chatgpt_system_prompt
A collection of ChatGPT and custom GPT system prompts with material on protecting instructions and prompt injection
Low risk
We rate an entry low when it mostly gives the agent instructions and reference material.
Why this level
- This is a reference collection of prompts to read; it runs no code
- The prompt injection section is for protecting your own projects, not for attacking others
Install
Manual install
Open the repository, find the prompt you need via the TOC.md table of contents by GPT name; the prompts live in the prompts folder. For local search in a clone there is the idxtool helper script. The material is for reading; no install is required.
This is third-party code. Review the repository files before installing.
What it does
The repository collects ChatGPT and many custom GPT system prompts in one place to study how they are written. The prompts are arranged in folders, with a master table of contents searchable by GPT name and a helper script for searching a clone. Alongside are a getting-started guide, a breakdown of techniques for protecting GPT instructions and a section on prompt injection as a risk authors learn to defend against. The material is meant for learning to write system prompts and raising security awareness.
Who it is for. For prompt authors and developers studying how system prompts are built and how to protect custom GPTs.
Good fit when
- You want system prompt examples to learn from
- You are studying how to protect your GPT instructions
- You are learning about prompt injection risks
Not a fit when
- You need an installable tool, not a collection of examples
- The goal is to bypass others' protections or extract their instructions without permission
Example request
Show examples of custom GPT system prompts and the techniques authors use to protect their instructionsLimitations
This is a reference collection to read, not an installable tool. The prompts were gathered by many people, reflect the state at publication time and may be outdated. The prompt injection section is for protecting your own projects and for awareness; using it against others' systems without permission is not allowed.
How to disable. No separate install is needed. Stop using the material or delete the cloned copy of the repository.
Security check
- This is a reference collection of prompts to read; it runs no code
- The prompt injection section is for protecting your own projects, not for attacking others
README in short
The README describes a collection of ChatGPT and custom GPT system prompts for learning prompt writing. You find a prompt through the TOC.md table of contents by GPT name; the prompts live in the prompts folder, with an idxtool script for local search. There is a getting-started guide, a guide on protecting GPT instructions and material on prompt injection. The author stresses the material is for reference and security awareness. MIT licensed.
FAQ
How do I find a specific prompt?
Through the TOC.md table of contents by GPT name; the prompts themselves live in the prompts folder.
Why the prompt injection section?
So authors understand the risks and protect their GPT instructions better, which is the project's stated goal.
Related
Open-source personal AI assistant on your own machine: answers in Telegram, Slack, Discord and WhatsApp, extended with skills and plugins
A self-improving agent from Nous Research with a TUI, messaging gateway, cron jobs and skills it writes itself
An open source coding agent for the terminal and desktop with build and plan modes
Open prompt library with a Claude Code plugin, MCP server and CLI: search, fetch and improve prompts and skills from an agent