Skip to main content

​
Why Prompt Engineering Matters

The difference between a mediocre AI integration and a great one often comes down to how you communicate with the model. Well-crafted prompts lead to:
  • More accurate responses — The model understands exactly what you need
  • Consistent behavior — Reliable outputs across different inputs
  • Better tool usage — Agents call the right tools with correct arguments
  • Reduced costs — Fewer retries and failed interactions

​
What You’ll Learn

System Prompts

Design system prompts that define agent behavior, personality, and constraints.

Tool Descriptions

Write tool descriptions that help agents make smart decisions about when and how to use tools.

Message Formatting

Structure your messages for clarity and optimal model understanding.

Common Patterns

Proven prompt patterns for common agent workflows and use cases.

​
Quick Reference

​
The Prompt Quality Checklist

Specific — Avoids vague terms like “good” or “appropriate”
Structured — Uses clear formatting (lists, sections, examples)
Contextual — Provides relevant background information
Constrained — Defines boundaries and limitations
Testable — Can verify if the output meets requirements

​
Before You Start

The Agent API uses three key areas where prompt engineering applies:
  1. System Prompt — Defines the agent’s role, behavior, and constraints
  2. Tool Descriptions — Tells the agent what each tool does and when to use it
  3. User Messages — The actual requests and context you send
Each area has different best practices, covered in the guides above.
Start with: System Prompts →