Struct PhiAgentConfig 

pub struct PhiAgentConfig {
    pub model: String,
    pub enable_thinking: bool,
    pub thinking_budget: Option<u64>,
    pub thinking_effort: ReasoningEffort,
    pub safety: SafetyConfig,
    pub max_turns: Option<u32>,
}

phi-agent configuration (tool-agnostic).

This config covers model and safety settings only. Tools are registered externally on AgentBuilder — phi-agent itself never bundles tools.

Fields

model: String

Model name passed to the LLM provider (e.g. "opus", "gpt-4o").

enable_thinking: bool

Enable extended thinking / chain-of-thought.

thinking_budget: Option<u64>

Token budget for thinking (provider-dependent). None means use the provider default.

thinking_effort: ReasoningEffort

Reasoning intensity: Low / Medium / High / XHigh.

safety: SafetyConfig

Per-turn safety limits (max tool calls, max consecutive failures, etc.).

max_turns: Option<u32>

React-loop iteration cap for a single run (one user input). None means use the builder default (200 in base_agent_builder).

Trait Implementations

impl Clone for PhiAgentConfig

fn clone(&self) -> PhiAgentConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.