Struct AgentRunOutput 

pub struct AgentRunOutput {
    pub answer: String,
    pub history: Vec<AgentMessage>,
    pub total_iterations: usize,
    pub total_completion_tokens: Option<u64>,
}

Output returned by a successful AgentLoopPort::run invocation.

Using a named struct instead of a bare tuple keeps call sites self-documenting and allows new fields to be added without breaking existing destructures.

Fields

answer: String

The final answer text produced by the agent.

history: Vec<AgentMessage>

Full accumulated conversation history: the caller-supplied messages plus every assistant and tool-result message appended during the loop, including the final assistant reply.

Safe to pass directly as the messages argument for the next turn.

total_iterations: usize

Number of loop iterations consumed before the agent produced its final answer. Always ≥ 1. Useful for logging and telemetry.

total_completion_tokens: Option<u64>

Total completion tokens generated across every iteration, summed from each response’s trailing usage report.

None when no upstream response reported usage — distinct from zero, which would be a real (if strange) measurement. Feeds the benchmark speed axis (completion tokens over wall-clock time).

Trait Implementations

impl Debug for AgentRunOutput

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

Formats the value using the given formatter.