Trait Middleware 

pub trait Middleware: Send + Sync {
    // Required method
    fn name(&self) -> &'static str;

    // Provided methods
    fn incompatible_middleware(&self) -> &'static [&'static str] { ... }
    fn context_limit(&self, _context_window: i64) -> Option<i64> { ... }
    fn register(
        &self,
        _catalog: &mut Catalog,
        _runtime: &RuntimeContext,
    ) -> Result<()> { ... }
    fn prompt_section(
        &self,
        _runtime: &RuntimeContext,
    ) -> Result<Option<PromptSection>> { ... }
    fn frontend(&self, _session_id: &str) -> FrontendContribution { ... }
    fn render(
        &self,
        _event: &EventMsg,
        _session_id: &str,
    ) -> Option<FrontendBlock> { ... }
    fn command<'a>(
        &'a self,
        context: MiddlewareCommandContext<'a>,
    ) -> BoxFuture<'a, Result<MiddlewareCommandOutput>> { ... }
    fn session_start<'a>(
        &'a self,
        _context: &'a mut SessionStartContext<'_>,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn handles_messages(&self) -> bool { ... }
    fn route_message<'a>(
        &'a self,
        _context: &'a mut MessageRouteContext<'_>,
    ) -> BoxFuture<'a, Result<SubmissionResult>> { ... }
    fn message_boundary_events(&self, _submission_id: &str) -> Vec<EventMsg> { ... }
    fn message_submit<'a>(
        &'a self,
        _context: &'a mut MessageSubmitContext<'_>,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn active_command<'a>(
        &'a self,
        _context: &'a mut ActiveCommandContext<'_>,
    ) -> BoxFuture<'a, Result<Option<SubmissionResult>>> { ... }
    fn pre_model<'a>(
        &'a self,
        _context: &'a mut ModelContext<'_>,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn model_request<'a>(
        &'a self,
        _context: &'a mut ModelRequestContext<'_>,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn tool_exposure<'a>(
        &'a self,
        _context: &'a mut ToolExposureContext<'_>,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn pre_tool_use<'a>(
        &'a self,
        _context: &'a mut PreToolUseContext<'_>,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn permission_request<'a>(
        &'a self,
        _context: &'a mut PermissionRequestContext<'_>,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn post_tool_use<'a>(
        &'a self,
        _context: &'a mut PostToolUseContext<'_>,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn pre_compact<'a>(
        &'a self,
        _context: &'a mut CompactContext<'_>,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn prepare_compacted_input(
        &self,
        _original: ModelInput<'_>,
        _compacted: &mut Vec<Arc<Value>>,
    ) { ... }
    fn post_compact<'a>(
        &'a self,
        _context: &'a mut CompactContext<'_>,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn stop<'a>(
        &'a self,
        _context: &'a mut StopContext<'_>,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn turn_end<'a>(
        &'a self,
        _context: &'a mut TurnEndContext<'_>,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn session_end<'a>(
        &'a self,
        _runtime: &'a RuntimeContext,
    ) -> BoxFuture<'a, Result<()>> { ... }
    fn main_agent_only(&self) -> bool { ... }
    fn has_background_work<'a>(
        &'a self,
        _session_id: &'a str,
    ) -> BoxFuture<'a, Result<bool>> { ... }
}

A capability contribution to the single ordered agent pipeline.

Required Methods

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

Stable ID used to reject duplicate registrations.

Provided Methods

fn incompatible_middleware(&self) -> &'static [&'static str]

Middleware IDs that cannot share a stack with this configured capability.

fn context_limit(&self, _context_window: i64) -> Option<i64>

Returns the input-token limit this capability enforces for a model context window.

fn register( &self, _catalog: &mut Catalog, _runtime: &RuntimeContext, ) -> Result<()>

Adds tools to the catalog while the agent is created.

Errors

Returns an error if validation or an operation required by this function fails.

fn prompt_section( &self, _runtime: &RuntimeContext, ) -> Result<Option<PromptSection>>

Contributes one immutable system-prompt section while the agent is created.

Errors

Returns an error if validation or an operation required by this function fails.

fn frontend(&self, _session_id: &str) -> FrontendContribution

Declares commands and status data that any frontend may render for one session.

fn render(&self, _event: &EventMsg, _session_id: &str) -> Option<FrontendBlock>

Renders an event owned by this capability for the destination session.

