Struct AgentDefinition 

pub struct AgentDefinition {

    pub slug: String,
    pub display_name: String,
    pub description: String,
    pub backend: AgentBackendSpec,
    pub system_prompt: String,
    pub allowed_mcp_tools: Vec<String>,
    pub max_sensitivity: u8,
    pub outcome_sharing: OutcomeSharing,
    pub semantic_profile: AgentSemanticProfile,
    pub context_policy: AgentContextPolicy,
    pub data_policy: AgentDataPolicy,
    pub execution_policy: AgentExecutionPolicy,
    pub roster_version: u16,
    pub enabled: bool,
    pub created_at_unix: u64,
    pub updated_at_unix: u64,
}

A single agent in a principal’s roster.

Keyed by slug, a stable identifier that survives renames of display_name. All fields are public so callers (e.g. the command/API layer) can build and edit definitions directly; use AgentDefinition::new for a conservatively-defaulted, timestamped entry.

Fields

slug: String

Stable identifier for this agent (unique within the roster).

display_name: String

Human-friendly label shown in the UI.

description: String

Free-form description of what the agent is for.

backend: AgentBackendSpec

The inference backend this agent uses.

system_prompt: String

The agent’s persona / system prompt.

allowed_mcp_tools: Vec<String>

MCP tool allowlist. Empty = no tools allowed; a single "*" entry = all tools allowed. See AgentDefinition::has_tool.

max_sensitivity: u8

Ceiling of material sensitivity this agent may handle (0 = public).

outcome_sharing: OutcomeSharing

How this agent’s processed outcomes may be shared with peers.

semantic_profile: AgentSemanticProfile

Pointed ontology-linked study graph: classification, specialization, geographic/language scope, methods, datasets, tools and constraints.

context_policy: AgentContextPolicy

Fine-grained, directional context contract. Missing in pre-existing roster files means the safe default above, preserving compatibility.

data_policy: AgentDataPolicy

Narrower per-agent data-source boundary, intersected with session permissions at inference time.

execution_policy: AgentExecutionPolicy

Model residency and remote-consent preferences, interpreted by the job scheduler rather than by the inference hot path.

roster_version: u16

Persisted schema marker for future non-breaking roster migrations.

enabled: bool

Whether the agent is currently available for invocation.

created_at_unix: u64

Unix seconds at which this agent was first created.

updated_at_unix: u64

Unix seconds at which this agent was last written.

Implementations

impl AgentDefinition

pub fn new( slug: impl Into<String>, display_name: impl Into<String>, description: impl Into<String>, backend: AgentBackendSpec, system_prompt: impl Into<String>, ) -> Self

Construct a new agent with a conservative default posture: empty tool allowlist, public-only sensitivity (SENSITIVITY_PUBLIC), default (owner-only) outcome sharing, and enabled = true. Both timestamps are stamped with the current wall-clock. Raise the sensitivity ceiling or widen the allowlist deliberately after construction.

pub fn has_tool(&self, tool: &str) -> bool

Whether this agent is permitted to use the MCP tool named tool.

An empty allowlist permits nothing; a "*" entry permits everything; otherwise the tool must be listed by exact name.

Trait Implementations

impl Clone for AgentDefinition

fn clone(&self) -> AgentDefinition

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for AgentDefinition

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for AgentDefinition

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for AgentDefinition

fn eq(&self, other: &AgentDefinition) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for AgentDefinition

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

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl Eq for AgentDefinition

impl StructuralPartialEq for AgentDefinition