Good default pattern 1. Task 2. Critical rule 3. Exact step order 4. Edge cases or clarification behavior 5. Output format 6. One correct example Avoid - Implied next steps - Unspecified edge cases - Schema-only prompts for tool workflows - Generic instructions without structure

General principles - Be clear and direct - Add context to improve performance - Use examples effectively - Structure prompts with XML tags - Give Claude a role - Long context prompting - Model self-knowledge

Clear and specific instructions - Input - Constraints - Response format - Zero-shot vs few-shot prompts - Optimal number of examples - Consistent formatting - Add context - Break down prompts into components - Experiment with model parameters

Basics - Few-shot learning - Non-chat scenarios - Start with clear instructions - Repeat instructions at the end - Prime the output - Add clear syntax - Break down the task

ICIO pattern - Full name: Instruction-Context-Input Data-Output Indicator Core elements - Instruction - Context - Input data - Output indicator

CRISPE pattern - Full name: Capacity-Role-Insight-Statement-Personality-Experiment Core elements - Role - Insight - Statement - Personality - Experiment

LangGPT pattern - Full name: Language for GPT Core elements - Role - Profile - Rules - Workflow - Tools

RISEN pattern - Full name: Role-Instructions-Steps-End goal-Narrow Core elements - Role - Instructions - Steps - End goal - Scope narrowing

Ask AI pattern Core elements - Tailor the writing style to the model being used - Generate prompts for the specific scenario - Let AI produce a directly usable prompt draft