Trait SkillExecutor 

pub trait SkillExecutor: Send + Sync {
    // Required methods
    fn execute_tool<'life0, 'life1, 'life2, 'async_trait>(
        &'life0 self,
        skill_id: &'life1 str,
        tool_name: &'life2 str,
        args: Value,
    ) -> Pin<Box<dyn Future<Output = Result<ToolOutput, SkillError>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait,
             'life2: 'async_trait;
    fn tools(&self, skill_id: &str) -> Result<Vec<ToolDefinition>, SkillError>;
}

Skill tool executor.

Executes a single named tool within a skill identified by skill_id. This trait does not run an LLM loop — it only handles one tool call at a time. Callers are responsible for any higher-level orchestration.

Example

use async_trait::async_trait;
use xz_skill_core::traits::{SkillExecutor, ToolOutput};
use xz_skill_core::SkillError;
use xz_skill_core::types::skill::ToolDefinition;

struct EchoExecutor;

#[async_trait]
impl SkillExecutor for EchoExecutor {
    async fn execute_tool(
        &self,
        _skill_id: &str,
        tool_name: &str,
        args: serde_json::Value,
    ) -> Result<ToolOutput, SkillError> {
        Ok(ToolOutput {
            output: format!("{}: {}", tool_name, args),
            is_error: false,
        })
    }

    fn tools(&self, _skill_id: &str) -> Result<Vec<ToolDefinition>, SkillError> {
        Ok(vec![ToolDefinition {
            name: "echo".into(),
            description: "Echoes input".into(),
            input_schema: serde_json::json!({}),
            tool_type: crate::types::skill::ToolType::Builtin {
                handler: "echo".into(),
            },
        }])
    }
}

Required Methods

fn execute_tool<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, skill_id: &'life1 str, tool_name: &'life2 str, args: Value, ) -> Pin<Box<dyn Future<Output = Result<ToolOutput, SkillError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Execute a single named tool with the given arguments.

• skill_id — Identifies which skill provides the tool.
• tool_name — The name of the tool to invoke.
• args — JSON arguments passed to the tool.

fn tools(&self, skill_id: &str) -> Result<Vec<ToolDefinition>, SkillError>

List all tools available for a given skill.

Returns the tool definitions so callers can discover capabilities without executing anything.

Dyn Compatibility

This trait is dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementors