Struct AgentBehaviorConfig 

pub struct AgentBehaviorConfig {
    pub visibility: Visibility,
    pub instructions: Option<String>,
    pub persona: Option<String>,
    pub greeting: Option<String>,
    pub conversation_workflow: Option<ConversationWorkflow>,
    pub enabled_tools: Vec<EnabledTool>,
}

The resolved per-agent behavior knobs. Every field is optional so a partial or malformed agents row degrades cleanly to the org default.

Fields

visibility: Visibility

Where the agent is reachable — gates tool auth. Defaults to Public.

instructions: Option<String>

instructions.prompt — the agent’s system prompt / persona. When present it overrides the org default persona for this agent’s conversations.

persona: Option<String>

personality.persona — an optional custom-persona addendum appended to the system prompt.

greeting: Option<String>

greeting — an optional opening line, injected into the prompt only on the first turn of a conversation (see [greeting_section]).

conversation_workflow: Option<ConversationWorkflow>

conversation_workflow — the optional stepped guided flow. None (or a malformed / empty-steps value) means the agent runs freeform.

enabled_tools: Vec<EnabledTool>

tool_config.enabledTools — a tool allow-list. When non-empty, this agent’s turns are restricted to the enabled == true entries’ tool_id (empty ⇒ the full server tool set). Unknown tool ids are ignored.

Implementations

impl AgentBehaviorConfig

pub fn is_empty(&self) -> bool

true when the row carried nothing usable — the runner should stay on the org default persona and take no workflow path.

pub fn system_prompt(&self) -> Option<String>

Build the per-agent system prompt from instructions (+ optional persona), or None when there are no instructions to anchor it.

None is the signal to fall back to the org default persona — a bare persona with no instructions is not enough to define an agent. The greeting is handled separately (greeting_section) because it is injected first-turn-only.

pub fn greeting_section(&self) -> Option<String>

The <GreetingAwareness> prompt section, or None when no greeting is set. The caller injects it only on the FIRST turn (empty prior history), so the agent opens with it once. Mirrors the sibling lanes’ first-turn greeting.

pub fn enabled_tool_ids(&self) -> Option<Vec<String>>

The enabled tool-id allow-list, or None when unrestricted (no tool_config / empty enabledTools ⇒ the full server tool set). Some(ids) restricts the turn to those snake_case ids (enabled == true entries only); unknown ids simply match nothing.

pub fn auth_level_for(&self, tool_id: &str) -> AuthLevel

The configured AuthLevel for a tool id (from its enabledTools entry), or AuthLevel::None when unconfigured.

pub fn tool_configs(&self) -> HashMap<String, Value>

The per-tool config object delivered to a tool at execution (the enabledTools entry’s config), for every entry that carries one. Empty when no tool has config. Mirrors registry.ts’s toolSpecificConfig.

pub fn from_row_values( instructions: Option<Value>, personality: Option<Value>, greeting: Option<String>, conversation_workflow: Option<Value>, tool_config: Option<Value>, visibility: Option<String>, ) -> Self

Parse from the raw agents-row jsonb / text columns, tolerating any malformed shape (a bad value drops just that field — never an error).

• instructions — jsonb { "prompt": string },
• personality — jsonb { "persona"?: string, ... },
• greeting — text,
• conversation_workflow — jsonb { goal, steps: [...] },
• tool_config — jsonb { enabledTools: [{ toolId, enabled, authLevel, config }] },
• visibility — text public | internal (defaults public).

Trait Implementations

impl Clone for AgentBehaviorConfig

fn clone(&self) -> AgentBehaviorConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for AgentBehaviorConfig

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

Formats the value using the given formatter.

impl Default for AgentBehaviorConfig

fn default() -> AgentBehaviorConfig

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for AgentBehaviorConfig

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Eq for AgentBehaviorConfig

impl PartialEq for AgentBehaviorConfig

fn eq(&self, other: &AgentBehaviorConfig) -> 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 AgentBehaviorConfig

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

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for AgentBehaviorConfig