# Agent desk, source, print and preservation review

## Scope and source receipt

This is an agent author/desk review of ten optional Year 11 Queensland General English Unit 1 starts, not an independent human teaching trial. Current official English **2025 v1.3**, January 2026, resource 231399, © State of Queensland (QCAA) 2026 and completion cohort from 2026 were checked on 30 September 2026 using the full relevant primary indexed sections. Exact Unit 1 objective wording is at printed 18; Unit 1 description/subject matter at 17, 19–20; course/text/assessment/reporting limits at 5–10 and 14–16. Direct PDF request returned HTTP 403: current remote bytes/hash/layout remain unverified. [Source register](source-register.json), [source rights](SOURCE-AND-RIGHTS.md) and [crosswalk](CURRICULUM-CROSSWALK.md) preserve this distinction. Senior work uses QCAA objectives, not F–10 ACARA codes.

A complete 1,148-word original fictional story, first-person companion and review support substantive literary reading/composition. A single supplied story is not the required complete short-story collection, Australian/First Nations authored text, category mix, 55-hour unit or full assessment programme. Local teachers can run the originals immediately and use the fully prepared lawful text-substitution route. No actual creator biography, school event, student response or cultural permission is invented.

## Independent publishing corrections · 1 October 2026

A separate author-agent read the complete sources, all ten plans and learner
routes, fresh A/B, rubric/key and current primary Unit 1 guidance. The private
`year-11-english-w07-08-independent-review-2026-10-01.md` records two findings
against the unchanged author freeze. Both are corrected in this publishing copy:

- **C02 chronology:** Lee now writes Mara's request beside the existing arrow,
  matching S04/S06. The third-person story, held sources and other contrast
  details are unchanged. The contrast data, prose, SVG/PDF and exact alternatives
  agree; its specific source pin is deliberately renewed.
- **Day 31 mode:** an actually composed oral/AAC opening is composed text. Only
  a proposed future opening is a plan. This does not supply written attainment,
  handwriting or an unobserved performance. The editable data and full script
  give the same corrected instruction.

The publishing agent genuinely regenerated all 17 print/alternative outputs.
Only the contrast SVG/PDF and the combined exact alternatives differ; the other
14 output bytes, including both held pairs, remain identical. The actual revised
contrast PDF was rendered and viewed in full: all C01–C03/R01 wording fits with
no clipping/collision. The original eight-page review remains dated 30 September;
`render-review.json` binds the seven unchanged pages and the explicit 1 October
contrast follow-up separately. No classroom, listening or human trial follows.

## Ten-day role walkthrough

All six teaching actions, three access routes, two worked contexts and home/extension per day were read. The following is reasoning about use and burden, not observed pacing or attainment.

| Day | Educator/use and learner-choice check | Particular limit/next move |
| --- | --- | --- |
|31|One opening begins in five minutes; complete S01–S12 is available across three actual source pages. A phrase/initial reading can be useful now.|Whole-story reading may exceed eight minutes for some students. Access can span the two sittings; read surrounding context and return to the whole text before whole-story claims.|
|32|Thought/speech/narrator words are differentiated; a prepared two-sentence transformation illustrates detail selection, not just pronouns.|New invented knowledge is labelled creative authorship; alternate coherent narrator choices remain valid.|
|33|Clock/heading/chair/window recurrence allows two source phrases and an own motif line; a worked competing interpretation keeps meaning open.|A device label alone is narrower than explaining its work. No old photograph date/meaning is invented.|
|34|A prepared analytical paragraph connects evidence and thesis with an alternative reading. Students choose their own argument.|250–350 words is a flexible eventual guide; one developed component can continue later. Literary analysis is not reduced to fact checking.|
|35|New complete A narrative/note is released only for the intended first response. Neutral access, own reading and supported counter-reading are possible.|Public sources cannot guarantee novelty. Log actual previous exposure/support; avoid model answer before capture. Partial responses have descriptive next teaching.|
|36|Two prepared opening proposals illustrate voice; an own third approach and imagined/private purpose are available.|350–500 words is an eventual guide, not a fifteen-minute minimum. No identity disclosure, performed scene or borrowed community voice is required.|
|37|Real direct/reported-speech alternatives and sequence choices give a practical craft trial.|Only actual existing versions support before/after comparison. Student direction and adult wording/acting are distinguished.|
|38|Integrated quotation/reference-word models make syntax/cohesion editing runnable. An own critical introduction relates source and two craft choices.|No earlier draft is assumed. Current composition can begin now; handwriting/accent is not a proxy for analysis.|
|39|An edition can start with an annotated excerpt and own passage. Prepared credit/reading-order choices work on paper or accessible text.|No completed portfolio, public upload, device, decoration or willing-reader response is necessary. Self-review is sufficient; feedback is recorded only if actual.|
|40|Genuinely new B springboard allows different narrator/time/ending/mode with commentary. Missing earlier work is acknowledged.|No compulsory reveal, agreement, title or happy ending. First response precedes feedback; a script is not actual performance. No whole-unit/year grade follows.|

