Prompt Debugging System Prompts

This repository is a small collection of system prompts designed for debugging prompts — whether system prompts, user prompts, or instructional prompts — across different use cases.

The collection spans both conversational and instructional contexts. It includes focused system prompts for areas such as tool/function calling, RAG (Retrieval-Augmented Generation) behavior, search result relevance, and structured output validation (e.g., JSON schemas).

While AI tools are traditionally seen as the responders to prompts, they can also serve as highly effective analyzers — applying their language understanding abilities to help users perform "self-examination" of prompts and identify why deviations occurred between expected and actual model outputs.
Of course, human intelligence and strong prompt engineering skills remain essential, but AI can serve as a useful assistant in this process.

These system prompts can be configured as AI assistants on any platform that supports setting custom system prompts (such as OpenAI, OpenRouter, LM Studio, etc.).

System Prompts Index

This repository contains the following system prompts for debugging different types of prompts:

Prompt	Description
	A general-purpose debugger for analyzing and improving both system and user prompts
	Specifically focused on debugging and improving user prompts
	Specifically focused on debugging and improving system prompts
	For diagnosing and correcting issues related to context retrieval and RAG behavior
	For debugging prompts that drive search functionality and improving search result relevance
	For debugging issues with tool, function, or MCP invocation in conversational contexts
	For debugging prompts that produce structured outputs (such as JSON) according to an intended schema