DOCUMENTATION.md

PROMPT-TOOL 🤖

Overview

Prompt-tool is a Python-based utility designed to enhance and structure prompts for Large Language Models (LLMs) and AI agents, especially for coding and developer productivity tasks. It provides a Gradio web interface and an MCP server endpoint for generating context-aware, high-quality prompts, optionally leveraging available tools in the host environment.

Features

• Prompt Engineering: Generates detailed, structured prompts for LLMs, focusing on coding tasks.
• Tool Awareness: Can incorporate available tools (e.g., Playwright) into prompt instructions.
• Multiple Models: Supports different LLM models (gpt-4.1-nano, gpt-4.1-mini) for basic and advanced prompt generation.
• Web Interface: Gradio-based UI for interactive prompt generation.
• MCP Server Integration: Exposes prompt-tool as an MCP tool for programmatic access.
• Customizable Instructions: Uses a configurable prompt template (see prompts/coding.txt).

Installation

1. Clone the repository and navigate to the work/prompt-tool directory.
2. Install dependencies:

   pip install -r requirements.txt
   

3. Create a .env file with your OpenAI API key:

   OPENAI_API_KEY=your_openai_api_key_here
   

Usage

Gradio Web Interface

Run:

python mcp_gradio.py

• Enter your prompt, select the model (A: advanced, B: basic, N: no tooling), and specify available tools (comma-separated).
• Click "Generate" to receive an enhanced prompt.

Note: When running the Gradio interface, an MCP server is also created and exposed automatically.

MCP Server

Run:

python mcp_server.py

This exposes the prompt-tool as an MCP tool endpoint for integration with other MCP-compatible systems.

Programmatic Usage

You can import and use the core functions in your own Python scripts:

from generator import prompt_tool
result = prompt_tool("Your prompt here", tool="A", tools="playwright"

Architecture

• generator.py: Core logic for prompt generation, model selection, and OpenAI API interaction.
• mcp_gradio.py: Gradio web interface for interactive use.
• mcp_server.py: MCP server exposing the prompt-tool as an endpoint.
• prompts/coding.txt: Template and guidelines for prompt generation.
• requirements.txt: Python dependencies.

Prompt Template

The prompt-tool uses a template (prompts/coding.txt) that enforces best practices in prompt engineering, such as specificity, structured frameworks, tool specification, output format, and constraints.

This allows the user to use prompt-tool for whatever they want—this means that the tool can be tailored for tasks other than coding just by changing the template and a couple of lines of code.

Cursor integration

In order to use prompt-tool in Cursor, you need to add the following to your .cursor/mcp.json file:

    {
    "mcpServers": {
        "prompt-tool": {
        "url": "your_mcp_server_url"
        }
    }
    }

Development & Contribution

• Exclude .env, venv/, and __pycache__/ from version control (see .gitignore).
• To contribute, fork the repository, create a feature branch, and submit a pull request.
• Please ensure code is well-documented and tested.

License

MIT License

---

This documentation will be updated to include images and diagrams in the future.