This is a template repo. Fork it, fill in your idea, and turn it into a personalized study guide.
A system that takes your project idea from rough brainstorm to a structured, phased study guide — generated by an AI assistant, tailored to your background, language choice, and constraints.
You do not have to know what you're building before you start. That's the point.
Click Fork on GitHub. This becomes your personal project roadmap repo — commit your study guide, track your progress, and build your Git history as a portfolio artifact.
Open BRAINSTORM.md and answer the questions. Do not overthink it — rough notes are fine. This file is for you.
- Open any capable AI assistant (Claude, GPT-4, Gemini, etc.)
- Paste the full contents of
doc-format-spec.mdas the system prompt (or first message) - Share your filled-in
BRAINSTORM.mdas context: "Here's my project idea:" followed by your notes - The agent will guide you through refining your idea, confirming your stack, and generating a personalized curriculum
As the agent generates phase files, save them into this repo (e.g., study-guide/phase-1.md). Commit as you go. Your Git history becomes the record of your learning.
- A phased study guide in Markdown, tailored to your project and background
- One Markdown file per phase, with sessions structured as: Reading → Build → Review → Exercises → Further Reading
- Quizzes (hidden answers) that target real misconceptions at each stage
- Three graded exercises per session: easy, medium, and challenge — all with working solutions
- A git commit suggestion at the end of every Build section
- A final phase covering production concerns: deployment, observability, security
See examples/sample-python-session.md for a complete example of what one session looks like.
The agent works best with these languages, but you can request others:
| Language | Best for |
|---|---|
| Python | AI/ML, data pipelines, rapid prototyping, scripting |
| TypeScript / Node.js | Full-stack web, APIs, frontend tooling |
| Java | Enterprise backends, microservices, Spring ecosystem |
| Go | High-throughput APIs, CLIs, systems tooling |
| Rust | Systems programming, performance-critical tools |
| Other | Tell the agent — it will work with you |
After forking, your repo will look like this as you work through the guide:
your-project-roadmap/
├── BRAINSTORM.md ← Fill this in first (your idea, background, constraints)
├── doc-format-spec.md ← The agent prompt — paste this into your AI assistant
├── study-guide/ ← Create this folder; save phase files here as you generate them
│ ├── phase-1.md
│ ├── phase-2.md
│ └── ...
├── examples/
│ └── sample-python-session.md ← Complete example: intake → curriculum → session
├── CONTRIBUTING.md
├── LICENSE
└── README.md
- One project, continuous evolution — the same codebase grows through every phase; you never rewrite from scratch
- Concepts before tools — understand why something exists before learning the library that implements it
- Sessions are units of mastery, not time — a session ends when the checklist is complete, not when the clock says so
- The quiz is a gate — every question targets a specific, common misconception so you know exactly what to revisit
See CONTRIBUTING.md for how to submit example study guides, add language stacks, or propose improvements to the session format.
MIT — see LICENSE.