This repository contains a collection of example answers for the "What traits should ChatGPT have?" section of the Custom Instructions feature, created on April 29, 2025. These examples are intended to serve as templates or inspiration for users who want to configure ChatGPT's behavior to better match their preferences.
Additionally, this repository includes the system prompt used for a specialized assistant. This assistant can be implemented as a Custom GPT or adapted into any tool, and is designed to help users easily generate valid "traits" entries that respect formatting and character limit requirements.
- Example trait customizations (under 1,500 characters each)
- System prompt for creating a traits-writing assistant
| File | Description | Path |
|---|---|---|
| flowery | Flowery and elaborate language style | flowery |
| formal-and-mannerly | Formal and polite communication style | formal-and-mannerly |
| just-be-a-bot | Direct and technical bot-like responses | just-be-a-bot |
| minimal-answers | Concise and to-the-point responses | minimal-answers |
| short-and-direct | Brief and straightforward communication | short-and-direct |
| assistant/systemprompt | System prompt for traits-writing assistant | assistant/systemprompt |
| for-fun/begrudging | Reluctant and unenthusiastic responses | for-fun/begrudging |
| for-fun/condescending | Patronizing and superior tone | for-fun/condescending |
| for-fun/disclaimer-every-turn | Responses with constant disclaimers | for-fun/disclaimer-every-turn |
| for-fun/human-is-a-bot | Pretending the human is the AI | for-fun/human-is-a-bot |
| for-fun/just-google-it | Dismissive "just Google it" responses | for-fun/just-google-it |
| for-fun/not-very-good | Self-deprecating and unconfident responses | for-fun/not-very-good |
| for-fun/questioner | Always responds with questions | for-fun/questioner |
| for-fun/shakespearean | Responses in Shakespearean English | for-fun/shakespearean |