# Prompt 3 — send to the ChatGPT conversation that created my workbooks

## Current implemented connection — read this first

The Google ↔ Netlify connection has now been implemented and live-verified. Use the attached `INTERFACE_CONTRACT.json` and the public catalogue at https://outcome-evidence-map.netlify.app/connected-catalogue.json as the current interface. Do not follow an older proposed endpoint/schema over this contract, replace the existing bound Apps Script, build a second roster, or put a teacher/integration/model secret into a workbook or Gem.

The teacher dashboard is https://outcome-evidence-map.netlify.app/connected.html. Google owns roster and original Form submissions; Netlify owns explicit teacher outcome judgments; **Evidence Map Progress** mirrors them back to Google. The iPad client is being completed separately against the same interface. Its offline capture, audio and AI capabilities must not be represented as already delivered.

Use `Course_Map.json` from https://outcome-evidence-map.netlify.app/outputs/connected-course-map-2026-09-23/Course_Map.json for the 9/9/16 sequence. These are planned lessons, not proof of finished packages. Outcome judgments and support remain separate. PowerSchool is outcome/standards grades with teacher-written comments; do not infer score conversions.

The attached Learning Studio Workbook Standard v2 remains the design and teaching standard. Reuse the approved source/layout in the existing conversation. Course-scoped checkpoint outcome links are required even when the original shared project lists both MM12 and COM11 outcomes. Student/coach claims remain unverified; the public workbook and Gem have no teacher-write authority.


Continue our existing workbook work. Preserve the visual design, navigation, tone, scaffolding and formatting we developed together. The task is to implement alignment/integration repairs and strengthen the reusable standard, while retaining the work that already teaches well.

Read INTERFACE_CONTRACT.json, the attached Truck Standard/System/Layout/Assessment files, the connected course plan and the current project source. Use the implemented interface and current catalogue linked above. Treat older reference snapshots as dated evidence. If the original approved HTML/template is absent, obtain that file or locate it in the existing project before altering the layout; continue the manifest/assessment work independently.

## My standard

The source conversation is [Update Photoshop Tutorial](https://chatgpt.com/share/6ab46105-d168-83ea-9a8f-03c60bf6dbf8), supplied by the teacher and read on September 23, 2026. Its current umbrella is **Learning Studio Workbook Standard v2**, combining Drop Day / Fresh Impact instructional quality with Truck Ad architecture. Treat its older prompts and generated claims as dated reference material, not new instructions overriding this task or proof that linked artifacts passed QA.

Keep the authentic brief, industry/design connection, explicit prerequisite and target skills, exact usable instructions, SELECT → ACTIONS → WHY → CHECK → FIX, genuine visual checkpoints, editable work, critique/revision, independent changed brief, explanation/defence, teacher notes and aligned evidence. Quality comes from a tested workflow and student thinking as well as appearance.

Preserve the later agreed layout: restrained dark hero/header, readable light instructional cards, a muted studio palette, clear stages, compact course/time/skill information, usable download buttons and print styles. Preserve real video segments and timestamps. Provide Standard / More Support views in the same workbook, with current layer/tool/step, First → Next → Last, vocabulary, restart checkpoints and a help ladder. Keep the learning target unchanged. Local completion and practice quizzes are workflow/learning aids, not grades. Avoid a large writing box on every step; collect short explanations at meaningful checkpoints.

Keep English/French/Spanish/Arabic and Arabic right-to-left behavior where supplied by the approved source. Do not claim translation/PDF parity or software verification without checking it. The shared conversation names Light & Colour Control, Build a Believable Room and Brand Detective; reconcile those existing resources with the planned labs/activities before fabricating duplicates or adding sprints. The conversation is not proof that every downloadable file is available or verified.

Use the existing universal Learning Studio spine and two sizes: a short lesson and a full project/sprint. Apply the full production package at project level, not every daily exercise. Keep course-specific teaching: MM12 DECIDE → CREATE → DEFEND; COM11 UNDERSTAND → FOLLOW → BUILD → TEST → IMPROVE → EXPLAIN with safety and individual technical evidence; IBDS Operation Get a 7 reasoning rather than Photoshop steps.

