Skip to content

PromptCraft

A public collection of system prompts from ChatGPT-based products and custom GPTs, for studying prompt writingLouisShark/chatgpt_system_prompt

Stars
10.8k
Forks
1.5k
Licence
MIT
Language
HTML
Last update
24 days ago
Type
Learning resource

What it is

This repository gathers system prompts (the hidden instructions that shape how an AI assistant behaves) from ChatGPT and various other AI products and custom GPTs. It also includes material on prompt injection and prompt leaking, which are ways people try to extract those hidden instructions. The README presents it mainly as an educational resource for learning how system prompts are written and how to protect them.

  • Browse the table of contents in TOC.md to find the instructions for a specific custom GPT
  • Study how different GPTs structure their instructions as a model for your own prompts
  • Read the security guide to learn how to protect your own GPT instructions
  • Explore the linked learning resources for prompt engineering

Great for

  • People learning how to write system prompts for custom GPTs
  • Developers who want examples of how others structure AI instructions
  • Security-minded builders studying how prompt leaking works so they can guard their own GPTs
  • Students of prompt engineering looking for real-world examples

Not ideal for

  • Anyone expecting ready-to-run software, since the README describes a collection of text, not an application
  • People who need a verified or current list of official vendor prompts, because the README does not state how or when the prompts were obtained or confirmed
  • Users looking for a library or tool to integrate into a product
Step-by-step guide: how to use PromptCraftInstall, first run, examples, tips and FAQs

How to get it

  1. 1

    Clone and build from source

    Clone the repository

    git clone https://github.com/LouisShark/chatgpt_system_prompt
  2. 2

    Run it

    • Open the repository on GitHub or clone it locally
    • Open TOC.md and use Ctrl + F to search for the GPT you want
    • Read the prompt or knowledge file for that GPT
    • Read SECURITY.md or GETTING_STARTED.md for guidance on the topic
  3. 3

    Get the most out of it

    • Use the idxtool mentioned in the README (scripts/README.md) to search the collection if you cloned the repo
    • Read the Getting Started Guide to learn how system prompts and knowledge files are obtained
    • Read the Security Guide to understand how to protect your own GPT instructions
    • Check the Learning Resources page for tools and materials on prompt engineering

Downloads

Strengths

  • Large, popular collection with over 10,000 stars on GitHub
  • Released under the permissive MIT license
  • Organized with a searchable table of contents and linked guides on getting prompts, security, and contributing

Trade-offs

  • The README frames the material as reference and awareness content, and it does not explain how each prompt was verified
  • The README is truncated here, so the full scope and maintenance details are not visible in this analysis
  • It is a collection of text files rather than software, so there is nothing to install or run beyond reading

How it compares

ProjectStarsTypeLevelUpdatedBest for
chatgpt_system_prompt (this)10.8kLearning resourcebeginner24 days agoPeople learning how to write system prompts for custom GPTs
promptflow11.3kDev toolintermediate6 days agoTeams with existing Prompt flow applications that need to keep running until migration
generative-ai-for-beginners123kLearning resourcebeginner5 days agoDevelopers who know some Python or TypeScript and want a guided introduction to generative AI
botpress15kDev toolintermediate6 days agoDevelopers who want to write integrations and bots as TypeScript code using the Botpress CLI and SDK
Prompt-Engineering-Guide79.8kLearning resourcebeginner8 months agoPeople who are new to prompting and want structured lessons with examples
awesome-gpt-image-235.4kLearning resourcebeginner23 days agoPeople who want working GPT Image 2 prompts to copy and adapt

Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.