Built-in Tools

Built-in tools provide ready-to-use functionality for common agent tasks. These tools are pre-built, tested, and optimized for integration with ADK TypeScript agents, allowing you to add powerful capabilities without custom development.

How Built-in Tools Work

Built-in tools follow a simple integration pattern:

1. Import the desired tool from the ADK TypeScript library
2. Configure the tool with any required parameters
3. Register the tool in your agent's tools list
4. Use the tool automatically based on agent instructions and user prompts

Available Built-in Tools

Google Search Tool

Enables agents to perform web searches and retrieve real-time information from the internet.

Key Features

• Real-time Results: Access current web information
• Gemini Integration: Optimized for Gemini models
• Automatic Parsing: Results formatted for LLM consumption
• Simple Integration: Works with minimal configuration

Use Cases

• Current Events: Get latest news and information
• Fact Checking: Verify information against web sources
• Research Tasks: Gather information on specific topics
• Market Data: Access current pricing and market information

Configuration

import { GoogleSearch } from '@iqai/adk';

const searchTool = new GoogleSearch();

File Operations Tool

Enables agents to interact with files and documents for reading, writing, and manipulation tasks.

Key Features

• File I/O: Read and write various file formats
• Format Support: Text, JSON, CSV, and other common formats
• Safe Operations: Controlled file system access with security validation
• Multiple Operations: Read, write, append, delete, exists, list, mkdir

Use Cases

• Document Processing: Read and analyze document contents
• Data Import/Export: Load data from files or save results
• Configuration Management: Read settings and configuration files
• Report Generation: Create output files and reports

Configuration

import { FileOperationsTool } from '@iqai/adk';

const fileTool = new FileOperationsTool({
  basePath: '/path/to/working/directory' // Optional, defaults to cwd
});

Security Considerations

• Sandboxed Access: Limited to authorized directories
• Path Validation: Prevents access outside base directory
• Permission Checks: Proper file system permissions
• Error Recovery: Graceful handling of access failures

HTTP Request Tool

Make HTTP requests to external APIs and web services for data integration and communication.

Key Features

• Multiple Methods: Support for GET, POST, PUT, DELETE, PATCH, HEAD, OPTIONS
• Header Management: Custom headers and authentication
• Request Body: JSON and form data support
• Query Parameters: URL parameter handling
• Timeout Control: Configurable request timeouts

Use Cases

• API Integration: Connect to REST APIs and web services
• Data Fetching: Retrieve information from external sources
• Webhook Communication: Send data to external systems
• Service Integration: Connect with third-party platforms

Configuration

import { HttpRequestTool } from '@iqai/adk';

const httpTool = new HttpRequestTool();

User Interaction Tool

Prompt users for input during agent execution for human-in-the-loop workflows.

Key Features

• Interactive Prompts: Request user input with custom messages
• Choice Options: Present multiple choice selections
• Default Values: Pre-populate input fields
• Long Running: Supports asynchronous user response patterns

Use Cases

• Decision Making: Get user approval for actions
• Data Collection: Gather additional information from users
• Confirmation: Verify critical operations before execution
• Preference Setting: Allow users to specify preferences

Configuration

import { UserInteractionTool } from '@iqai/adk';

const userTool = new UserInteractionTool();

Load Memory Tool

Search and load relevant memories from conversation history for context-aware interactions.

Key Features

• Memory Search: Query past conversations and interactions
• Contextual Retrieval: Find relevant information based on current needs
• Automatic Integration: Seamlessly integrate memory results
• Efficient Querying: Optimized search across conversation history

Use Cases

• Context Continuity: Maintain conversation context across sessions
• Information Recall: Reference past discussions and decisions
• Personalization: Access user preferences and history
• Knowledge Building: Build on previous interactions and learning

Configuration

import { LoadMemoryTool } from '@iqai/adk';

const memoryTool = new LoadMemoryTool();

Load Artifacts Tool

Load uploaded files and documents into agent context for processing and analysis.

Key Features

• File Management: Access uploaded documents and media
• Context Integration: Automatically add artifacts to LLM requests
• Dynamic Loading: Load artifacts based on agent needs
• Content Processing: Handle various file types and formats

Use Cases

• Document Analysis: Process uploaded documents and files
• Content Review: Analyze images, PDFs, and text files
• Data Processing: Work with uploaded datasets and spreadsheets
• Media Handling: Process images and multimedia content

Configuration

import { LoadArtifactsTool } from '@iqai/adk';

const artifactsTool = new LoadArtifactsTool();

Additional Built-in Tools

Get User Choice Tool

Present options to users and collect their selection for decision-making workflows.

import { GetUserChoiceTool } from '@iqai/adk';

const choiceTool = new GetUserChoiceTool();

Transfer to Agent Tool

Transfer control to another agent for specialized processing or workflow delegation.

import { TransferToAgentTool } from '@iqai/adk';

const transferTool = new TransferToAgentTool();

Exit Loop Tool

Exit execution loops in workflow agents when completion conditions are met.

import { ExitLoopTool } from '@iqai/adk';

const exitTool = new ExitLoopTool();

Tool Integration Patterns

Single Tool Usage

Use individual tools for specific capabilities:

• Search Only: Agents focused on information retrieval
• Files Only: Document processing specialists
• Memory Only: Context-aware conversation agents

Multi-Tool Combinations

Combine tools for complex workflows:

• Search + Files: Research with document storage
• Memory + User Interaction: Personalized interactive agents
• Files + HTTP: Data processing with external API integration
• All Tools: Comprehensive capability agents

Conditional Tool Selection

Agents choose tools based on task requirements:

• Task Analysis: LLM determines required capabilities
• Dynamic Selection: Different tools for different queries
• Fallback Strategies: Alternative approaches when tools fail

Performance and Limitations

Rate Limiting

• Google Search: Subject to API quotas and service availability
• HTTP Requests: Limited by target service rate limits
• File Operations: File system and storage limits
• User Interaction: Depends on user response time

Error Handling

Built-in tools include robust error handling:

• Network Failures: Graceful handling of connectivity issues
• Authentication Errors: Clear messaging for credential problems
• Rate Limit Responses: Appropriate backoff and retry logic
• Result Validation: Verification of tool outputs

Performance Optimization

• Async Operations: Non-blocking execution where possible
• Resource Management: Efficient use of system resources
• Error Recovery: Automatic retry logic for transient failures
• Timeout Handling: Configurable timeouts for external operations

Best Practices

Tool Selection

• Match Capabilities: Choose tools that match your use case requirements
• Consider Environment: Ensure tools work in your deployment environment
• Plan Fallbacks: Design alternative approaches for tool failures
• Monitor Usage: Track tool usage patterns and performance

Agent Instructions

• Clear Guidance: Provide specific instructions for when to use each tool
• Error Handling: Define how agents should handle tool failures
• Result Processing: Guide agents on interpreting tool outputs
• User Communication: Instruct agents on communicating tool usage to users

Security and Privacy

• Data Sensitivity: Consider data sensitivity when using external tools
• Access Controls: Implement appropriate access restrictions
• Audit Trails: Log tool usage for security and compliance
• Error Information: Avoid exposing sensitive data in error messages

Related Topics