07-built-in-tools

07. Built-in Tools

Google Search and Code Executor tools in ADK.

Blog Post: https://arjunprabhulal.com/adk-built-in-tools/

Table of Contents

1. Overview
2. Prerequisites
3. Setup Steps
4. Available Tools
5. Google Search Tool
6. Code Executor Tool
7. Running the Agent
8. Project Structure
9. Next Steps

Overview

ADK provides built-in tools that agents can use out of the box:

• Google Search - Real-time web search with grounding
• Code Executor - Safe Python code execution in a sandbox

Prerequisites

• Python 3.10+
• Gemini API key from AI Studio

Setup Steps

1. Navigate to this module:

cd 07-built-in-tools

2. Create and activate a virtual environment:

python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install dependencies:

pip install -r ../requirements.txt

4. Create .env file in google_search_agent/:

echo "GOOGLE_GENAI_USE_VERTEXAI=0" > google_search_agent/.env
echo "GOOGLE_API_KEY=your-api-key-here" >> google_search_agent/.env

Available Tools

Tool	Import	Description
google_search	from google.adk.tools import google_search	Web search via Google
BuiltInCodeExecutor	from google.adk.code_executors import BuiltInCodeExecutor	Python code execution

Google Search Tool

The Google Search tool enables real-time web searches with grounding:

from google.adk.agents import Agent
from google.adk.tools import google_search

agent = Agent(
    model="gemini-2.5-flash",
    name="search_agent",
    instruction="You are a helpful assistant with web search.",
    tools=[google_search],
)

Features:

• Real-time search results
• Grounded responses with citations
• Automatic query formulation

Code Executor Tool

The Code Executor runs Python code in a secure sandbox:

from google.adk.agents import Agent
from google.adk.code_executors import BuiltInCodeExecutor

agent = Agent(
    model="gemini-2.5-flash",
    name="code_agent",
    instruction="You can write and execute Python code.",
    code_executor=BuiltInCodeExecutor(),
)

Features:

• Sandboxed execution (safe)
• Supports common libraries (math, json, etc.)
• Returns stdout, stderr, and results
• Great for calculations, data processing

Example queries:

• "Calculate the factorial of 10"
• "Generate a list of prime numbers up to 50"
• "Parse this JSON and extract the names"

Running the Agent

Using ADK Web

adk web

Open http://127.0.0.1:8000 and select google_search_agent.

Test Queries:

• "What's the latest news on AI?"
• "Search for Python 3.12 new features"

Using ADK CLI

adk run google_search_agent

Project Structure

07-built-in-tools/
├── README.md
├── google_search_agent/
│   ├── __init__.py
│   ├── agent.py          # Agent with google_search tool
│   └── .env
├── code_agent.py          # Standalone Code Executor example
└── search_agent.py        # Standalone Search example

Next Steps

Continue to 08. Vertex AI RAG