Recap and Feedback

In this part, we extended basic LLM applications into tool-using systems.

The most important lessons were:

• tools give the model access to deterministic capabilities,
• tool schemas reduce ambiguity,
• the application, not the model, executes the tools,
• and useful systems stay inside explicit boundaries.

The tutorial chapter showed how these ideas work in a command-line assistant with local issue data. The framework-variant chapter then showed that the same ideas can be expressed through LangChainJS without removing the application’s responsibility for boundaries and validation.

In the next part, we continue by looking at another important capability: giving applications access to external knowledge and then evaluating whether the results are actually good enough.

Next, please take a moment to reflect on your work in this part and provide feedback. Your input helps us improve the course materials.