This handbook provides a structured guide for "ChatGPT Prompt Engineers" to maximize efficiency when collaborating with ChatGPT for programming tasks. It offers a curated list of effective prompts and techniques for various development workflows, aiming to elicit more targeted and consistent responses from the AI.

How It Works

The handbook categorizes prompts based on their intended use, such as setting conversational context, refining AI output, and generating code or documentation. It emphasizes techniques like using delimiters (e.g., 「」, {}) to isolate content for processing and leveraging ChatGPT's contextual memory for iterative refinement. The guide also notes that English prompts generally yield better results for code-related tasks, even when requesting Chinese output.

Quick Start & Requirements

This is a documentation repository; no installation or execution is required. Users interact directly with ChatGPT via its web interface or API.

Highlighted Details

• Provides specific prompt examples for tasks like code optimization, text refinement, requirements analysis, and code generation in various languages.
• Includes methods for managing conversational context and setting default response languages.
• Offers prompts for code review, error analysis, refactoring, and generating unit tests and mock data.
• Discusses "command stability," a subjective measure of prompt reliability, categorizing prompts as strong, medium, or weak.

Maintenance & Community

The project is maintained by wuchangming and welcomes contributions via Pull Requests to expand the prompt library. Links to related resources like OpenAI's documentation and GitHub Copilot are provided.

Licensing & Compatibility

The repository itself is not associated with a specific open-source license. Usage of ChatGPT is governed by OpenAI's terms of service.

Limitations & Caveats

The effectiveness of prompts, particularly for code generation, is noted as "weak" and highly dependent on the clarity of the request. The handbook acknowledges that discovering more reliable prompting patterns requires ongoing exploration.