skills.rest
Tgoldi
TgoldiCOMMUNITY
Explore all skills by this author →
GitHub

prompt-engineering-patterns

Design, test, and optimize LLM prompts using few-shot, chain-of-thought, and structured output patterns.

1|Updated May 10, 2026
Search Tags:#prompt engineering#pydantic#prompt optimization#chain-of-thought#llm#few-shot learning#structured output
One-click install
npx skills add https://github.com/Tgoldi/claude-skills --skill prompt-engineering-patterns-tgoldi
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: prompt-engineering-patterns
Source: https://github.com/Tgoldi/claude-skills/tree/main/prompt-engineering-patterns
Command: npx skills add https://github.com/Tgoldi/claude-skills --skill prompt-engineering-patterns-tgoldi

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires numpy, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve? LLM prompts often produce inconsistent, unparseable, or low-quality outputs in production. This Skill provides proven prompt engineering patterns and tooling to make LLM outputs reliable, structured, and measurable. ## Core Features & Use Cases - Prompt Pattern Library: Implement few-shot learning, chain-of-thought reasoning, tree-of-thought, self-consistency, and structured JSON outputs with Pydantic schema enforcement. - Prompt Optimization: Run A/B tests, measure accuracy/latency/token metrics, and iteratively refine prompts using the included optimization script. - Template Systems: Build reusable, composable prompt templates with variable interpolation, conditional sections, and role-based system prompts. - Use Case: You are building a sentiment analysis feature and getting inconsistent outputs. Use this Skill to apply structured output with a Pydantic schema, add dynamically selected few-shot examples, and A/B test prompt variations until accuracy exceeds your target. ## Quick Start Ask the AI to rewrite your prompt using chain-of-thought reasoning and structured JSON output with a Pydantic schema, then evaluate it against your test cases.

Frequently Asked Questions about prompt-engineering-patterns

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get structured JSON output from Claude in Python?▼

Use Pydantic models with LangChain's with_structured_output method or instruct the model to respond with JSON matching an explicit schema, then parse with json.loads. Wrap parsing in try/except to handle malformed outputs with a fallback prompt.

How to implement few-shot learning with dynamic example selection?▼

Embed your examples and queries with a model like VoyageAI, then use SemanticSimilarityExampleSelector with a vector store like Chroma to retrieve the k most similar examples per query. Insert selected examples into the prompt before the user input.

When should I use chain-of-thought prompting?▼

Use chain-of-thought for math, logical reasoning, multi-step planning, and code debugging tasks. Skip it for simple factual lookups, creative writing, or latency-sensitive applications where step-by-step reasoning adds unnecessary tokens.

How do I A/B test two prompt variations?▼

Run both prompts against the same test suite, measuring accuracy, latency, and token usage per variant. Use a statistical significance test like a t-test on the results to confirm the winner rather than relying on small sample differences.

Why does my prompt produce inconsistent outputs?▼

Inconsistency usually comes from vague instructions, missing output format constraints, or untested edge cases. Add explicit format requirements, representative few-shot examples, and a verification step, then measure consistency across repeated runs.

How can I reduce prompt token usage without losing quality?▼

Remove filler phrases, consolidate multi-sentence instructions into numbered steps, and use prompt caching for repeated system prompts via cache_control. Measure accuracy before and after to confirm quality is preserved.

Related Skills

Explore More Software Engineering Skills→
tanveerriaztanveerriaz
1

llm-eval-golden-set

Builds a golden set of test cases and an eval runner to detect LLM output regressions.

Community
Intermediate
SanketAdlakSanketAdlak

prompt-engineering-patterns

Design, test, and optimize LLM prompts using few-shot, chain-of-thought, and structured output patterns.

Community
Advanced
JCETools-PetraJCETools-Petra
16

ai-llm-engineering

Design RAG pipelines with chunking, embedding, retrieval, and evaluation metrics.

Community
Advanced
lebsrallebsral
11

ai-fine-tuning

Fine-tune models on labeled datasets using BootstrapFinetune workflows.

Community
Advanced