Learn Prompting
Master the techniques that transform how you communicate with AI models
Prompting Techniques
Zero-Shot Prompting
Direct instructions without examples. Works well for simple, well-defined tasks.
Few-Shot Prompting
Provide examples to guide format and style. Perfect for consistent outputs.
Chain-of-Thought
Ask for step-by-step reasoning. Essential for complex math and logic problems.
Role-Based Prompting
Assign a persona or expertise to shape tone and depth of response.
XML Tags & Delimiters
Use structured formatting to separate instructions from content.
System Prompts
Set persistent behavior and constraints for conversations.
Prompt Chaining
Break complex tasks into sequential steps for better results.
Tree of Thoughts
Explore multiple reasoning paths and select the best solution.
Meta-Prompting
Use AI to generate and improve prompts automatically.
Reflection Prompting
Have AI critique and improve its own outputs.
Use Cases
Model-Specific Guides
Real-World Prompt Library
Study system prompts used in real open-source coding tools. Each entry links directly to source files and highlights model-specific flavor differences.
Explore Real-World PromptsOpenCode
Default system prompt used in anomalyco/opencode sessions.
Fundamentals
1. Give context
Tell AI who you are and what you're working on. Explain the background so it understands your needs.
2. Be specific
Say exactly what you want. Vague prompts get vague results—be clear about format, length, and tone.
3. Try again
First attempts are starting points. Refine based on what AI returns—it's a conversation, not a command.