This project gives you 10 prompt engineering patterns based on Claude Code behavior. It is a Python app, so you run it with Python on Windows. It helps you work with prompts in a more structured way.
Use it if you want:
- clearer prompts
- more consistent AI output
- reusable prompt templates
- a local Python tool you can run on your PC
Before you start, make sure you have:
- Windows 10 or Windows 11
- Internet access
- Python 3.10 or newer
- Git, if you want to clone the repo
- A folder where you can save files
If you do not have Python yet, install it first from the official Python site. During setup, keep the option to add Python to PATH turned on.
Go to this page to download the project:
On the page, click the green Code button, then choose one of these options:
- Download ZIP
- Open with GitHub Desktop
- Copy the repository link for Git
If you want the easiest path, choose Download ZIP. Then unzip the file into a folder on your computer.
After you download the project, open the folder.
You should see files for a Python project, such as:
- Python source files
- a readme file
- dependency files
- prompt pattern data or examples
Keep the folder in a place you can find again, such as:
- Downloads
- Documents
- Desktop
If you use ZIP, right-click the file and choose Extract All.
Open Command Prompt in the project folder.
You can do this by:
- Opening the folder in File Explorer
- Clicking the address bar
- Typing cmd
- Pressing Enter
Then install the packages with pip if the project includes a requirements file:
pip install -r requirements.txtIf there is no requirements file, the project may run with standard Python only.
Use the main Python file in the project folder. Common file names are:
- main.py
- app.py
- run.py
Run it from Command Prompt like this:
python main.pyIf the project uses a different file name, replace main.py with the correct one.
If the app opens in a terminal menu, follow the on-screen steps. If it opens a web page or local interface, keep the window open while you use it.
This tool focuses on prompt structure. It uses patterns that help you shape requests for AI tools in a better way.
You may see patterns such as:
- role setup
- task framing
- context blocks
- output rules
- multi-step instructions
- constraint-based prompts
- review and correction prompts
- agent-style prompts
These patterns help you turn a vague request into a clear one.
You can use this project for tasks like:
- writing better AI prompts
- building prompt templates
- testing different prompt styles
- organizing work for AI agents
- improving consistency in Claude Code workflows
- learning how system prompts shape output
If you use AI tools for coding, writing, or analysis, these patterns can help you get cleaner results.
A simple way to use the project is:
- Pick a prompt pattern
- Add your task
- Add the context the AI needs
- Set output rules
- Review the result
- Refine the prompt if needed
For example, if you want help with code, you can give:
- the goal
- the language
- the file type
- the expected result
- any limits on the output
This reduces guesswork and makes the result easier to use.
This project is built for people who want a practical way to work with prompt design. It aims to help you:
- write prompts with less trial and error
- keep prompt structure consistent
- reuse patterns across tasks
- improve output from AI agents
- understand what makes a prompt work
Because it is a Python project, it is easy to inspect and adapt if you later want to edit the code.
If the app does not start, check these items:
- Python is installed
- Python is added to PATH
- you are in the right folder
- the required packages are installed
- you typed the file name correctly
If Windows says Python is not found, close Command Prompt and reopen it after you install Python.
If pip fails, try:
python -m pip install -r requirements.txtIf the app still does not run, look for the main Python file in the project folder and try that file name instead.
This repository is centered on:
- AI agents
- Anthropic
- Claude
- Claude Code
- prompt engineering
- system prompts
- Python
- multi-agent workflows
- cybersecurity-aware use
- LLM output control
It is aimed at people who want a clear set of patterns they can apply in real work.
For the smoothest setup on Windows:
- keep the project in a simple folder path
- avoid folders with special characters
- use Command Prompt or PowerShell
- install Python once and keep it updated
- unzip the project before running it
If Windows blocks the files after download, right-click the ZIP or folder, open Properties, and check for an unblock option if it appears.
When you open the project, look for files that tell you how it runs:
- README.md
- requirements.txt
- main.py
- app.py
- examples
- prompt patterns
- config files
The README file in the repo may also show the exact command to start the app.
When you use the patterns, keep your prompt simple and specific.
Good prompts usually include:
- what you want done
- what the input is
- what form the output should take
- what the AI should avoid
- any length or style limits
If the output is not right, change one part at a time. That makes it easier to see what helped.
If you need the project again, use this page:
Check the repository for the license before you use the code in your own projects. The project description says it is a Python implementation with no proprietary code, so you can inspect and adapt it for local use within the terms set by the repo
Open the project, run the main Python file, and try one prompt pattern with a task you already know well