Tutorial
Prompt Mastery: The quality of OpenClaw's responses depends heavily on how you phrase your requests. Learn professional prompting techniques.
The Anatomy of a Great Prompt
Effective prompts for OpenClaw include:
- Clear context: What situation are you in?
- Specific task: What do you need done?
- Constraints: Any limitations or requirements?
- Output format: How should the response be structured?
Context-First Prompting
"I'm working on a React Native app for iOS. I need help debugging a navigation issue where the back button doesn't work after deep linking."The Chain-of-Thought Technique
"Think step by step. First, analyze the problem. Then, identify possible causes. Finally, suggest solutions with code examples."Example-Driven Prompts
"Create a function similar to this: [example]. Apply the same pattern to: [new scenario]."Setting Output Formats
"Format your response as: 1) Issue 2) Root Cause 3) Solution 4) Prevention"Iterative Refinement
Don't accept the first response if it's not perfect:
"That's close, but can you make the error handling more robust?"System Prompt Customization
openclaw config set personality.system_prompt="You are a senior software engineer. Always consider security, performance, and maintainability in your responses."Common Prompt Patterns
- Code review: "Review for security vulnerabilities, performance issues, and code style"
- Documentation: "Write JSDoc comments for each function with parameter descriptions"
- Debugging: "Explain why this error occurs and how to prevent it in the future"
- Refactoring: "Refactor this code to be more readable while maintaining functionality"
Temperature and Thinking Settings
openclaw config set model.temperature=0.7 # Higher for more creativity
openclaw config set model.thinking="high" # Deeper reasoning