# Teacher, child, assessor and funder desk run-through · Weeks 19–20

**Teacher desk run:** Day 94 totals 25 minutes (3 source, 4 prediction, 5 model, 7 child record, 4 check, 2 exit). The actual apparatus must be school-approved and adult-pretested; four safe tries may be impractical in seven minutes, so prepare/rehearse the adult operation beforehand and never rush it. If unsafe, noisy, inaccessible or interrupted, use M and write **REAL OBSERVATION NOT AVAILABLE**. If only one valid actual attempt occurs, keep it but say comparison insufficient. The integrated Week 19 ring test is a separate plan; do not count its fictional record as this pack's run. Day 99 offers a private teacher audience.

**Child desk run:** M says broad stops START and MIDDLE, narrow stops MIDDLE and MIDDLE, plus one tipped rim entry labelled INTERRUPTED. I can say what that *invented* list contains but not that a real carton behaves that way. I can ask which conditions stayed the same and speak, sign, draw, use AAC, point, direct an adult, receive an attributed ADULT REPORT, or use a secured broad tactile strip. I do not have to touch, hear or watch the carton. I can say “I don't know yet.”

**Assessor desk run:** Held R's HOME/BRIDGE/GATE fabric-covered block and held S's MARKET DISPLAY PANEL MODEL/A-B crossed entries are absent from teaching M/N, learner cards, swaps and print aids before first response. R tests confounding start lines; S tests mixed valid stops and a caption added after results. First answer and support are saved before key. Paper cases cannot credit an actual AC9SFI02 investigation or actual movement-factor result. A prediction recorded after results cannot be compared as prior. One check cannot establish Foundation year/term mastery or improvement.

**Funder/public desk run:** Seven exact pinned ACARA rows and a QCAA Prep entry, original CC BY teaching/printables, source and rights gates, six access-equivalent A4 aids, no-cost family invitations and hash receipt are visible. Educator, disability/accessibility, local curriculum, apparatus/safety, privacy and classroom review remain due. There is no field pilot, measured effectiveness, endorsed syllabus, translation or cultural authority claim.

**Mechanical/visual receipt:** [Verifier](verify_pack.py) checks ten six-step 25-minute scripts, 30 same-target routes, 20 worked swaps, ten optional family prompts, two separate fresh checks/keys, seven exact official rows/workbook hash, local links, six A4 SVG/PDF pairs and manifest hashes. Rendered PDFs receive a human visual desk check. This is a desk run-through, not a classroom trial.