What is agentic-engineering-mentor?

Agentic Engineering Mentor, also called Engineering Workspace OS, is a framework of files and instructions designed to turn an AI coding assistant into a long-term mentor for a software project rather than a tool you ask isolated questions. Instead of starting fresh in every conversation, it stores project knowledge in a permanent workspace folder that lives inside your own repository, so the AI can remember what has been built, what has been taught, and what comes next, even across different chat sessions. The core idea is a division of roles: you own the vision and direction of the project, while the AI owns the engineering process, including creating a learning roadmap, breaking work into milestones, running structured engineering sessions, and tracking progress over time. The project is meant to work with any kind of engineering project, not just web development. To use it, you copy a folder called .ai into the root of your own project, and optionally add any existing documentation you have, such as a README, requirements, diagrams, or notes, into a specific subfolder. If you have no documentation, the AI helps you create some and initializes the workspace from scratch. You then start a project by giving the AI a specific instruction that tells it to read the workspace files and follow the defined process instead of behaving generically. Each engineering session follows a clear pattern: the AI teaches only what is needed for that step, explains the objective, assigns one task with clear success criteria, and waits for you to complete it before reviewing your work and updating the workspace. The stated goal of the project is to help you learn by building, acting like a senior engineering mentor that guides you through planning, implementation, and review while gradually reducing how much you depend on the AI over time.