Skip to content

🚀 Quickstart: 5-Minute Onboarding

Welcome to the Proompts repository! If you're a new developer wondering where to begin, this guide will help you understand the architecture and run your first AI workflows in under 5 minutes.

🧭 What is this repository?

Proompts treats "Prompts as Code". Instead of copying and pasting text into ChatGPT, we store AI instructions as structured .prompt.yaml files. This allows us to version control them, test them with testData, and chain them together into complex Workflows.

🤔 Why do we do this?

  • Reproducibility: A prompt in YAML behaves the same way for everyone.
  • Testing: We can simulate logic without spending money on LLM API calls.
  • Automation: We can link prompts to build autonomous agents (e.g., passing a codebase to an Architect prompt, and its output to a Developer prompt).

🛠️ How to run your first simulation

1. Setup your environment

First, ensure you have Python 3 and install the required tooling:

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

2. Search for a Prompt

Let's find a prompt to test. Use the search script to find a "Code Reviewer":

python3 tools/scripts/search_prompts.py "code review"

3. Run a Workflow Simulation

We don't need an API key to test our logic! Our run_workflow.py engine simulates LLM responses using the testData embedded in our prompts.

Let's simulate the Agentic Coding Workflow (which chains multiple engineering prompts together):

# Run the workflow and provide a starting variable (-i)
python3 tools/scripts/run_workflow.py workflows/technical/agentic_coding.workflow.yaml -i product_concept="A new time-tracking app"

[!TIP] Add the -v (verbose) flag to see exactly how inputs map to the prompts at each step! 

python3 tools/scripts/run_workflow.py workflows/technical/agentic_coding.workflow.yaml -i product_concept="A new time-tracking app" -v

4. Verify your changes

Before pushing any new prompts or documentation, you must run the validation suite to ensure schemas and links are intact:

python3 tools/scripts/test_all.py

📚 What's Next?

Now that you understand the mechanics, check out these deeper guides: - Usage Guide: How to integrate these prompts with real LLM APIs in Python. - Workflow Guide: Learn how to build your own .workflow.yaml files. - Best Practices: The definitive guide to writing high-quality prompts.