AI-enhanced Educational Game Design

A twelve-session microcredential handbook —
objectives, playtest, evidence, credential.

What it is · In the handbook · Technical niches · Architecture · Quick start · Design notes

---

What it is

A self-contained, static handbook for The University of Alabama's AI-enhanced Educational Game Design microcredential (v2). Twelve weekly sessions, each pairing one learning objective with one designed artifact. Five non-compensatory deliverables, 25 rubric criteria, one Proficient floor.

The repository is also a reference implementation of a credential stack — Open Badges 3.0 / W3C Verifiable Credentials 2.0 / CLR 2.0 / xAPI — so the same scaffolding can be lifted into other competency-based programs without re-deriving the spec choices.

---

In the handbook

The front of the house. Crimson-branded hero, 12-session grid, and a Resource Library entry — one handout per session.	The left rail. Four parts, twelve sessions, live progress counter (hb:done in localStorage).
The companion library. Twelve facilitator handouts, indexed and filtered by course placement.	Embedded minigames. Each session ships an interactive check-for-understanding — answers lock on first click, no partial credit.

---

Technical niches

Specifications chosen on purpose. Each one does work the next one cannot.

Open Badges 3.0 + VC 2.0 Assertions conform to OpenBadgeCredential on a VerifiableCredential base. Data Integrity Proof placeholders are ready for eddsa-rdfc-2022 / Ed25519 signing; issuer binds to a did:web path so proofs verify without contacting us. See credential/badge-class-v3.json, assertion-example-v3.json.	ESCO + Lightcast crosswalk All 25 rubric criteria are mapped to ESCO v1.1.1 and Lightcast Open Skills URIs. The badge's alignment[] block reads machine-verifiably against the same taxonomies state workforce agencies and LER-RS pipelines consume. See credential/skills-crosswalk.json, alignment.html.	CLR 2.0 portfolio export Whole-portfolio transcripts as a ClrCredential, with per-deliverable achievement blocks and an embedded _xapi_statements array so cohort context travels with the artifact. Cohort or per-learner export from analytics.html. See clr.js.
xAPI 1.0.3 analytics Cohort heatmap, completion funnel, Cohen's κ inter-rater reliability, and a pre/post eight-skill self-assessment stream with growth deltas — rendered against the same taxonomy the badge alignment uses. See xapi.js, analytics.html.	Non-compensatory rubric One Developing on any criterion blocks the deliverable. scoreGiven is binary (0 or 1); no partial credit buys over a floor. Double-rating on ≥ 20% of submissions feeds κ. See rubrics.html, docs/L3-evaluation-plan.md.	Endorsement credentials Third-party endorsers (external faculty, industry partners) sign their own EndorsementCredential against our BadgeClass — per-cohort or per-learner. The handbook stays the source of truth; the endorser contributes an orthogonal signal. See endorsement-template-v3.json.

Design notes shipped alongside, not code. LTI 1.3 production wiring, Credential Engine / CTDL registry submission, AI-assisted skill tagging (prompt template v0.3 with negative-space signal), and a Kirkpatrick × CIPP program-evaluation framework — see docs/.

---

Architecture

Deliberately boring stack. Static HTML / CSS / vanilla JS, no build step for the shell, no server component, no user analytics. Hosted on Cloudflare Pages at teachplay.dev. Two interactive labs are self-contained SCORM packages under minigames/.

Learner browser                                        Issuer (durable URL)
────────────────                                       ─────────────────────
index.html → shell.js → localStorage                   teachplay.dev/credential/
   │                                                     ├── issuer-v3.json
   ├── xapi.js ──────── xAPI 1.0.3 statements          ├── badge-class-v3.json
   │                         (local queue)             ├── assertion-example-v3.json
   ├── analytics.html ── κ · funnel · heatmap          ├── endorsement-template-v3.json
   │                     · skills growth chart         └── skills-crosswalk.json
   ├── clr.js ───────── CLR 2.0 portfolio export                │
   └── role.js ──────── student / instructor surface           ▼
                                                        ESCO + Lightcast
                                                        (taxonomy refs)

---

Quick start

# clone
git clone https://github.com/Educatian/TeachPlay.git
cd TeachPlay

# serve (any static server; python is zero-dep)
python -m http.server 8099

Open http://localhost:8099/. The preview launch config at .claude/launch.json starts the same server for in-editor preview.

To deploy your own instance: push to a fork, connect the fork to Cloudflare Pages (or any static host), and rewrite the canonical URL references in credential/*.json and the xAPI extension IRIs to your domain.

