What is the 'Progressive Disclosure' pattern?

It is an iterative approach included in this skill that starts with simple prompts and adds layers of constraints and reasoning only as needed to prevent over-engineering.

How does this skill improve LLM output reliability?

It implements structured patterns like Chain-of-Thought, few-shot examples, and verification steps that guide the model through logical stages, reducing hallucinations and inconsistent formatting.

Does this skill support dynamic data in prompts?

Absolutely. It provides template patterns for variable interpolation and conditional logic, allowing you to inject real-time data or retrieved context into your prompts.

Is this skill specific to a certain programming language?

While the examples are provided in Python, the prompt engineering patterns themselves are language-agnostic and can be applied to any application calling an LLM API.

Can I use this to reduce my API token costs?

Yes, the skill includes performance optimization patterns specifically designed to consolidate instructions and remove redundant content without sacrificing output quality.

Prompt Engineering Patterns

Name: Prompt Engineering Patterns
Author: kivilaid

bykivilaid

•

データサイエンスとML

Implements advanced prompt engineering techniques to maximize LLM performance, reliability, and controllability in production environments.

The Prompt Engineering Patterns skill provides a comprehensive framework for building production-grade LLM applications using sophisticated prompting strategies. It enables developers to implement complex reasoning patterns such as Chain-of-Thought, manage dynamic few-shot learning systems, and create modular, version-controlled prompt templates. By moving beyond basic instructions to structured architectural patterns, this skill ensures AI outputs are consistent, cost-effective, and highly reliable for critical software tasks and automated workflows.

主な機能

01Dynamic few-shot learning with semantic example selection

02Structured reasoning elicitation using Chain-of-Thought and verification steps

03Reusable prompt template systems with variable interpolation

04Specialized system prompt design for role-based AI behavior

056 GitHub stars

06Iterative optimization workflows for token efficiency and latency

ユースケース

01Designing complex reasoning flows for production LLM applications

02Scaling AI systems with dynamic context and example retrieval

03Optimizing prompt consistency and reducing output variance

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add kivilaid/plugin-marketplace prompt-engineering-patterns

For use in Claude.ai and ChatGPT

Download Skill