prompt-engineeringlisted

# Prompt Engineering Skill > Craft effective prompts that get the best results from language models. ## Core Principle Prompts are programming for probabilistic systems. Clear instructions, good examples, and structured output formats dramatically improve results. ## Prompt Anatomy ``` ┌─────────────────────────────────────────┐ │ SYSTEM PROMPT (Role & Constraints) │ │ "You are a senior code reviewer..." │ ├─────────────────────────────────────────┤ │ CONTEXT (Background Information) │ │ "The codebase uses TypeScript..." │ ├─────────────────────────────────────────┤ │ EXAMPLES (Few-Shot Learning) │ │ Input: X → Output: Y │ ├─────────────────────────────────────────┤ │ TASK (What to Do) │ │ "Review this pull request for..." │ ├─────────────────────────────────────────┤ │ FORMAT (Output Structure) │ │ "Respond in JSON with fields..." │ └─────────────────────────────────────────┘ ``` ## Prompting Techniques ### Zero-Shot Direct instruction without examples: ``` Classify this customer feedback as positive, negative, or neutral: "The product arrived late but works great." ``` **Best for**: Simple, well-defined tasks the model understands. ### Few-Shot Provide examples to demonstrate the pattern: ``` Classify customer feedback: Input: "Love it! Best purchase ever!" Output: positive Input: "Broken on arrival. Waste of money." Output: negative Input: "The product arri