InstructionsConfig`

`property` extra_fields

`property` model_extra

`property` model_fields_set

`classmethod` `from_mapping`

`method` `to_mapping`

`class` `RuntimeConfig`

`property` extra_fields

`property` model_extra

`property` model_fields_set

`classmethod` `from_mapping`

`method` `to_mapping`

`class` `EnvironmentConfig`

`property` extra_fields

`property` model_extra

`property` model_fields_set

`classmethod` `from_mapping`

`method` `to_mapping`

`class` `ModelConfig`

`property` extra_fields

`property` model_extra

`property` model_fields_set

`classmethod` `from_mapping`

`method` `to_mapping`

`class` `SkillConfig`

`property` extra_fields

`property` model_extra

`property` model_fields_set

`method` `add_path`

`classmethod` `from_mapping`

`method` `remove_path`

`method` `to_mapping`

`class` `McpServerConfig`

`property` extra_fields

`property` model_extra

`property` model_fields_set

`classmethod` `from_mapping`

`method` `to_mapping`

`class` `McpConfig`

`property` extra_fields

`property` model_extra

`property` model_fields_set

`method` `add_server`

`classmethod` `from_mapping`

`method` `remove_server`

`method` `to_mapping`

`class` `RelayConfigPolicy`

`property` extra_fields

`property` model_extra

`property` model_fields_set

`classmethod` `from_mapping`

`method` `to_mapping`

`class` `RelayAtofFileSinkConfig`

`property` extra_fields

`property` model_extra