Struct AgentConfig 

pub struct AgentConfig {

    pub name: String,
    pub description: String,
    pub system_prompt: String,
    pub model: Option<String>,
    pub max_iterations: usize,
    pub tool_choice: ToolChoice,
    pub thinking: bool,
    pub members: Vec<String>,
    pub skills: Vec<String>,
    pub mcps: Vec<String>,
    pub tools: Vec<String>,
    pub compact_threshold: Option<usize>,
    pub compact_tool_max_len: usize,
}

Serializable agent configuration.

Contains all parameters for an agent: identity, system prompt, model, iteration limits, and delegation scope. Used both as the TOML deserialization target and the runtime agent definition.

Fields

name: String

Agent identifier. Derived from TOML key, not stored in TOML.

description: String

Human-readable description.

system_prompt: String

System prompt sent before each LLM request. Loaded from .md file.

model: Option<String>

Model to use from the registry. None = registry’s active/default.

max_iterations: usize

Maximum iterations before stopping.

tool_choice: ToolChoice

Controls which tool the model calls.

thinking: bool

Whether to enable thinking/reasoning mode.

members: Vec<String>

Agents this agent can delegate to via spawn_task. Empty = no delegation.

skills: Vec<String>

Skill names this agent can access. Empty = all skills (crabtalk default).

mcps: Vec<String>

MCP server names this agent can access. Empty = all MCPs (crabtalk default).

tools: Vec<String>

Computed tool whitelist. Empty = all tools. Not serialized.

compact_threshold: Option<usize>

Token count threshold for automatic context compaction. When history exceeds this, the agent compacts automatically. None = disabled. Defaults to 100_000.

compact_tool_max_len: usize

Max byte length to keep from tool-role messages when compacting. Longer results are truncated before sending to the compaction LLM.

Implementations

impl AgentConfig

pub fn new(name: impl Into<String>) -> AgentConfig

Create a new config with the given name and defaults for everything else.

pub fn system_prompt(self, prompt: impl Into<String>) -> AgentConfig

Set the system prompt.

pub fn description(self, desc: impl Into<String>) -> AgentConfig

Set the description.

pub fn model(self, name: impl Into<String>) -> AgentConfig

Set the model to use from the registry.

pub fn thinking(self, enabled: bool) -> AgentConfig

Enable or disable thinking/reasoning mode.

Trait Implementations

impl Clone for AgentConfig

fn clone(&self) -> AgentConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for AgentConfig

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

Formats the value using the given formatter.

impl Default for AgentConfig

fn default() -> AgentConfig

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for AgentConfig

fn deserialize<__D>( __deserializer: __D, ) -> Result<AgentConfig, <__D as Deserializer<'de>>::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for AgentConfig

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

where __S: Serializer,

Serialize this value into the given Serde serializer.