Cursor rule

1---
2description: "Prompt engineering templates and best practices for AI model interactions"
3alwaysApply: false
4---
5 
6# Prompt Engineering Templates
7 
8Guidelines and templates for effective AI prompt engineering. Use `@prompt-engineering` to include this rule.
9 
10## Prompt Engineering Best Practices
11 
12### 1. Prompt Structure Templates
13```typescript
14// System prompt template
15const SYSTEM_PROMPT = `
16You are an expert assistant specialized in [DOMAIN].
17Your responses should be:
18- Accurate and factual
19- Concise but comprehensive
20- Professional in tone
21- Focused on [SPECIFIC_GOAL]
22
23Context: [CONTEXT_INFORMATION]
24`;
25 
26// User prompt template
27const createUserPrompt = (input: string, context?: string) => `
28Task: ${input}
29${context ? `Additional context: ${context}` : ''}
30
31Please provide a response that follows the system guidelines.
32`;
33```
34 
35### 2. Prompt Versioning
36- Store prompts in separate configuration files
37- Version prompts for A/B testing
38- Track prompt performance and iterate
39- Document prompt changes and rationale
40 
41### 3. Dynamic Prompt Generation
42```typescript
43interface PromptConfig {
44  system: string;
45  temperature: number;
46  maxTokens: number;
47  stopSequences?: string[];
48}
49 
50const generatePrompt = (
51  task: string, 
52  userInput: string, 
53  config: PromptConfig
54) => {
55  // Build context-aware prompts
56  // Include relevant examples
57  // Optimize for token efficiency
58};
59```
60 
61### 4. Context Management
62- Implement conversation memory
63- Manage context window limits
64- Prioritize relevant information
65- Handle context overflow gracefully
66 
67### 5. Prompt Optimization Techniques
68- Use few-shot learning examples
69- Implement chain-of-thought prompting
70- Apply role-based prompting
71- Include output format specifications
72 
73### 6. Testing and Validation
74- Create test suites for prompt effectiveness
75- Measure response quality metrics
76- Compare different prompt versions
77- Monitor prompt performance in production