Session-bound handles must only be exposed when they belong to session_id.

fn command<'a>( &'a self, context: MiddlewareCommandContext<'a>, ) -> BoxFuture<'a, Result<MiddlewareCommandOutput>>

Handles a command declared by this middleware’s frontend contribution.

fn session_start<'a>( &'a self, _context: &'a mut SessionStartContext<'_>, ) -> BoxFuture<'a, Result<()>>

Starts or re-enters a session lifecycle.

fn handles_messages(&self) -> bool

Reports whether this middleware is the session’s conversation-message preparer.

fn route_message<'a>( &'a self, _context: &'a mut MessageRouteContext<'_>, ) -> BoxFuture<'a, Result<SubmissionResult>>

Validates and prepares one conversation message for its lifecycle boundary.

Errors

Returns an error if validation or an operation required by this function fails.

fn message_boundary_events(&self, _submission_id: &str) -> Vec<EventMsg>

Produces capability UI cleanup when one queued message reaches its boundary.

fn message_submit<'a>( &'a self, _context: &'a mut MessageSubmitContext<'_>, ) -> BoxFuture<'a, Result<()>>

Intercepts a prepared conversation message at its delivery boundary.

fn active_command<'a>( &'a self, _context: &'a mut ActiveCommandContext<'_>, ) -> BoxFuture<'a, Result<Option<SubmissionResult>>>

Handles a capability command while a turn is active.

The active model, tool, or hook future is not polled until this returns. Implementations must keep work bounded and must not await a resource held by that active future. Return None when the command is unavailable during the active turn.

fn pre_model<'a>( &'a self, _context: &'a mut ModelContext<'_>, ) -> BoxFuture<'a, Result<()>>

Mutates durable context before the next model request is assembled.

fn model_request<'a>( &'a self, _context: &'a mut ModelRequestContext<'_>, ) -> BoxFuture<'a, Result<()>>

Applies request-only context after every durable transform has completed.

fn tool_exposure<'a>( &'a self, _context: &'a mut ToolExposureContext<'_>, ) -> BoxFuture<'a, Result<()>>

Hides registered tools whose capability is unavailable at this boundary.

fn pre_tool_use<'a>( &'a self, _context: &'a mut PreToolUseContext<'_>, ) -> BoxFuture<'a, Result<()>>

Intercepts one normalized tool call before authorization and persistence.

fn permission_request<'a>( &'a self, _context: &'a mut PermissionRequestContext<'_>, ) -> BoxFuture<'a, Result<()>>

Decides whether an approval request should be deferred, allowed, or denied.

fn post_tool_use<'a>( &'a self, _context: &'a mut PostToolUseContext<'_>, ) -> BoxFuture<'a, Result<()>>

Transforms model-visible feedback after a tool has executed.

fn pre_compact<'a>( &'a self, _context: &'a mut CompactContext<'_>, ) -> BoxFuture<'a, Result<()>>

Intercepts context immediately before compaction and may stop the active turn.

fn prepare_compacted_input( &self, _original: ModelInput<'_>, _compacted: &mut Vec<Arc<Value>>, )

Repairs or discards capability-owned projections before compacted input is committed. The original input is read-only; changes are validated before the rewrite is saved.

fn post_compact<'a>( &'a self, _context: &'a mut CompactContext<'_>, ) -> BoxFuture<'a, Result<()>>

Intercepts the committed compacted context and may stop the active turn.

fn stop<'a>( &'a self, _context: &'a mut StopContext<'_>, ) -> BoxFuture<'a, Result<()>>

Decides whether a naturally stopped model should continue once more.

fn turn_end<'a>( &'a self, _context: &'a mut TurnEndContext<'_>, ) -> BoxFuture<'a, Result<()>>

Observes terminal bookkeeping for a completed or aborted turn.

fn session_end<'a>( &'a self, _runtime: &'a RuntimeContext, ) -> BoxFuture<'a, Result<()>>

Releases session-local state when the agent runtime stops.

fn main_agent_only(&self) -> bool

Keeps this capability out of child agents created from the root’s configuration.

fn has_background_work<'a>( &'a self, _session_id: &'a str, ) -> BoxFuture<'a, Result<bool>>

Reports work this capability still runs for the session outside an active turn.

Dyn Compatibility

This trait is dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementors

impl Middleware for Artifacts

impl Middleware for Attachments

impl Middleware for Compaction

impl Middleware for ComputerControl

impl Middleware for Extensions

impl Middleware for ImageGeneration

impl Middleware for Instructions

impl Middleware for Messages

impl Middleware for Questions

impl Middleware for Sandbox

impl Middleware for Scratchpad

impl Middleware for Sessions

impl Middleware for Subagents

impl Middleware for Tasks

impl Middleware for Tools

impl Middleware for Voice