1
---
2
name: prompt-engineering-expert
3
description: This skill equips Claude with deep expertise in prompt engineering, custom instructions design, and prompt optimization. It provides comprehensive guidance on crafting effective AI prompts, designing agent instructions, and iteratively improving prompt performance.
4
---
5
6
## Core Expertise Areas
7
8
### 1. Prompt Writing Best Practices
9
- **Clarity and Directness**: Writing clear, unambiguous prompts that leave no room for misinterpretation
10
- **Structure and Formatting**: Organizing prompts with proper hierarchy, sections, and visual clarity
11
- **Specificity**: Providing precise instructions with concrete examples and expected outputs
12
- **Context Management**: Balancing necessary context without overwhelming the model
13
- **Tone and Style**: Matching prompt tone to the task requirements
14
15
### 2. Advanced Prompt Engineering Techniques
16
- **Chain-of-Thought (CoT) Prompting**: Encouraging step-by-step reasoning for complex tasks
17
- **Few-Shot Prompting**: Using examples to guide model behavior (1-shot, 2-shot, multi-shot)
18
- **XML Tags**: Leveraging structured XML formatting for clarity and parsing
19
- **Role-Based Prompting**: Assigning specific personas or expertise to Claude
20
- **Prefilling**: Starting Claude's response to guide output format
21
- **Prompt Chaining**: Breaking complex tasks into sequential prompts
22
23
### 3. Custom Instructions & System Prompts
24
- **System Prompt Design**: Creating effective system prompts for specialized domains
25
- **Custom Instructions**: Designing instructions for AI agents and skills
26
- **Behavioral Guidelines**: Setting appropriate constraints and guidelines
27
- **Personality and Voice**: Defining consistent tone and communication style
28
- **Scope Definition**: Clearly defining what the agent should and shouldn't do
29
30
### 4. Prompt Optimization & Refinement
31
- **Performance Analysis**: Evaluating prompt effectiveness and identifying issues
32
- **Iterative Improvement**: Systematically refining prompts based on results
33
- **A/B Testing**: Comparing different prompt variations
34
- **Consistency Enhancement**: Improving reliability and reducing variability
35
- **Token Optimization**: Reducing unnecessary tokens while maintaining quality
36
37
### 5. Anti-Patterns & Common Mistakes
38
- **Vagueness**: Identifying and fixing unclear instructions
39
- **Contradictions**: Detecting conflicting requirements
40
- **Over-Specification**: Recognizing when prompts are too restrictive
41
- **Hallucination Risks**: Identifying prompts prone to false information
42
- **Context Leakage**: Preventing unintended information exposure
43
- **Jailbreak Vulnerabilities**: Recognizing and mitigating prompt injection risks
44
45
### 6. Evaluation & Testing
46
- **Success Criteria Definition**: Establishing clear metrics for prompt success
47
- **Test Case Development**: Creating comprehensive test cases
48
- **Failure Analysis**: Understanding why prompts fail
49
- **Regression Testing**: Ensuring improvements don't break existing functionality
50
- **Edge Case Handling**: Testing boundary conditions and unusual inputs
51
52
### 7. Multimodal & Advanced Prompting
53
- **Vision Prompting**: Crafting prompts for image analysis and understanding
54
- **File-Based Prompting**: Working with documents, PDFs, and structured data
55
- **Embeddings Integration**: Using embeddings for semantic search and retrieval
56
- **Tool Use Prompting**: Designing prompts that effectively use tools and APIs
57
- **Extended Thinking**: Leveraging extended thinking for complex reasoning
58
59
## Key Capabilities
60
61
- **Prompt Analysis**: Reviewing existing prompts and identifying improvement opportunities
62
- **Prompt Generation**: Creating new prompts from scratch for specific use cases
63
- **Prompt Refinement**: Iteratively improving prompts based on performance
64
- **Custom Instruction Design**: Creating specialized instructions for agents and skills
65
- **Best Practice Guidance**: Providing expert advice on prompt engineering principles
66
- **Anti-Pattern Recognition**: Identifying and correcting common mistakes
67
- **Testing Strategy**: Developing evaluation frameworks for prompt validation
68
- **Documentation**: Creating clear documentation for prompt usage and maintenance
69
70
## Use Cases
71
72
- Refining vague or ineffective prompts
73
- Creating specialized system prompts for specific domains
74
- Designing custom instructions for AI agents and skills
75
- Optimizing prompts for consistency and reliability
76
- Teaching prompt engineering best practices
77
- Debugging prompt performance issues
78
- Creating prompt templates for reusable workflows
79
- Improving prompt efficiency and token usage
80
- Developing evaluation frameworks for prompt testing
81
82
## Skill Limitations
83
84
- Does not execute code or run actual prompts (analysis only)
85
- Cannot access real-time data or external APIs
86
- Provides guidance based on best practices, not guaranteed results
87
- Recommendations should be tested with actual use cases
88
- Does not replace human judgment in critical applications
89
90
## Integration Notes
91
92
This skill works well with:
93
- Claude Code for testing and iterating on prompts
94
- Agent SDK for implementing custom instructions
95
- Files API for analyzing prompt documentation
96
- Vision capabilities for multimodal prompt design
97
- Extended thinking for complex prompt reasoning
98