Struct CodeTool 

pub struct CodeTool { /* private fields */ }

A tool that executes code through language-specific pipelines.

CodeTool wraps a RustExecutor and implements adk_core::Tool, making code execution with compiler diagnostics available to LLM agents. Phase 1 supports Rust only; other languages return a descriptive error.

Error Handling

Executor errors are never propagated as ToolError. Instead, they are converted to structured JSON with a "status" field. Compile errors include a "diagnostics" array with structured compiler output.

Example

use adk_code::{CodeTool, RustExecutor, RustExecutorConfig};
use adk_sandbox::ProcessBackend;
use std::sync::Arc;

let backend = Arc::new(ProcessBackend::default());
let executor = RustExecutor::new(backend, RustExecutorConfig::default());
let tool = CodeTool::new(executor);
assert_eq!(tool.name(), "code_exec");

Implementations

impl CodeTool

pub fn new(executor: RustExecutor) -> Self

Creates a new CodeTool wrapping the given Rust executor.

Trait Implementations

impl Tool for CodeTool

fn name(&self) -> &str

fn description(&self) -> &str

fn required_scopes(&self) -> &[&str]

Returns the scopes required to execute this tool.

fn parameters_schema(&self) -> Option<Value>

fn execute<'life0, 'async_trait>( &'life0 self, _ctx: Arc<dyn ToolContext>, args: Value, ) -> Pin<Box<dyn Future<Output = Result<Value>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

fn enhanced_description(&self) -> String

Returns an enhanced description that may include additional notes. For long-running tools, this includes a warning not to call the tool again if it has already returned a pending status. Default implementation returns the base description.

fn is_long_running(&self) -> bool

Indicates whether the tool is a long-running operation. Long-running tools typically return a task ID immediately and complete the operation asynchronously.

fn response_schema(&self) -> Option<Value>