Struct FactoryRuntime 

pub struct FactoryRuntime { /* private fields */ }

Everything a tool call needs, wired to a real factory.

Owns rather than borrows, because this has to live in an HTTP handler that outlives any particular call and is shared across threads.

Implementations

impl FactoryRuntime

pub fn new(wiring: Wiring) -> Self

Wires a runtime to a factory definition and the places it keeps things.

Trait Implementations

impl Runtime for FactoryRuntime

fn peers(&self, session: &Session) -> Vec<Peer>

Who session’s agent may send to.

fn send( &self, session: &Session, to: &AgentName, body: &str, ) -> Result<String, ToolError>

Sends work to another agent, returning the flight’s identifier.

fn report( &self, session: &Session, headline: &str, body: &str, ) -> Result<(), ToolError>

Records what this run concluded.

fn help( &self, session: &Session, summary: &str, detail: &str, fatal: bool, ) -> Result<(), ToolError>

Records that this run could not get past something.

fn memory_read(&self, session: &Session) -> Result<String, ToolError>

Reads this agent’s own notes.

fn memory_write(&self, session: &Session, text: &str) -> Result<(), ToolError>

Adds to this agent’s own notes.

fn wait( &self, session: &Session, until: &str, because: &str, ) -> Result<String, ToolError>

Sets this work down to be picked up later, returning when it will be.

fn learn(&self, session: &Session, text: &str) -> Result<String, ToolError>

Proposes something future runs of this agent should know.

fn logbook_append(&self, session: &Session, text: &str) -> Result<(), ToolError>

Adds a line to the factory’s shared memory, which every agent can read.