What is 'prefilling' in this context?

Prefilling is a Claude-specific technique where you provide the start of the assistant's response to force a specific output format, such as JSON, without any conversational preamble.

Can I use this for models other than Claude?

Yes, the skill includes specific workflows and formatting patterns for both Claude (XML-based) and GPT (Markdown-based) to ensure optimal performance across different providers.

Why does the skill emphasize XML tags?

Claude was specifically trained to understand and respect boundaries defined by XML tags, making them the most reliable way to separate context, tasks, and constraints.

How does this skill improve Claude's responses?

It applies research-backed structures like XML tagging and chain-of-thought instructions that help the model reason more effectively and follow constraints accurately.

Meta Prompt Creator

Name: Meta Prompt Creator
Author: Melvynx

byMelvynx

•

128

•

開発者ツール

Generates high-performance prompts for Claude, GPT, and other LLMs using advanced prompt engineering techniques and research-backed structures.

The Meta Prompt Creator skill transforms Claude into an expert prompt engineer, enabling the creation of structured, effective prompts for any task. It leverages sophisticated methodologies like XML tagging, chain-of-thought reasoning, and few-shot examples to optimize model performance and output reliability. Whether you are building complex system prompts for AI agents or refining simple user instructions, this skill provides a standardized workflow to ensure clarity, handle edge cases, and enforce specific output formats across different AI providers.

主な機能

01Model-specific optimization for Anthropic Claude and OpenAI GPT architectures

02Standardized XML and Markdown prompt structuring for multi-model compatibility

03128 GitHub stars

04Anti-pattern detection to eliminate vague language and negative prompting

05Few-shot example generation to guide specific output formatting

06Chain-of-thought reasoning and step-by-step logic implementation

ユースケース

01Creating complex reasoning instructions for data extraction and multi-step analysis

02Building robust system prompts for specialized AI agents and automated workflows

03Optimizing existing prompts that produce inconsistent or hallucinated results

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add melvynx/aiblueprint meta-prompt-creator

For use in Claude.ai and ChatGPT

Download Skill