Enum BusMessage 

pub enum BusMessage {
    AgentReady {
        agent_id: String,
        capabilities: Vec<String>,
    },
    AgentShutdown {
        agent_id: String,
    },
    AgentMessage {
        from: String,
        to: String,
        parts: Vec<Part>,
    },
    TaskUpdate {
        task_id: String,
        state: TaskState,
        message: Option<String>,
    },
    ArtifactUpdate {
        task_id: String,
        artifact: Artifact,
    },
    SharedResult {
        key: String,
        value: Value,
        tags: Vec<String>,
    },
    ToolRequest {
        request_id: String,
        agent_id: String,
        tool_name: String,
        arguments: Value,
    },
    ToolResponse {
        request_id: String,
        agent_id: String,
        tool_name: String,
        result: String,
        success: bool,
    },
    Heartbeat {
        agent_id: String,
        status: String,
    },
    RalphLearning {
        prd_id: String,
        story_id: String,
        iteration: usize,
        learnings: Vec<String>,
        context: Value,
    },
    RalphHandoff {
        prd_id: String,
        from_story: String,
        to_story: String,
        context: Value,
        progress_summary: String,
    },
    RalphProgress {
        prd_id: String,
        passed: usize,
        total: usize,
        iteration: usize,
        status: String,
    },
    ToolOutputFull {
        agent_id: String,
        tool_name: String,
        output: String,
        success: bool,
        step: usize,
    },
    AgentThinking {
        agent_id: String,
        thinking: String,
        step: usize,
    },
    VoiceSessionStarted {
        room_name: String,
        agent_id: String,
        voice_id: String,
    },
    VoiceTranscript {
        room_name: String,
        text: String,
        role: String,
        is_final: bool,
    },
    VoiceAgentStateChanged {
        room_name: String,
        state: String,
    },
    VoiceSessionEnded {
        room_name: String,
        reason: String,
    },
}

The set of messages the bus can carry.

Variants

AgentReady

An agent has come online and is ready to receive work

Fields

agent_id: String

capabilities: Vec<String>

AgentShutdown

An agent is shutting down

Fields

agent_id: String

AgentMessage

Free-form message from one agent to another (mirrors A2A Message)

Fields

from: String

to: String

parts: Vec<Part>

TaskUpdate

Task status changed

Fields

task_id: String

state: TaskState

message: Option<String>

ArtifactUpdate

A new artifact was produced for a task

Fields

task_id: String

artifact: Artifact

SharedResult

A sub-agent published a shared result (replaces raw ResultStore publish)

Fields

key: String

value: Value

tags: Vec<String>

ToolRequest

Tool execution request (for shared tool dispatch)

Fields

request_id: String

agent_id: String

tool_name: String

arguments: Value

ToolResponse

Tool execution response

Fields

request_id: String

agent_id: String

tool_name: String

result: String

success: bool

Heartbeat

Heartbeat (keep-alive / health signal)

Fields

agent_id: String

status: String

RalphLearning

An agent shares learnings from a Ralph iteration (insights, blockers, patterns discovered) so subsequent iterations or co-ordinating agents can build on them.

Fields

prd_id: String

story_id: String

iteration: usize

learnings: Vec<String>

context: Value

RalphHandoff

Context handoff between sequential Ralph stories. The finishing story publishes what the next story should know.

Fields

prd_id: String

from_story: String

to_story: String

context: Value

progress_summary: String

RalphProgress

PRD-level progress update (aggregated across all stories).

Fields

prd_id: String

passed: usize

total: usize

iteration: usize

status: String

ToolOutputFull

Full, untruncated tool output from an agent loop step. Published before the result is truncated for the LLM context window, so other agents and the bus log see the complete output.

Fields

agent_id: String

tool_name: String

output: String

success: bool

step: usize

AgentThinking

Internal reasoning / chain-of-thought from an agent step. Published so the training pipeline captures the model’s thinking process alongside its actions.

Fields

agent_id: String

thinking: String

step: usize

VoiceSessionStarted

A voice session (LiveKit room) has been created.

Fields

room_name: String

agent_id: String

voice_id: String

VoiceTranscript

A transcript fragment from a voice session.

Fields

room_name: String

text: String

role: String

“user” or “agent”

is_final: bool

VoiceAgentStateChanged

The voice agent’s state changed (e.g. listening, thinking, speaking).

Fields

room_name: String

state: String

VoiceSessionEnded

A voice session has ended.

Fields

room_name: String

reason: String

Trait Implementations

impl Clone for BusMessage

fn clone(&self) -> BusMessage

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for BusMessage

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for BusMessage

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for BusMessage

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.