Module 9 Wrap-up: The Toolsmith

You have learned that LLMs don't have to be "Shut-in researchers." By providing Custom Tools and Community Integrations, you have given your AI the ability to interact with the world. This is the Final Ingredient required to build an autonomous Agent.

---

Hands-on Exercise: The Calculator-Researcher

1. The Goal

Create a list of 2 tools:

1. A custom @tool that calculates the area of a circle given a radius.
2. A standard Wikipedia search tool.

2. The Implementation Plan

1. Define the calculate_area function with a clear docstring.
2. Import WikipediaQueryRun.
3. Store them in a list called tools = [...].
4. Test: Call calculate_area.invoke(5) and see the result.

---

Module 9 Summary

• Tools: Abstractions that allow LLMs to take action.
• Descriptions: The "Manual" the AI reads to understand the tool.
• Decorator: Using @tool to instantly convert Python code.
• Community: Using pre-built tools like Tavily or Wikipedia.
• Safety: Understanding the risks of tools like PythonREPL.

Coming Up Next...

In Module 10, we bring everything together into Agents. We will learn about the Reasoning Loop, and how a model uses these tools to solve complex, multi-step problems autonomously.

---

Module 9 Checklist

• I have converted a Python function into a tool using @tool.
• My tool docstrings are descriptive and helpful for an AI.
• I understand the JSON schema that is generated by the decorator.
• I have used the Tavily or Wikipedia tool in a script.
• I know why small, atomic tools are better than large, complex ones.