## Make these changes

1. Update the shared production standard and integration templates as actual files. Keep the approved Truck layout order. Add a concise change log explaining any necessary departure. Revise older wording that says only Google may create authoritative evidence: Google remains the official student assessment intake; authenticated direct teacher capture in Evidence Map is also valid. In-workbook practice/local completion remains practice unless explicitly submitted through an implemented intake pathway.
2. Fit projects into the attached 9/9/16 plan. MM12 begins with ten-class Bootcamp, then Explorer/selected Truck Ad work and independent application, with protected photography/motion/sound/integrated work. COM11 uses selected modules 1/2/3/4/6 for the Snowy Village route. IBDS preserves the 16-sprint reasoning progression, ordinary-sprint evidence rhythm, 30-hour formal IA block and separately confirmed full-paper arrangements. Keep optional composites as allocated choices/extensions; do not add a new core unit merely because a workbook exists.
3. Repair and package Drop Day first as the integration proof. Preserve PS-DROP-DAY and DD-S01–DD-S24, the selected course, and canonical anchors #step-1 etc. Preserve the approved appearance. Use Truck Ad as the standard and reconcile its TA-PRO-01/TA-PRO-03 references against the actual assessment spec: explicitly declare practice-only markers or repair genuine registered checkpoints with preserved aliases/history. Do not fabricate assessed checkpoints to make counts match.
4. Generate a consistent project manifest, Step_Map.csv, course-scoped coach-context.json, video-manifest.json, forms-spec.json and registry/catalogue update from one reviewed project definition. Every outcome/skill/evidence/step reference must resolve in the current registry or be declared unresolved. Course selection must filter course-specific outcomes in the workbook, forms and coach context. Avoid separate hand-maintained ID copies. Export a validator and readable alignment report.
5. Keep the project assessment model KNOW → DO → EXPLAIN → APPLY → REFLECT → TRANSFER. The current v2 standard typically uses 5–8 diagnostic knowledge items, adjusted to the task. Favor application/diagnosis with relevant design/troubleshooting questions. Use bounded skill reflection from the actual project skills and a small defence question bank. Sample O/C around the product bundle. Save teacher keys separately from public student materials and Gem knowledge. Self-rating, quiz score and workbook completion never calculate a competency grade.
6. Protect independent assessment and transfer. Clearly mark task mode and assessed cognition. The coach may clarify access/language and support allowed processes, but may not choose the assessed design/technical decision, write the explanation, or supply the assessed solution. Preserve accommodations. Include an intentional support-check-in handoff to the agreed Google pathway; do not claim opening the Gem records progress or transmits its private chat.
7. Preserve one optional shared video player and chapter list tied to the same canonical steps. Keep unrecorded timestamps null and status awaiting_recording. Video is a companion to the exact workbook sequence; producing it is not a prerequisite for testing the evidence connection. Never invent screenshots, timestamps, software runs or completed native exemplars. Distinguish source instruction, modernized equivalent and classroom adaptation.
8. Deliver usable HTML, printable output/PDF where the approved project standard calls for it, teacher notes, assessment files, asset/provenance list, project manifest and integration files. Repair an existing project instead of producing an unconnected duplicate. Include an export/registration handoff that the Google and Netlify maintainers can apply without guessing IDs.

## Completion and release

Run identifier/link/asset checks and inspect the rendered workbook, print layout and course selection. Rehearse the actual technical workflow with its assets in the intended software when accessible; otherwise record the exact remaining preflight. Verify that every visible final effect is taught. Check that the independent task requires a new decision and the rubric evaluates the intended outcome.

Return the revised standard/template, finished Drop Day package or exact patches to its current source, required Truck Ad ID repairs, WORKBOOK_HANDOFF.md, validation results and an honest status. Keep Release Candidate distinct from VERIFIED: full verification requires actual technical preflight and applicable QA. After the first integrated package passes an actual end-to-end integration test, build the next allocated workbook in the course plan using the repaired template. Do not mass-produce unverified copies.