louisshark/chatgpt_system_prompt

★ 10,563HTMLAudience · generalComplexity · 1/5Setup · easy

Mindmap

mindmap
  root((chatgpt-prompts))
    What it does
      Prompt collection
      Education resource
      Security examples
    Content
      Custom GPT prompts
      Injection examples
      Protection guides
    Use Cases
      Learn prompt design
      Build custom GPTs
      Security study
    Tools
      Table of contents
      CLI search tool

mindmap root((chatgpt-prompts)) What it does Prompt collection Education resource Security examples Content Custom GPT prompts Injection examples Protection guides Use Cases Learn prompt design Build custom GPTs Security study Tools Table of contents CLI search tool

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Study real system prompts from popular custom GPTs to learn how to write effective instructions for your own AI assistant

USE CASE 2

Find examples of how GPT builders protect their system prompts from being revealed through prompt injection

USE CASE 3

Reference existing prompts when building a custom GPT for a specific use case like customer support or coding help

Tech stack

HTML

Getting it running

Difficulty · easy Time to first run · 5min

In plain English

This repository is a collection of system prompts used by ChatGPT custom GPTs and other AI products. A system prompt is the hidden set of instructions that shapes how an AI assistant behaves in a particular context, for example telling it to act as a customer support agent or a coding helper. The collection here is shared for educational purposes, specifically to help people understand how these prompts are written and to illustrate security considerations around keeping them private. The main content is a table of contents file listing the available prompts organized by the name of the custom GPT they came from. You can search that file for a specific GPT name to find its prompt. There is also a command-line tool in the scripts folder called idxtool for searching if you have cloned the repository locally. Separate guides cover how to retrieve system prompts, how to protect your own GPT instructions from being leaked, how to contribute new prompts, and where to find additional learning resources. The project also documents prompt injection, which is a technique where someone tries to override or reveal a system prompt by crafting specific inputs. The README states that many GPT authors have responded to this project by improving how they protect their prompts, which the project treats as a positive outcome aligned with its educational purpose. The repository is primarily useful to people who want to study how effective system prompts are structured, anyone building their own custom GPTs who wants examples to learn from, or developers thinking about how to keep their own instructions secure. No installation is needed to browse the content, and the table of contents file is the main entry point for exploring what is available.

Copy-paste prompts

Prompt 1

Search this collection for a system prompt from a coding assistant GPT and adapt it to create my own code review bot.

Prompt 2

Show me examples from this repo of how custom GPT system prompts handle attempts to get the AI to reveal its instructions.

Prompt 3

Based on the system prompts in this collection, write a system prompt for a customer support bot that keeps its instructions confidential.

Prompt 4

Find a writing assistant system prompt in this repo and explain what makes its instruction structure effective for that use case.

Open on GitHub → Explain another repo

← louisshark on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.