skills.rest
tronghieu
tronghieuCOMMUNITY
Explore all skills by this author →
GitHub

system-prompt-creator

Create structured system prompts for LLMs using layered drafting and review.

36|16|Updated May 1, 2026
Search Tags:#prompt engineering#prompt optimization#agent design#system prompts#llm instructions#model alignment#structured prompting
One-click install
npx skills add https://github.com/tronghieu/lumina-wiki --skill system-prompt-creator
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: system-prompt-creator
Source: https://github.com/tronghieu/lumina-wiki/tree/main/.agents/skills/system-prompt-creator
Command: npx skills add https://github.com/tronghieu/lumina-wiki --skill system-prompt-creator

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill removes guesswork from prompt engineering by helping you create clear, structured, model-aware system prompts that reliably steer an LLM’s behavior.

Core Features & Use Cases

  • Target-model prompting: Produces prompts tailored for major LLM families (Claude, GPT, Gemini, and model-agnostic setups).
  • Structured, research-backed workflow: Uses an interview→analysis→layered drafting→principle-based optimization→review loop to improve quality.
  • Reusable system prompt architecture: Generates prompts using consistent layered sections (role, context, instructions, output format, examples, tools, guardrails).
  • Playbook support for multi-scenario agents: Incorporates named operational playbooks so behavior stays consistent across different workflows.
  • Validation readiness: Includes model-independent checklists and guidance to make the prompt testable with validation prompts.

Quick Start

Provide the target model and your assistant’s use case, then request a complete system prompt following a structured, layered format with role, instructions, output format, examples, tools guidance, and guardrails.

Frequently Asked Questions about system-prompt-creator

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

FAQPage Schema
How do I create a system prompt that works reliably across different LLMs?▼

Model-aware system prompts require structured, layered drafting with role, context, instructions, output format, examples, tools, and guardrails to steer LLM behavior reliably. This approach tailors instructions for major model families while maintaining core functionality.

How do I design system prompts for LLM tool-use workflows and code assistants?▼

Design LLM tool-use workflows and code assistants using a layered system prompt architecture that includes explicit tools guidance, operational playbooks, and guardrails. This ensures consistent behavior across different workflows and edge-case scenarios.

Can I generate model-specific prompts for Claude, GPT, and Gemini?▼

Yes, you can generate model-specific prompts tailored for Claude, GPT, Gemini, and model-agnostic setups. The drafting process targets major LLM families to align instructions with specific model behaviors and capabilities.

How do I validate and optimize an LLM system prompt before deployment?▼

Validate and optimize an LLM system prompt using an explicit review checklist and test prompts. The process includes principle-based optimization and validation readiness checks to ensure the prompt performs correctly across common and edge-case scenarios.

Why does my LLM assistant behave inconsistently across different operational scenarios?▼

Inconsistent LLM behavior often stems from lacking named operational playbooks in the system prompt. Incorporating structured playbooks ensures behavior stays consistent across different multi-scenario agent workflows and edge cases.

Related Skills

Explore More Content & Communication Skills→
retz8retz8

lead-research-assistant

Analyze ICP and source target companies to produce outreach-ready lead briefs.

Community
Intermediate
gaoqiongxiegaoqiongxie
1

prompt-driven-dev

Version AI coding prompts in Git for deterministic, testable code generation.

Community
Advanced
Andre-cardiaAndre-cardia

aiox-master

Orchestrate cross-domain AI components and workflows with a unified command system.

Community
Advanced
jokken79jokken79
1

prompt-engineer

Design and optimize LLM prompts including system, few-shot, and chain-of-thought.

Community
Advanced