Struct MicroAgent 

pub struct MicroAgent<Ctx> {
    pub history: Vec<Message>,
    pub tools: HashMap<String, Arc<dyn ToolFunction<Ctx>>>,
    pub skills: HashMap<String, String>,
    pub provider: SupportedProvider,
    pub model: String,
    pub system: String,
    pub tool_context: Arc<ToolExecutionContext<Ctx>>,
    pub storage: Box<dyn AgentStorage>,
    pub parallel_tool_calls: bool,
    /* private fields */
}

A fully-configured agent ready to generate responses or run conversations.

Created via MicroAgentBuilder. Holds the conversation history, tool registry, and LLM client configuration.

Fields

history: Vec<Message>

tools: HashMap<String, Arc<dyn ToolFunction<Ctx>>>

skills: HashMap<String, String>

provider: SupportedProvider

model: String

system: String

tool_context: Arc<ToolExecutionContext<Ctx>>

storage: Box<dyn AgentStorage>

parallel_tool_calls: bool

Trait Implementations

impl<Ctx: Send + Sync + 'static> Agent for MicroAgent<Ctx>

fn generate<'life0, 'async_trait>( &'life0 mut self, ) -> Pin<Box<dyn Future<Output = Result<GenerationStream, AgentError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Generate the next assistant response as a raw token stream.

The stream yields [StreamChunk]s that may contain text deltas or partial tool calls. Higher-level orchestration (e.g. [run]) is responsible for buffering and acting on tool calls.

fn run<'async_trait>( self, prompt: String, session_id: Option<String>, ) -> Pin<Box<dyn Future<Output = Result<RunStream, AgentError>> + Send + 'async_trait>>

where Self: 'async_trait,

Run a complete conversation turn.

If session_id is Some the conversation history is restored from storage; otherwise a new session is created. The returned stream yields high-level events (AgentEventAny) including deltas, tool calls, results, and the final stop event.

impl<Ctx: Debug> Debug for MicroAgent<Ctx>

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

Formats the value using the given formatter.