Skip to content

Python Path

This is the shortest useful route through Harness Guides if you are trying to build an agent harness in Python.

What this path optimizes for

  • Python-first interfaces and examples
  • OpenAI Responses API as the concrete model and tool layer
  • architecture that stays portable even if the API provider changes later
  1. What Is An Agent Harness?
  2. OpenAI Responses API
  3. Chapter 1: Tool System
  4. Chapter 3: Query Engine
  5. Chapter 5: Permissions
  6. Chapter 6: Session and State
  7. Chapter 8: MCP Integration

What you should have by the end

  • a clear tool interface
  • a model loop that can call tools and continue
  • a permission layer that is separate from validation
  • a transcript and state model
  • a path to streaming, MCP, and long-running work

Working rule

Do not copy surface details from any one product. Copy the harness patterns that remain useful when the UI, model, or provider changes.