Update the universal definition of done to include prompt engineer part as part of the standard duties #280
Description
As we increasingly leverage AI tools like GitHub Copilot, our effectiveness is directly tied to the quality of our "prompts"—the instructions we provide to the AI.
Currently, this knowledge is often ad hoc and siloed per developer. To improve development efficiency and ensure the entire team benefits from best practices, we propose establishing "prompt engineering" as a formal, standard, and trackable work item.
The objective is to treat our project-specific prompts as reusable assets, just like code or documentation. This involves creating and refining a shared library of prompts to improve development and train all team members on how to use AI tooling effectively within this project.
By "prompts," we mean an umbrella term that includes:
-
GitHub Copilot Custom Instructions: Guiding Copilot's behavior for our specific codebase.
-
Reusable Chat Prompts: Effective, copy-pasteable prompts for everyday tasks (debugging, refactoring, documentation).
-
Defined "Agents" or "Personas": A collection of roles we can ask the AI to adopt (e.g., "You are a senior QA engineer reviewing this code for test cases...").
-
A potential AGENTS.md: A file to formalize and share these personas.
We want to integrate this idea in the uDoD.