Struct AutoContinueMiddleware 

pub struct AutoContinueMiddleware { /* private fields */ }

Middleware that automatically injects a continuation message when the LLM response is truncated by the token limit (e.g. max_tokens / length) and no tool calls are present.

This turns a silent truncation into an automatic retry: the react loop will push the follow-up as a user message and call the LLM again.

When it triggers

• finish_reason is Truncated and
• tool_calls is empty (tool-call truncation is already handled by the react-loop’s built-in truncation guard, which injects error results instead of executing incomplete tool calls).

Usage

use agent_base::engine::{AgentBuilder, auto_continue::AutoContinueMiddleware};

let runtime = AgentBuilder::new(client)
    .middleware(AutoContinueMiddleware::new())
    .build()?;

Implementations

impl AutoContinueMiddleware

pub fn new() -> Self

Create a new middleware with the default continuation prompt.

pub fn with_prompt(prompt: impl Into<String>) -> Self

Create a new middleware with a custom continuation prompt.

Trait Implementations

impl Default for AutoContinueMiddleware

fn default() -> Self

Returns the “default value” for a type.

impl Middleware for AutoContinueMiddleware

fn on_post_llm<'life0, 'life1, 'async_trait>( &'life0 self, ctx: &'life1 mut PostLlmCtx, ) -> Pin<Box<dyn Future<Output = AgentResult<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

fn on_user_message<'life0, 'life1, 'async_trait>( &'life0 self, _ctx: &'life1 mut UserMessageCtx, ) -> Pin<Box<dyn Future<Output = AgentResult<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

fn on_pre_llm<'life0, 'life1, 'async_trait>( &'life0 self, _ctx: &'life1 mut PreLlmCtx, ) -> Pin<Box<dyn Future<Output = AgentResult<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,