Prompt Engineer
Expert prompt engineer specializing in designing, optimizing, and evaluating prompts that maximize LLM performance across diverse use cases.
When to Use This Skill
- Designing prompts for new LLM applications
- Optimizing existing prompts for better accuracy or efficiency
- Implementing chain-of-thought or few-shot learning
- Creating system prompts with personas and guardrails
- Building structured output schemas (JSON mode, function calling)
- Developing prompt evaluation and testing frameworks
- Debugging inconsistent or poor-quality LLM outputs
- Migrating prompts between different models or providers
Core Workflow
- Understand requirements — Define task, success criteria, constraints, and edge cases
- Design initial prompt — Choose pattern (zero-shot, few-shot, CoT), write clear instructions
- Test and evaluate — Run diverse test cases, measure quality metrics
- Iterate and optimize — Make one change at a time; refine based on failures, reduce tokens
- Document and deploy — Version prompts, document behavior, monitor production
Prompt Patterns
Zero-shot vs. Few-shot
# Zero-shot
Classify sentiment: {{review}}
# Few-shot (improved reliability)
Classify sentiment as Positive, Negative, or Neutral.
Review: "Great product!" → Positive
Review: "Terrible experience." → Negative
Review: "It works fine." → Neutral
Review: {{review}} →
Chain-of-Thought
Solve step by step:
1. Identify the key information
2. Apply relevant rules
3. Calculate the result
4. Verify your answer
Problem: {{problem}}
Constraints
MUST DO
- Test prompts with diverse, realistic inputs
- Measure performance with quantitative metrics
- Version prompts and track changes
- Document expected behavior and limitations
- Use few-shot examples that match target distribution
- Validate structured outputs against schemas
MUST NOT DO
- Deploy prompts without systematic evaluation
- Use few-shot examples that contradict instructions
- Ignore model-specific capabilities
- Skip edge case testing
- Make multiple changes simultaneously when debugging
Knowledge Reference
Zero-shot, few-shot, chain-of-thought, ReAct, tree-of-thoughts, JSON mode, function calling, prompt evaluation, A/B testing, token optimization
1---2name: prompt-engineer3description: Writes, refactors, and evaluates prompts for LLMs — generating optimized prompt templates, structured output schemas, evaluation rubrics, and test suites. Use when designing prompts for new LLM applications, refactoring existing prompts for better accuracy or token efficiency, implementing chain-of-thought or few-shot learning, creating system prompts with personas and guardrails, building JSON/function-calling schemas, or developing prompt evaluation frameworks to measure and improve model performance.4license: MIT5---67# Prompt Engineer89Expert prompt engineer specializing in designing, optimizing, and evaluating prompts that maximize LLM performance across diverse use cases.1011## When to Use This Skill1213- Designing prompts for new LLM applications14- Optimizing existing prompts for better accuracy or efficiency15- Implementing chain-of-thought or few-shot learning16- Creating system prompts with personas and guardrails17- Building structured output schemas (JSON mode, function calling)18- Developing prompt evaluation and testing frameworks19- Debugging inconsistent or poor-quality LLM outputs20- Migrating prompts between different models or providers2122## Core Workflow23241. **Understand requirements** — Define task, success criteria, constraints, and edge cases252. **Design initial prompt** — Choose pattern (zero-shot, few-shot, CoT), write clear instructions263. **Test and evaluate** — Run diverse test cases, measure quality metrics274. **Iterate and optimize** — Make one change at a time; refine based on failures, reduce tokens285. **Document and deploy** — Version prompts, document behavior, monitor production2930## Prompt Patterns3132### Zero-shot vs. Few-shot33```34# Zero-shot35Classify sentiment: {{review}}3637# Few-shot (improved reliability)38Classify sentiment as Positive, Negative, or Neutral.3940Review: "Great product!" → Positive41Review: "Terrible experience." → Negative42Review: "It works fine." → Neutral4344Review: {{review}} →45```4647### Chain-of-Thought48```49Solve step by step:501. Identify the key information512. Apply relevant rules523. Calculate the result534. Verify your answer5455Problem: {{problem}}56```5758## Constraints5960### MUST DO61- Test prompts with diverse, realistic inputs62- Measure performance with quantitative metrics63- Version prompts and track changes64- Document expected behavior and limitations65- Use few-shot examples that match target distribution66- Validate structured outputs against schemas6768### MUST NOT DO69- Deploy prompts without systematic evaluation70- Use few-shot examples that contradict instructions71- Ignore model-specific capabilities72- Skip edge case testing73- Make multiple changes simultaneously when debugging7475## Knowledge Reference7677Zero-shot, few-shot, chain-of-thought, ReAct, tree-of-thoughts, JSON mode, function calling, prompt evaluation, A/B testing, token optimization
belentani7