Agents
YAML definitions and configurations for native agent development.
A standardized template for building responsible AI Agents with watsonx Orchestrate using Agent Development Kit (ADK) in IBM Bob.
Start with a short use case and let Bob build the agent workflow step by step.
I would like to develop an AI agent with watsonx Orchestrate here is my use case:
Title: Market analysis agent
Description: Tracks financial news, analyzes historical data, and forecasts trends. I want to also use the Tavily search for web search.
Example prompts:
“What's the recent trend in tech stocks?”
“How has IBM's stock performed over last week?”
“How does the S&P 500 outlook look for 2026?”
“How does the US equity market compare to the rest of world?”
Value:
- Productivity improvement of a financial advisor by 10%
Please propose the agent, tools, toolkit, plugins, knowledge base and connections first and wait for my approval before making changes.
This template provides the knowledge needed to build, test, manage, and evaluate AI agents in watsonx Orchestrate.
YAML definitions and configurations for native agent development.
Custom Python tools and plugins to extend agent capabilities.
RAG-ready knowledge base definitions for domain grounding.
MCP toolkit implementations for connecting reusable capabilities.
Configuration for external systems, credentials, and environments.
Safety and compliance controls to ensure responsible AI behavior.
Evaluation configurations for testing and improving agent quality.
External model configurations to support tailored AI workflows.
The repository is designed to help team start building with standardized code in a productive way, eliminating setup hastles.
After cloning the repo, open it in Bob, scripts are executed automatically to setup wxO ADK environment.
Prompt Bob with details about your use case, and bob will build the agents & tools in an iterative way.
Once agents are built, tested and evaluated, bob can deploy the agents directly to wxO.
Copy and paste one of these example prompts into Bob to get started with building your AI agent:
Improve productivity by up to 50% by using pre-built templates and components that are already tested and optimized for accuracy and reliability.
Get startedWe're here to help! View existing bugs or raise a new issue.