Concept
The model picks a tool by reading its description, not your intent.
Two vaguely-worded tools are indistinguishable to it, however obvious the difference seems to you.
Concept
Tool Definitions Models Do Not Misuse
Learn to write tool definitions an LLM calls correctly, with the right arguments, at the right moment. Free, one sitting.
Start Quick LessonFree to explore · ~3 minutes
Why this matters
When an agent calls the wrong tool or passes nonsense arguments, the fix is almost never a better system prompt — it is a clearer tool definition. Names, descriptions and required parameters are the actual interface between reasoning and action, and most are written as an afterthought.
What you'll cover
The Description Is the Interface
- A tool definition is a prompt
- How the model chooses
- The cost of ambiguity
- Why the system prompt is not the fix
- What good definitions look like
Naming and Describing
- Naming for distinction
- Writing the description
- Stating the boundaries
- Examples in descriptions
- Length and clarity
Parameter Design
- Required versus optional
- Types and enums
- Describing each parameter
- Defaults and their risks
- Validating arguments
How Many Tools
- Accuracy versus tool count
- Splitting an overloaded tool
- Merging near-duplicates
- Dynamic tool sets
- Tool definitions and token cost
Debugging Wrong Calls
- Reading the call that went wrong
- Wrong tool chosen
- Right tool, wrong arguments
- Tool not called at all
- Testing tool selection
What you'll understand
- Write tool names and descriptions the model reads correctly
- Design parameters that prevent invalid calls
- Say when to split or merge tools
- Diagnose a wrong tool call from the definition
A look inside
Three moments from this Quick Lesson
The real thing — not a mockup of it.
How it fits
Name + Description + Schema → Model choice → Call
Every wrong call traces back to one of the first three.
Apply
An agent keeps calling the wrong tool. First fix?
Add more instructions to the system prompt
Clarify the tool names and descriptions
Switch to a bigger model
How it works
01
Understand the concept
A plain-language walkthrough of the idea itself, no prior context assumed.
02
See it in practice
A simple diagram or example showing how it actually fits together.
03
Apply what you learned
One quick check that you can recognise it, not just recall it.
Useful for
DevelopersAI engineersProduct managersFounders
Ready to understand Tool Definitions Models Do Not Misuse?
Want to go deeper? Explore Agent Architect →
Frequently asked
How many tools is too many?+
Accuracy drops as similar tools multiply. Prefer fewer, clearly distinct tools over many overlapping ones.
What will I learn?+
How to write tool names, descriptions and schemas that lead to correct calls, and how to debug incorrect ones.
Is this Quick Lesson free?+
Yes. Quick Lessons are free, short, and do not require a paid plan — sign in only to save your progress.