Add Wish product intent folder

AGENTS.md

@@ -0,0 +1,7 @@
+# Agent entrypoint
+
+This repository is a public teaching Space. Before changing its learner-facing purpose, copy, or flow, read [docs/wish/LATEST.md](docs/wish/LATEST.md). For navigation, page roles, or a major interaction, also read [docs/wish/DESIGN.md](docs/wish/DESIGN.md).
+
+The Wish records product intent. The current code, Space configuration in `README.md`, and verified Hugging Face behavior own implementation facts. Do not treat [docs/wish/ARCHIVE.md](docs/wish/ARCHIVE.md) as current direction; read it only when tracing an older decision.
+
+The shared Wish lifecycle is maintained in [WISH_PROTOCOL.md](https://github.com/mykcs/.codex/blob/main/website-governance/WISH_PROTOCOL.md). Keep this file as a short route to the local Wish rather than a second copy of it.

README.md

@@ -44,6 +44,8 @@ The source code and the live demo are the same project:
 - `detector.js` — the AI model call
 - `examples/` — tested COCO validation images
 
+The maintainer's current product intent is in [docs/wish/LATEST.md](docs/wish/LATEST.md), with the Space-specific design interpretation in [docs/wish/DESIGN.md](docs/wish/DESIGN.md). The [Wish folder](docs/wish/README.md) explains the other files.
+
 ## Student assignment
 
 Students can duplicate this Space as a working starting point, then build their own AI mini app. They may extend object detection or adapt the project to a different AI task. A submission should make the app's purpose, input, output, and AI model or service easy to understand and demonstrate at least one working example.

docs/wish/ARCHIVE.md

@@ -0,0 +1,3 @@
+# Wish Archive
+
+No earlier center Wish has been superseded. When the central product intent changes, record the old Wish here with its active dates and the reason it was replaced before updating `LATEST.md`.

docs/wish/DESIGN.md

@@ -0,0 +1,31 @@
+# How the Wish Shapes This Space
+
+This document interprets the [current Wish](LATEST.md) for this Space. Code and the Space configuration remain the source of truth for runtime behavior.
+
+## L0 — Center
+
+The Space lowers the effort of publishing a first AI mini app and the effort of reviewing it. A visitor should see a runnable example; a student should see a path to make a personal copy; an instructor should see what a submitted app does without searching through files.
+
+## L1 — Real situations
+
+1. **A student has computing basics but has not used Hugging Face.** They need visible account, duplicate, edit, and submit steps rather than an assumption that they know Git or Space settings.
+2. **A student wants a different AI idea.** The starter should show a possible shape for an AI app, while allowing a different model, task, input, or output.
+3. **An instructor opens many submissions.** Each public Space should explain its purpose quickly, offer an obvious interaction, and state the student's contribution.
+4. **A visitor tries an uploaded image.** They need to understand what the browser does with their input and see a clear result or failure.
+
+## L2 — Product response
+
+- Keep a working example with an immediate way to try it, including known sample inputs.
+- Keep the browser guide and PDF aligned. Use real platform screenshots where a button or menu is hard to locate.
+- Let students copy the whole Space into their own account and then choose whether to extend object detection or make another AI mini app.
+- Make the finished app explain its input, output, model or service, the student's change, and at least one test or limitation.
+- Describe upload and model behavior accurately. Never ask students to put a secret in public source files.
+
+## L3 — Page and file roles
+
+- **App (`index.html`)** is the live example and the first explanation of what it does. It links to How to Start.
+- **How to Start (`how-to-start.html`)** is the editable, step-by-step browser guide. **`how-to-start.pdf`** is its shareable counterpart with the same instructional content.
+- **`README.md`** explains the starter, Space configuration, and assignment context.
+- **`docs/wish/`** tells maintainers why these pieces exist and how to judge later changes. It is not the student's setup guide.
+
+When a design choice is unclear, prefer the option that makes a student's next action or an instructor's first review easier to understand.

docs/wish/LATEST.md

@@ -0,0 +1,13 @@
+# Current Wish
+
+Updated: 2026-09-26
+
+> I want this Space to give graduate students who are new to Hugging Face a clear path from an AI idea to a public, working mini app under their own account. When they submit the link, an instructor should be able to understand and try their work within a few moments.
+
+Keep three things central:
+
+- **A working starting point.** Students can duplicate the object-detection demo, then improve it or adapt it to another AI task. Training a model from scratch is not required.
+- **A beginner-readable path.** The editable How to Start page and matching PDF should show the actual platform steps, including sign-up, duplication, editing, testing, and submitting a Space link.
+- **A legible submission.** Each app should make its purpose, input, output, model or service, and the student's own changes easy to find. A real result and an honest limitation matter more than decoration or code volume.
+
+This Space should stay a reusable teaching launch point, not become a single prescribed AI assignment or a gallery that hides what each student actually built.

docs/wish/README.md

@@ -0,0 +1,9 @@
+# Course Space Wish
+
+This folder records the current product intent for this teaching Space.
+
+- [LATEST.md](LATEST.md) — the current Wish and priorities.
+- [DESIGN.md](DESIGN.md) — how the Wish shapes the learner and instructor experience.
+- [ARCHIVE.md](ARCHIVE.md) — earlier Wishes when the central intent changes.
+
+The shared update and archiving rules live in [WISH_PROTOCOL.md](https://github.com/mykcs/.codex/blob/main/website-governance/WISH_PROTOCOL.md). This folder owns this Space's intent; it does not duplicate the cross-project protocol.