refine-prompt

Optimize LLM prompts and append the final prompt to .ai/PROMPT.md.

69|3|Updated Feb 5, 2026

Search Tags:#prompts #optimize #prompt-engineering #prompt-design #LLMs #PROMPT.md

Ready to run

npx skills add https://github.com/PaulRBerg/agent-skills --skill refine-prompt

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: refine-prompt
Source: https://github.com/PaulRBerg/agent-skills/tree/main/skills/refine-prompt
Command: npx skills add https://github.com/PaulRBerg/agent-skills --skill refine-prompt

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill optimizes prompts for LLMs to maximize clarity, structure, and actionable results.

Core Features & Use Cases

Prompt design for few-shot reasoning, role assignments, and explicit output formats across AI assistants.
Generates a complete, deployment-ready prompt block and appends it to .ai/PROMPT.md.
Supports configurable behavior via frontmatter to tailor prompts for different tasks and models.

Quick Start

Use the refine-prompt skill to optimize the prompt "Summarize quarterly results and outline next steps" and append the result to .ai/PROMPT.md.

Frequently Asked Questions about refine-prompt

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

FAQPage Schema

How do I optimize prompts for LLMs to get better results?▼

Optimizing prompts for LLMs involves structuring them with clear role assignments, few-shot examples, and explicit output formats to maximize clarity and actionability. This skill applies these principles systematically and saves the refined prompt to .ai/PROMPT.md for reuse across your project.

What's the best way to design prompts for few-shot reasoning?▼

Few-shot prompt design includes providing example input-output pairs and explicit instructions for the task. This skill generates deployment-ready prompts with few-shot patterns built in, then persists them to .ai/PROMPT.md so your team can reference and iterate on consistent prompt structures.

Can I save and reuse optimized prompts across my project?▼

Yes. This skill generates a complete prompt block and appends it to .ai/PROMPT.md, creating a centralized artifact your project can reference. This approach keeps prompts versioned and accessible to all AI assistants and team members working on the project.

How do I configure prompt optimization for different tasks?▼

The skill supports configurable behavior via YAML frontmatter, letting you tailor prompts for specific tasks and model requirements. Configuration controls how the prompt is structured, what role or context is assigned, and what output format is enforced before saving to .ai/PROMPT.md.

What makes a prompt structure more effective for AI assistants?▼

Effective prompts combine role clarity, structured examples, and explicit output specifications. This skill applies prompt-engineering best practices to transform ambiguous requests into well-formed, actionable prompts that produce consistent, higher-quality results from LLMs.