Struct DeepAgentPromptMiddleware

pub struct DeepAgentPromptMiddleware { /* private fields */ }

Deep Agent prompt middleware that injects comprehensive tool usage instructions and examples to force the LLM to actually call tools instead of just talking about them.

This middleware is inspired by Python’s deepagents package and Claude Code’s system prompt. It provides:

• Explicit tool usage rules with imperative language
• JSON examples of tool calling
• Workflow guidance for multi-step tasks
• Few-shot examples for common patterns

Implementations

impl DeepAgentPromptMiddleware

pub fn new(custom_instructions: impl Into<String>) -> Self

Trait Implementations

impl AgentMiddleware for DeepAgentPromptMiddleware

fn id(&self) -> &'static str

Unique identifier for logging and diagnostics.

fn modify_model_request<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, ctx: &'life1 mut MiddlewareContext<'life2>, ) -> Pin<Box<dyn Future<Output = Result<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Apply middleware-specific mutations to the pending model request.

fn tools(&self) -> Vec<ToolBox>

Tools to expose when this middleware is active.

fn before_tool_execution<'life0, 'life1, 'life2, 'life3, 'async_trait>( &'life0 self, _tool_name: &'life1 str, _tool_args: &'life2 Value, _call_id: &'life3 str, ) -> Pin<Box<dyn Future<Output = Result<Option<AgentInterrupt>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait, 'life3: 'async_trait,

Hook called before tool execution - can return an interrupt to pause execution.