microsoft.opentelemetry.a365.core.models.tool_call_schema module

Typed Agent365 execute-tool argument and result schema models.

The models mirror the Agent365 JSON contract used by the .NET distro. Payloads are serialized with serialize_tool_call_payload(), which never raises: any failure is replaced by TOOL_CALL_SERIALIZATION_ERROR_JSON so a span is never orphaned and no caller exception escapes the telemetry path.

microsoft.opentelemetry.a365.core.models.tool_call_schema.TOOL_CALL_SCHEMA_VERSION = '1.0': Schema version emitted on every typed execute-tool payload.

microsoft.opentelemetry.a365.core.models.tool_call_schema.TOOL_CALL_SERIALIZATION_ERROR_JSON = '{"serialization_error":"Failed to serialize execute tool payload."}': Diagnostic payload emitted when a typed execute-tool payload cannot be serialized.

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ToolCallAction(value)[source]

Bases: str, Enum

Well-known actions taken by an execute-tool call.

CREATE = 'create': Creates a resource.

READ = 'read': Reads a resource.

UPDATE = 'update': Updates a resource.

DELETE = 'delete': Deletes a resource.

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ToolCallOutcomeStatus(value)[source]

Bases: str, Enum

Well-known statuses for an execute-tool outcome.

SUCCESS = 'success': The tool call completed successfully.

FAILURE = 'failure': The tool call failed.

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ToolPolicyDecision(value)[source]

Bases: str, Enum

Well-known policy decisions for an execute-tool result resource.

ALLOW = 'allow': The policy allows the tool call.

DENY = 'deny': The policy denies the tool call.

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ToolCallIdentifier(identifier_type=None, value=None, extension_data=<factory>)[source]

Bases: object

Resource identifier for an execute-tool payload.

identifier_type: str | None = None

value: str | None = None

extension_data: dict[str, object]

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ToolCallContainer(container_id=None, uri=None, container_type=None, extension_data=<factory>)[source]

Bases: object

Container that owns or scopes an execute-tool resource.

container_id: str | None = None

uri: str | None = None

container_type: str | None = None

extension_data: dict[str, object]

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ToolCallResource(resource_id=None, uri=None, name=None, resource_type=None, provider=None, identifiers=None, container=None, extension_data=<factory>)[source]

Bases: object

Resource referenced by an execute-tool arguments payload.

resource_id: str | None = None

uri: str | None = None

name: str | None = None

resource_type: str | None = None

provider: str | None = None

identifiers: list[ToolCallIdentifier] | None = None

container: ToolCallContainer | None = None

extension_data: dict[str, object]

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ExecuteToolCallArguments(action=None, resources=None, parameters=None, schema_version='1.0', extension_data=<factory>)[source]

Bases: object

Structured arguments for an execute-tool call.

action: ToolCallAction | str | None = None

resources: list[ToolCallResource] | None = None

parameters: dict[str, object] | None = None

schema_version: str = '1.0'

extension_data: dict[str, object]

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ToolCallResultOutcome(status=None, code=None, provider_code=None, message=None, extension_data=<factory>)[source]

Bases: object

Outcome metadata for an execute-tool result.

status: ToolCallOutcomeStatus | str | None = None

code: str | None = None

provider_code: str | None = None

message: str | None = None

extension_data: dict[str, object]

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ToolCallResultSensitivity(label_id=None, extension_data=<factory>)[source]

Bases: object

Sensitivity metadata for an execute-tool result resource.

label_id: str | None = None

extension_data: dict[str, object]

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ToolCallResultPolicy(decision=None, policy_id=None, name=None, extension_data=<factory>)[source]

Bases: object

Policy metadata for an execute-tool result resource.

decision: ToolPolicyDecision | str | None = None

policy_id: str | None = None

name: str | None = None

extension_data: dict[str, object]

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ToolCallResultSecurity(xpia_detected=None, extension_data=<factory>)[source]

Bases: object

Security metadata for an execute-tool result resource.

xpia_detected: bool | None = None

extension_data: dict[str, object]

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ToolCallResultPagination(has_more=None, next_cursor=None, total_count=None, extension_data=<factory>)[source]

Bases: object

Pagination metadata for an execute-tool result.

has_more: bool | None = None

next_cursor: str | None = None

total_count: int | None = None

extension_data: dict[str, object]

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ToolCallResultResource(resource_id=None, uri=None, name=None, resource_type=None, provider=None, identifiers=None, container=None, outcome=None, sensitivity=None, policy=None, security=None, data=None, extension_data=<factory>)[source]

Bases: object

Resource included in an execute-tool result payload.

resource_id: str | None = None

uri: str | None = None

name: str | None = None

resource_type: str | None = None

provider: str | None = None

identifiers: list[ToolCallIdentifier] | None = None

container: ToolCallContainer | None = None

outcome: ToolCallResultOutcome | None = None

sensitivity: ToolCallResultSensitivity | None = None

policy: ToolCallResultPolicy | None = None

security: ToolCallResultSecurity | None = None

data: dict[str, object] | None = None

extension_data: dict[str, object]

class microsoft.opentelemetry.a365.core.models.tool_call_schema.ExecuteToolCallResult(outcome=None, resources=None, data=None, pagination=None, schema_version='1.0', extension_data=<factory>)[source]

Bases: object

Structured result for an execute-tool call.

outcome: ToolCallResultOutcome | None = None

resources: list[ToolCallResultResource] | None = None

data: dict[str, object] | None = None

pagination: ToolCallResultPagination | None = None

schema_version: str = '1.0'

extension_data: dict[str, object]

microsoft.opentelemetry.a365.core.models.tool_call_schema.serialize_tool_call_payload(value)[source]

Serialize a typed execute-tool payload using the Agent365 JSON contract.

The call never raises. When any value in the payload cannot be represented in the schema — an unsupported type, a reference cycle, a non-finite float, an undefined enum value, or invalid extension data — the whole payload is replaced by TOOL_CALL_SERIALIZATION_ERROR_JSON.

Parameters:: value (ExecuteToolCallArguments | ExecuteToolCallResult | None) – The typed payload to serialize, or None.
Return type:: str | None
Returns:: The JSON payload, the diagnostic payload on failure, or None when value is None.