Tool Trait

Tool 是 Awaken 暴露能力给 LLM 的主扩展点。tool 接收 JSON 参数和只读上下文，返回 ToolOutput。

Trait 定义

#[async_trait]
pub trait Tool: Send + Sync {
    fn descriptor(&self) -> ToolDescriptor;

    fn validate_args(&self, _args: &Value) -> Result<(), ToolError> {
        Ok(())
    }

    async fn execute(
        &self,
        args: Value,
        ctx: &ToolCallContext,
    ) -> Result<ToolOutput, ToolError>;
}

ToolDescriptor

描述 tool 的 ID、名称、说明和参数 schema。它会被注册到运行时，并在推理请求里暴露给 LLM。

pub struct ToolDescriptor {
    pub id: String,
    pub name: String,
    pub description: String,
    pub parameters: Value,
    pub category: Option<String>,
}

Builder 方法：

ToolDescriptor::new(id, name, description) -> Self
    .with_parameters(schema: Value) -> Self
    .with_category(category: impl Into<String>) -> Self

ToolResult

Tool::execute 返回的结构化结果。

pub struct ToolResult {
    pub tool_name: String,
    pub status: ToolStatus,
    pub data: Value,
    pub message: Option<String>,
    pub suspension: Option<Box<SuspendTicket>>,
}

ToolStatus

pub enum ToolStatus {
    Success,
    Pending,
    Error,
}

构造函数

方法	状态	用途
`ToolResult::success(name, data)`	`Success`	正常完成
`ToolResult::success_with_message(name, data, msg)`	`Success`	带补充说明的成功
`ToolResult::error(name, message)`	`Error`	可恢复失败
`ToolResult::error_with_code(name, code, message)`	`Error`	带 code 的结构化失败
`ToolResult::suspended(name, message)`	`Pending`	HITL 挂起
`ToolResult::suspended_with(name, message, ticket)`	`Pending`	带 `SuspendTicket` 的挂起

判定方法

is_success()
is_pending()
is_error()
to_json()

ToolError

ToolError 和 ToolResult::error(...) 的区别是：前者直接终止该次 tool call，后者会把失败信息回传给 LLM。

pub enum ToolError {
    InvalidArguments(String),
    ExecutionFailed(String),
    Denied(String),
    NotFound(String),
    Internal(String),
}

ToolCallContext

tool 执行期拿到的只读上下文：

pub struct ToolCallContext {
    pub call_id: String,
    pub tool_name: String,
    pub run_identity: RunIdentity,
    pub agent_spec: Arc<AgentSpec>,
    pub snapshot: Snapshot,
    pub activity_sink: Option<Arc<dyn EventSink>>,
}

方法

fn state<K: StateKey>(&self) -> Option<&K::Value>
async fn report_activity(&self, activity_type: &str, content: &str)
async fn report_activity_delta(&self, activity_type: &str, patch: Value)
async fn report_progress(
    &self,
    status: ProgressStatus,
    message: Option<&str>,
    progress: Option<f64>,
)

示例

最小 tool

use async_trait::async_trait;
use awaken::contract::tool::{Tool, ToolCallContext, ToolDescriptor, ToolError, ToolResult, ToolOutput};
use serde_json::{Value, json};

struct Greet;

#[async_trait]
impl Tool for Greet {
    fn descriptor(&self) -> ToolDescriptor {
        ToolDescriptor::new("greet", "greet", "Greet a user by name")
            .with_parameters(json!({
                "type": "object",
                "properties": {
                    "name": { "type": "string" }
                },
                "required": ["name"]
            }))
    }

    async fn execute(
        &self,
        args: Value,
        _ctx: &ToolCallContext,
    ) -> Result<ToolOutput, ToolError> {
        let name = args["name"]
            .as_str()
            .ok_or_else(|| ToolError::InvalidArguments("name required".into()))?;
        Ok(ToolResult::success("greet", json!({ "greeting": format!("Hello, {name}!") })).into())
    }
}

从上下文读取状态

use async_trait::async_trait;
use awaken::contract::tool::{Tool, ToolCallContext, ToolDescriptor, ToolError, ToolResult, ToolOutput};
use awaken::state::StateKey;
use serde_json::{Value, json};

struct GetPreferences;

#[async_trait]
impl Tool for GetPreferences {
    fn descriptor(&self) -> ToolDescriptor {
        ToolDescriptor::new("get_prefs", "get_preferences", "Get user preferences")
    }

    async fn execute(
        &self,
        _args: Value,
        ctx: &ToolCallContext,
    ) -> Result<ToolOutput, ToolError> {
        // let prefs = ctx.state::<UserPreferences>().cloned().unwrap_or_default();
        Ok(ToolResult::success("get_prefs", json!({})).into())
    }
}

Tool 执行钩子

每次 tool call 在真正执行前后都会经过插件钩子。

完整生命周期

LLM 选择 tool
  -> validate_args()
  -> BeforeToolExecute
     插件可调度 ToolInterceptAction:
       Block
       Suspend
       SetResult
  -> execute()
  -> AfterToolExecute

BeforeToolExecute

参数校验后运行。插件可以在这里阻断、挂起，或直接提供预构造结果。

拦截优先级：

Block > Suspend > SetResult

AfterToolExecute

tool 完成后运行。插件可以观察 ToolResult、更新状态、追加事件或调度后续 action。

ToolCallStatus 转移

New -> Running -> Succeeded
                  Failed
                  Suspended -> Resuming -> Running -> ...
                  Cancelled

终态（Succeeded、Failed、Cancelled）不能再向前转移。

Keyboard shortcuts

Awaken