Prompt Engineer
Design, test, and optimize prompts for reliable LLM performance
What You Can Do
You can transform underperforming or inconsistent prompts into reliable, production-ready specifications that work across diverse inputs and use cases. This skill guides you through diagnosing prompt failures, applying evidence-based optimization techniques (chain-of-thought reasoning, few-shot examples, output constraints), testing variants systematically, and documenting final prompts for team-wide reproducibility. You'll move beyond trial-and-error to engineering prompts that scale reliably across different model versions and maintain quality with thousands of inputs.
Features
structure prompts to break problems into logical steps for improved reasoning and transparency
craft and validate examples that demonstrate desired behavior patterns
define JSON, markdown, or custom schemas to ensure consistent, parseable results
create controlled test suites to compare prompt versions and measure performance gains
establish baseline metrics and track improvements across iterations
identify why prompts fail (ambiguous instructions, missing context, wrong model assumptions) and target fixes
generate final prompt specs with clear instructions, examples, constraints, and usage guidelines
test and adapt prompts for different LLM versions and families
Example Output
Example 1: Customer Support Classification Prompt
Input: Raw customer ticket text with high variability
Optimized prompt output:
## Classification Prompt Specification
### Purpose
Classify support tickets into: Billing, Technical, Product Feedback, Account Access
### Chain-of-Thought Structure
1. Identify the primary problem statement
2. Note any secondary concerns
3. Match against category definitions
4. Assign confidence score (0-1)
### Few-Shot Examples
- Ticket: "Can't log in after password reset" → Account Access (0.95)
- Ticket: "Charge appeared twice on invoice" → Billing (0.92)
### Output Format
{"category": "string", "confidence": 0.0-1.0, "reasoning": "2-3 sentences"}
### Quality Baseline
- Accuracy target: 92%+
- Consistency: 95%+ same-class agreement on retest
Example 2: Data Extraction Prompt Test Results
Variant A (baseline): 78% accuracy on 50-document sample Variant B (added CoT + constraints): 89% accuracy Variant C (added few-shot examples): 94% accuracy
Recommendation: Use Variant C across production pipeline
Example 3: JSON Structured Output Prompt
Input: Product review text Output:
{
"sentiment": "positive",
"topics": ["durability", "price"],
"actionable_feedback": true,
"priority_level": "medium",
"summary": "Customer praised build quality but flagged cost as concern"
}
What's Included
- SKILL.md: core instruction file with methodology and examples
- Prompt Testing Framework: template for creating and comparing prompt variants with consistency metrics
- Chain-of-Thought Design Checklist: step-by-step guide for structuring reasoning prompts
- Few-Shot Example Generator: worksheet for crafting representative examples and validating coverage
- Output Schema Templates: pre-built JSON and markdown structures for common tasks (classification, extraction, summarization)
- Specification Documentation Template: final prompt specification format for handoff and team use
Who It's For
- Prompt engineers designing and optimizing prompts for production LLM applications
- AI/ML teams scaling prompt quality across multiple use cases and models
- Product managers ensuring consistent AI feature performance at scale
- Data scientists building reliable extraction and classification pipelines
- Operations leads documenting and standardizing prompts across teams
Best For
- Diagnosing and fixing inconsistent or underperforming prompts
- Engineering customer-facing prompts (support, recommendations, content generation)
- Building data extraction and classification pipelines at scale
- Testing and comparing prompt variants with measurable quality metrics
- Documenting production prompts for reproducibility and team handoff
- Optimizing prompts for reliability across different model versions
You might also like
This skill helps you systematically diagnose failures in distributed ML training and serving infrastructure. You provide system logs, metrics, and error traces, and Claude performs structured root-cause analysis to identify the underlying issue—whether it's resource exhaustion, distributed system deadlock, data pipeline corruption, or model serving misconfiguration. You get a detailed diagnosis with remediation steps ranked by likelihood and implementation effort.
by moonwell-fi
You can systematically diagnose why your RAG system isn't returning relevant results and identify specific misalignments between queries and context. This skill walks you through analyzing retrieval quality, scoring context relevance, and optimizing prompt-retrieval alignment to improve your RAG accuracy. You'll get concrete, actionable recommendations for fixing bottlenecks in your pipeline.
by moonwell-fi
You can prepare your domain-specific datasets, train fine-tuned Claude models optimized for your use case, benchmark performance against base models, and deploy production-ready custom models. This skill automates the entire fine-tuning pipeline from raw data to live inference, complete with cost-benefit analysis and version control.
by moonwell-fi
You get structured guidance for designing controlled experiments, selecting appropriate metrics, and analyzing results with statistical rigor. Claude helps you identify confounding variables, calculate required sample sizes, interpret statistical significance, and document methodology for reproducible benchmarking that stands up to scrutiny.
by moonwell-fi
You can structure comprehensive health technology assessments (HTAs) that organize clinical evidence, economic analyses, and regulatory considerations into evidence-based coverage recommendations. This skill helps you synthesize clinical trial data, health economic models, and real-world evidence into clear, defensible payer coverage determinations. You'll generate professional HTA reports that align with major frameworks like ICER, CADTH, and NICE standards.
by moonwell-fi
This skill lets you leverage Gemini CLI as a complementary AI tool within your Claude-driven workflows. You can request code reviews from a different AI perspective, access current web information through Google Search integration, analyze complex codebase architectures, run parallel code generation tasks, and handle specialized operations like test suite generation or documentation—all while Claude coordinates the orchestration and results integration.
You can systematically diagnose performance issues across your ML infrastructure, identify root causes of distributed training failures, and receive actionable optimization recommendations. This skill analyzes metrics, logs, and configuration data to pinpoint resource constraints, communication overhead, and configuration misalignments that impact training speed and reliability.
by moonwell-fi
This skill transforms high-level feature requests into actionable implementation plans by conducting deep requirement interviews, analyzing existing codebases (or planning greenfield projects), and producing comprehensive handoff documents. It distinguishes between brownfield development (with existing code to integrate) and greenfield projects (starting from scratch), ensuring your technical strategy accounts for your actual constraints and patterns.
by moonwell-fi