Skip to content

About

Think in documents, review with a human, then build — a starter kit and interactive dashboard for students beginning projects with AI assistance.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

6 Commits

Folders and files

Repository files navigation

Project Startup Guide

Think in documents. Review with a human. Then build.

https://ethical-tech-colab.github.io/project-startup-guide/ — the whole guide as a clickable walkthrough with progress tracking, copy-ready prompts, and downloadable templates. Start there; this repository is the source behind it.


A complete starter kit for students and non-engineers beginning a project with AI assistance: four Markdown documents to write before any code, a human review gate, a recommended free toolchain, MCP server setup, and an interactive dashboard you can deploy to GitHub Pages.

Chat first. This kit assumes you work in VS Code Copilot Chat, Agent mode — roughly 95% of the work, including installing Git, committing, pushing, opening pull requests and deploying, is done by asking rather than typing terminal commands. Terminal equivalents are included as a fallback, not as the default path.


The method in one line

   💡                 📐                  🗺️                  ✅                  🔨
CONCEPT-IDEA  →  SPECIFICATION  →  PLAN + BACKLOG  →  human review  →  build
  (why/what)      (what exactly)    (how / in what        (gate)        (the easy part)
                                        order)
Document Answers Gate before moving on
CONCEPT-IDEA.md Why should this exist, and for whom? A human confirms the problem is real and the scope is narrow
SPECIFICATION.md What exactly does it do? A human confirms it is testable and complete
PLAN.md How is it built, in what order? A human confirms milestones are demonstrable and realistic
BACKLOG.md What is the very next task? A human confirms priorities and sizes

Documents are drafted with an AI assistant, read and rewritten by you, and then reviewed by a person against REVIEW-CHECKLIST.md before anything moves forward.


What's in here

project-startup-guide/
├─ docs/
│  ├─ index.html                    the interactive dashboard (GitHub Pages)
│  └─ .nojekyll
├─ guide/
│  ├─ 00-START-HERE.md              read this first
│  ├─ 01-DOCUMENT-WORKFLOW.md       the four documents and how they chain together
│  ├─ 02-TOOLING-SETUP.md           GitHub, Git, VS Code, Copilot, Pages, Claude/OpenAI
│  ├─ 03-MCP-SERVERS.md             MCP explained, the starter set, safety rules
│  ├─ 04-HUMAN-REVIEW.md            how to run the review gate
│  └─ 05-GLOSSARY.md                every term, in plain language
├─ templates/                       the four documents + the reviewer's checklist
├─ prompts/PROMPT-PACK.md           eight prompts covering the whole workflow
├─ examples/                        mcp.json, copilot-instructions.md, CLAUDE.md, .gitignore
├─ tools/build-templates.mjs        embeds templates into the dashboard
└─ .github/workflows/deploy-pages.yml

Quick start

  1. Read guide/00-START-HERE.md.
  2. Work through guide/02-TOOLING-SETUP.md — about 90 minutes, free.
  3. Add the MCP servers from guide/03-MCP-SERVERS.md — start with GitHub and Tavily.
  4. Copy the four templates into your own project's repository root.
  5. Run the prompts in prompts/PROMPT-PACK.md, in order, one review between each.

Or — recommended — just open the dashboard and let it walk you through it step by step.


Publishing the dashboard

The dashboard is a single self-contained HTML file — no build step, no dependencies. This repository is already published at https://ethical-tech-colab.github.io/project-startup-guide/ via the included workflow. To publish your own fork:

Option A — GitHub Actions (what this repo uses)

Settings → Pages → Source: GitHub Actions. The workflow publishes docs/ on every push to main.

Option B — deploy from a branch (simplest)

Settings → Pages → Source: Deploy from a branch, branch main, folder /docs. Live within a minute at https://<username>.github.io/<repository>/.

Preview it locally

npx serve docs
# then open the printed http://localhost:3000 address

Opening docs/index.html directly from the file system also works; copy-to-clipboard falls back to a legacy method outside a secure context.


Editing the content

The dashboard embeds the templates so a visitor can copy or download them without leaving the page. Those copies are generated — edit the files in templates/, prompts/ and examples/, then re-embed:

node tools/build-templates.mjs

The script injects each file into the matching <script type="text/markdown" data-src="…"> block in docs/index.html. Running it twice is safe.


What the dashboard covers

Tab Contents
Start here The method, the five phases, a 10-minute version
1 · Tools GitHub, GitHub Education, Git, VS Code, Copilot, GitHub Pages, Claude Code / Codex
2 · MCP servers What MCP is, where config lives, the starter set, safety rules, troubleshooting
3 · The four documents What goes in each, the prompt to draft it, when it is done
4 · Human review Who reviews, how, the checklist, verdicts
5 · Build & ship Walking skeleton, the build loop, weekly rhythm, verification, demo
Templates & prompts Copy or download every file in this kit
Glossary ~70 searchable terms for non-engineers

Progress ticks are stored in the visitor's browser (localStorage) — nothing is uploaded.


A note on accuracy

Pricing, free tiers, student benefits and MCP server addresses change over time. Everything here was checked when written, but verify anything that matters against the vendor's own documentation — and teach your AI assistant to do the same, with sources.

Licence

Use it, fork it, adapt it for your class or cohort.

About

Think in documents, review with a human, then build — a starter kit and interactive dashboard for students beginning projects with AI assistance.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages