Struct AgentRunState 

pub struct AgentRunState {

    pub run_id: RunId,
    pub conversation_id: ConversationId,
    pub message_history: Vec<ModelMessage>,
    pub usage: Usage,
    pub run_step: usize,
    pub status: RunStatus,
    pub output: Option<String>,
    pub structured_output: Option<Value>,
    pub latest_response: Option<ModelResponse>,
    pub pending_tool_calls: Vec<ToolCallPart>,
    pub pending_tool_returns: Vec<ToolReturnPart>,
    pub pending_approval_tool_returns: Vec<ToolReturnPart>,
    pub deferred_tool_returns: Vec<ToolReturnPart>,
    pub idle_messages: Vec<String>,
    pub parent_run_id: Option<RunId>,
    pub parent_task_id: Option<TaskId>,
    pub metadata: Metadata,
}

Checkpointable state owned by the graph loop.

Fields

run_id: RunId

Run identifier.

conversation_id: ConversationId

Conversation identifier.

message_history: Vec<ModelMessage>

Canonical model history.

usage: Usage

Accumulated usage.

run_step: usize

Completed model/tool loop steps.

status: RunStatus

Current status.

output: Option<String>

Final text output.

structured_output: Option<Value>

Final structured JSON output.

latest_response: Option<ModelResponse>

Latest model response awaiting classification.

pending_tool_calls: Vec<ToolCallPart>

Tool calls awaiting execution.

pending_tool_returns: Vec<ToolReturnPart>

Tool returns awaiting request preparation.

pending_approval_tool_returns: Vec<ToolReturnPart>

Tool calls that require approval before execution can proceed.

deferred_tool_returns: Vec<ToolReturnPart>

Tool calls deferred to another runtime or durable worker.

idle_messages: Vec<String>

Idle messages ready to redirect finalization.

parent_run_id: Option<RunId>

Parent run identifier when this run is delegated from another run.

parent_task_id: Option<TaskId>

Parent-scoped delegated task identifier when this run executes a lightweight task.

metadata: Metadata

Run metadata.

Implementations

impl AgentRunState

pub fn new(run_id: RunId, conversation_id: ConversationId) -> Self

Create empty run state.

pub fn apply_model_response(&mut self, response: ModelResponse)

Apply a model response to state.

pub fn replace_latest_response(&mut self, response: ModelResponse)

Replace the latest response after lifecycle hooks mutate it.

pub const fn has_pending_hitl(&self) -> bool

Return true when the run is waiting for approval or deferred tool results.

pub fn pending_approvals(&self) -> &[ToolReturnPart]

Return pending approval-required tool returns.

pub fn pending_deferred_tools(&self) -> &[ToolReturnPart]

Return pending deferred tool returns.

pub fn pending_hitl_tool_returns(&self) -> impl Iterator<Item = &ToolReturnPart>

Iterate all pending HITL tool returns in approval-then-deferred order.

Trait Implementations

impl Clone for AgentRunState

fn clone(&self) -> AgentRunState

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for AgentRunState

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for AgentRunState

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Eq for AgentRunState

impl PartialEq for AgentRunState

fn eq(&self, other: &AgentRunState) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for AgentRunState

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for AgentRunState