Struct HatConfig 

pub struct HatConfig {

    pub name: String,
    pub description: Option<String>,
    pub triggers: Vec<String>,
    pub publishes: Vec<String>,
    pub instructions: String,
    pub extra_instructions: Vec<String>,
    pub backend: Option<HatBackend>,
    pub backend_args: Option<Vec<String>>,
    pub default_publishes: Option<String>,
    pub max_activations: Option<u32>,
    pub disallowed_tools: Vec<String>,
    pub timeout: Option<u32>,
    pub concurrency: u32,
    pub aggregate: Option<AggregateConfig>,
}

Configuration for a single hat.

Fields

name: String

Human-readable name for the hat.

description: Option<String>

Short description of the hat’s purpose (required). Used in the HATS table to help Ralph understand when to delegate to this hat.

triggers: Vec<String>

Events that trigger this hat to be worn. Per spec: “Hats define triggers — which events cause Ralph to wear this hat.”

publishes: Vec<String>

Topics this hat publishes.

instructions: String

Instructions prepended to prompts.

extra_instructions: Vec<String>

Additional instruction fragments appended to instructions.

Use with YAML anchors to share common instruction blocks across hats:

_confidence_protocol: &confidence_protocol |
  ### Confidence-Based Decision Protocol
  ...

hats:
  architect:
    instructions: |
      ## ARCHITECT MODE
      ...
    extra_instructions:
      - *confidence_protocol

backend: Option<HatBackend>

Backend to use for this hat (inherits from cli.backend if not specified).

backend_args: Option<Vec<String>>

Custom args to append to the backend CLI when this hat is active.

Accepts both backend_args: and shorthand args:.

default_publishes: Option<String>

Default event to publish if hat forgets to write an event.

max_activations: Option<u32>

Maximum number of times this hat may be activated in a single loop run.

When the limit is exceeded, the orchestrator publishes <hat_id>.exhausted instead of activating the hat again.

disallowed_tools: Vec<String>

Tools the hat is not allowed to use.

Injected as a TOOL RESTRICTIONS section in the prompt (soft enforcement). After each iteration, a file-modification audit checks compliance when Edit or Write are disallowed (hard enforcement via scope_violation event).

timeout: Option<u32>

Execution timeout in seconds for this hat.

For wave workers, this controls how long each parallel worker can run. Defaults to the adapter-level timeout (typically 300s) if not set.

concurrency: u32

Maximum concurrent wave instances for this hat.

When > 1, the loop runner spawns multiple backend instances in parallel for wave events targeting this hat. Default is 1 (sequential execution).

aggregate: Option<AggregateConfig>

Aggregation configuration for this hat.

When set, this hat acts as an aggregator — it buffers wave results and activates only when all correlated results have arrived (or timeout). Cannot be set on a hat with concurrency > 1.

Implementations

impl HatConfig

pub fn trigger_topics(&self) -> Vec<Topic>

Converts trigger strings to Topic objects.

pub fn publish_topics(&self) -> Vec<Topic>

Converts publish strings to Topic objects.

Trait Implementations

impl Clone for HatConfig

fn clone(&self) -> HatConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for HatConfig

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for HatConfig

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for HatConfig

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

where __S: Serializer,

Serialize this value into the given Serde serializer.