Minimum Lab Deployment

Minimum Lab Deployment (MLD) is a way to learn a technology by deploying the smallest working example of its essential behavior. It is smaller in scope than a Minimum Viable Product: the result only needs to teach, not serve users or test a business idea.

Generative AI and coding agents accelerate implementation while the set of technologies a professional must understand keeps growing. MLD uses that division of work deliberately: the human defines the learning goal, then runs and validates the result; the agent builds and explains the lab. The learning loop is run → change → observe → understand.

The method is in MINIMUM_LAB_DEPLOYMENT.md. AGENTS.md gives Codex the standing instructions for a Lab. Copy both files into each new learning project.

Where the instructions belong

Location	Purpose
AGENTS.md	Short, persistent rules Codex should apply when creating or changing the Lab.
MINIMUM_LAB_DEPLOYMENT.md	The method and detailed requirements for the learning artifact.
Starter prompt	Only the chosen technology, learning goal, and any environment constraints for this Lab.

Codex loads project-level AGENTS.md automatically. Keeping it short avoids repeating the full method in every task and avoids imposing a long checklist on unrelated edits.

Start a new lab

1. Choose one technology and one fundamental behavior you want to understand.
2. Create a new Codex project and a new local directory, then point the project to that directory.
3. Copy AGENTS.md and MINIMUM_LAB_DEPLOYMENT.md into the directory.
4. Give Codex the prompt below, filling in the technology and learning goal.
5. When Codex finishes, run the documented exercise yourself. Change one input or setting, compare the observed result with the explanation, and use the generated mld/ guide to explore further. Ask for a deeper iteration only when you need one.

Each Lab produces a root-level mld/ directory containing eight guides on its chosen technology, named in sequence: 01-concepts-and-fundamentals.md, 02-tutorial.md, 03-use-cases.md, 04-essential-links.md, 05-github-repositories.md, 06-best-practices.md, 07-bad-practices.md, and 08-next-iterations.md. The concepts, tutorial, use cases, best practices, and bad practices guides explain the technology's general foundations beyond the Lab, using its deployment as a concrete example.

Starter prompt

I want to learn [TECHNOLOGY] by doing.

Learning goal: [ONE FUNDAMENTAL QUESTION OR BEHAVIOR]
Environment or constraints: [OPTIONAL; LOCAL BY DEFAULT]

Build and verify a Minimum Lab Deployment for this goal in this project.