Struct RunWorkflowTool 

pub struct RunWorkflowTool { /* private fields */ }

The run_workflow tool: the agent picks a recipe by name + args.

Implementations

impl RunWorkflowTool

pub fn new(registry: WorkflowRegistry, provider: Arc<BoxedProvider>) -> Self

Build the tool over a registry and the shared provider used to run the recipe’s flow/ agents.

pub fn with_agent_events(self, sink: Arc<OnEvent>) -> Self

Forward every recipe-internal agent’s event stream (tool calls, LLM responses, run lifecycle) to sink — e.g. the TUI trace. Without this, a recipe run is a single opaque tool call to any observer.

pub fn with_provider_factory(self, factory: Arc<ProviderFactory>) -> Self

Install the per-call model resolver threaded into every recipe ctx — lets recipes run stages on role-named models (.model("fast")).

pub fn with_workspace(self, root: PathBuf) -> Self

Workspace (git repo root) threaded into every recipe ctx — required for recipes whose agents ask for Isolation::Worktree.

pub fn with_journal_dir(self, dir: PathBuf) -> Self

Enable resume: each run journals its agent outputs under dir, keyed by recipe + canonicalized args. Re-invoking the SAME call replays the completed agents at zero cost and continues the rest — an interrupted workflow picks up where it left off (scope the dir per session to avoid stale cross-session replays).

Trait Implementations

impl Tool for RunWorkflowTool

fn definition(&self) -> ToolDefinition

Return the tool’s definition (name, description, input schema) for the LLM.

fn execute( &self, _ctx: &ExecutionContext, input: Value, ) -> Pin<Box<dyn Future<Output = Result<ToolOutput, Error>> + Send + '_>>

Execute the tool with the LLM-supplied input JSON.

fn redact_for_history(&self, output: &str) -> String

Optional override: produce a compact version of the tool output suitable for inclusion in conversation history sent to subsequent LLM turns. The default implementation returns the input verbatim.