---

Repository map

index.html              Landing: hero, sessions grid, deliverables
session-01.html …       12 session pages — one LO, one artifact each
  session-12.html
rubrics.html            25-criterion rubric, non-compensatory, AI-tagging stub
alignment.html          Crosswalk: objectives ↔ deliverables ↔ skill chips
credential.html         Public credential page (OBv3, stacking, endorsement)
analytics.html          Instructor analytics — κ, funnel, skills growth
facilitator.html        Facilitator-only views
resources.html          Resource library + #labs section
read.html               Markdown viewer for companion handouts

handbook.css            Shared styles (inc. @media print)
shell.js                Sidebar, checklists, tickets, self-assessment wiring
xapi.js                 xAPI 1.0.3 emitter
clr.js                  CLR 2.0 export adapter
role.js                 Student ⇄ Instructor role toggle
minigames.js            Inline minigame injector

credential/             Open Badges + VC + endorsement + crosswalk (JSON-LD)
docs/                   Design notes — CTDL, AI tagging, LTI, evaluation
resources/              12 markdown handouts (companion library)
minigames/              SCORM labs — orbit-sum-lab, electric-circuit-lab
screenshots/            Handbook screenshots (used in this README)
og-image.svg / .png     Social preview card (1200 × 630)

---

Credentials in detail

What the badge means and how it's shaped

• Achievement type: Badge with creditsAvailable: 3
• Alignment block: ISTE · UDL · NETP + ESCO + Lightcast — one per criterion (25 total)
• Evidence: credentialSubject.result[] carries one block per deliverable D1–D5 with rubric rollup
• Proof placeholder: DataIntegrityProof with cryptosuite: eddsa-rdfc-2022, proofPurpose assertionMethod
• Dual publishing: OBv2 (credential/badge-class.json) and OBv3 (credential/badge-class-v3.json) live side-by-side until consumers migrate

Analytics that instructors actually use

• Cohort heatmap — which rubric criteria get Developing calls most often
• Completion funnel — entered / attempted / completed / abandoned per session
• Cohen's κ — inter-rater reliability on double-rated deliverables, with calibration trigger at κ < 0.70
• Skills growth — pre/post 8-skill Likert delta per cohort, rendered as bars with Δ
• CLR export — whole-cohort or per-learner, as ClrCredential JSON

Persistence and privacy

All learner progress is stored in localStorage on the visitor's browser:

• hb:s{n}:chk:{id} — checklist item state per session
• hb:s{n}:ticket:{id} — exit-ticket textarea content
• hb:done — set of sessions marked complete
• hb:self-assess:{pre|post}:{skill} — 0–4 Likert self-ratings
• hb:xapi:queue — xAPI statement queue
• hb:xapi:actor — pseudonymous actor (UUID, browser-local)

No server component, no third-party analytics, no identifying data leaves the browser in the reference implementation. Clearing site data resets progress.

---

Design notes

Under-building on purpose; each note names the decisions before any code ships.

Note	What it covers
docs/L1-ai-skill-tagging-design.md	Prompt template (v0.3) for AI-assisted skill extraction. Evidence-linked tags, negative-space "claimed but not shown" signal, guardrails against over-reliance.
docs/L1-credential-engine-registration.md	CTDL resources to publish (ceterms:MicroCredential, ceasn:CompetencyFramework, …), field-by-field mapping, submission workflow.
docs/L2-lti-1.3-design.md	LTI 1.3 integration plan — OIDC launch, Assignment and Grade Services (AGS), Names and Role Provisioning (NRPS), role claim enforcement, estimated effort.
docs/L3-evaluation-plan.md	Program evaluation framework — Kirkpatrick × CIPP, EQ1–EQ5 (skill acquisition, inter-rater reliability, equity subgroups, transfer, external validity), IRB posture.

---

Session ↔ Lab pairings

Session	Lab	Role
S3 · Objectives and crosswalk	Orbit Sum Lab (SCORM, React/Vite)	Worked crosswalk case
S8 · Interaction spec	Electric Circuit Lab (SCORM, Three.js)	Reference implementation
S10 · Audit	both labs	Audit subjects — accessibility, QA, data
resources.html#labs	both labs	Library entry

---

Adding content

Add a new handout

1. Drop resources/NN-slug.md into the resources/ folder.
2. Add a .resource card to the relevant session's Companion reading section with:
   • href="read.html?doc=NN-slug.md" on the Read button
   • href="resources/NN-slug.md" download on the Download button
3. Add a matching entry to the resources.html library grid.

Add a new lab

1. Unpack the SCORM package into minigames/<slug>/.
2. Pick the session whose learning objective the lab exemplifies.
3. Add a .resource card inside that session's #reading block — use the A (green) or B (blue) badge convention established in S3/S8/S10.
4. Also add a card to resources.html#labs so the library stays complete.

---

Print / PDF

Each session page has a Print / PDF button. The print stylesheet at handbook.css:1886 hides sidebar / toolbar / nav, expands content to full width, and avoids page breaks inside blocks.

---

License and acknowledgments

The University of Alabama, College of Education. Non-compensatory rubric design informed by the competency-based assessment literature. Credential specifications draw on the 1EdTech Open Badges 3.0 / CLR 2.0 working groups, W3C Verifiable Credentials 2.0, Credential Engine's CTDL vocabulary, and the Digital Credentials Consortium's guidance on decentralized issuer binding.

_{Built slowly, on purpose.}