## Four-role critique and refinements

- **Educator desk lens:** The full originals, locators, prepared grammar/focalisation/thesis models and separate open key avoid requiring a new teacher to invent the demonstration. Two 25-minute sittings and one sampled private note control burden. During review, concrete ordinary modelling language was added for Days 31–34 and 36–39. The original and fresh source facts were unchanged. The school still checks its lawful text programme, required breadth, assessment/moderation and local access; the pack is a bounded teaching module.
- **Parent/home desk lens:** Each invitation is free, optional and independent of a family reader, personal history, purchase, account or returned evidence. No participation is treated as an error. The teenager may use private fiction/self-study or decline. Fresh sources stay out of pre-task homework.
- **Learner desk lens:** The first useful output is an own opening, then an interpretation or passage rather than a completed administrative form. Paper/typing/sign/speech/AAC/dictation/director options are concrete and switchable. AAC is expanded in the teacher/family entry points. One chosen component and a later sitting are useful; no forced speech/performance/peer approval. Interpretation and deliberate craft remain senior English goals across modes; adult-supplied language has narrower authorship evidence.
- **Funder desk lens:** Ten authored plans and eight original pairs provide usable resources and conditional curriculum opportunities, not measured delivery, student attainment, Australian/First Nations text coverage, full Unit 1/year/QCE or all-state certification. No practising teacher, student, family, funder or cultural authority took part in this agent desk review. A computer-assisted example is labelled and cannot substitute for an actual human creator/account.

## Actual local PDF inspection and exact alternatives

Existing installed Poppler rendered **all eight stored local PDFs** at 1,800 pixels on the longer edge. Each full page was individually viewed with the local image tool, plus a contact sheet. [Bound receipt](render-review.json) pins the exact viewed local PDF bytes. Three story pages retain all S01–S12 words; contrast/review retains C01–C03/R01; two blank practical aids have readable broad diagrams/panels; both fresh pages show their complete new texts without supplied results. No blocking clipping, overlap or hidden-picture invention was found. Source fonts are about 10.9 pt; larger/plain text and neutral read-aloud remain supplied options. [Exact alternatives](print/TEXT-ALTERNATIVES.md) include all visible wording/order, source descriptions, tactile and no-print instructions. Local PDFs are searchable but untagged; human screen-reader/tactile, actual printing and learner access tests remain open. The official remote PDF was not rendered.

## Read-only reproduction, faults and preservation

From this leaf:

```sh
python3 verify_pack.py
python3 verify_pack.py --negative-tests
```

The default reads source identities/eleven objective wordings/locators, complete original and held facts, all ten concrete scripts and 50-minute splits, thirty routes, twenty examples, ten home/ten extensions, openness/first-response/key/authorship/course limits, actual A4 geometry, searchable PDF text/order, exact alternatives, rights, relative links, viewed-byte receipt and manifest. It genuinely runs the original SVG/render/metadata recipe in a temporary directory and compares **all sixteen generated SVG/PDF bytes plus the exact-alternative file**. It does not copy stored PDFs as generated results. Python bytecode is disabled for checks; authored file SHA/size/mtime are compared before/after.

The negative suite mutates temporary copies, refreshes their manifests deliberately and must still reject substantive source, current-version/cohort/copyright, formal-result, lesson/pacing, fresh-answer/first-response, actual-version/authorship, home/privacy/substitution, course/text breadth, renderer/font rights, source leakage, local link, geometry, exact-alternative and render-scope faults. Harmless temporary runtime caches are controls, not authored resources; an unlisted authored file must fail. Final executed result: **37 refreshed-manifest faults rejected, three harmless cache controls passed, and one unlisted-file control rejected**. Default generation matched all 17 outputs; actual PDF words/order are read through existing Poppler (rather than treating a PDF-library kerning-space artefact as missing words). Black and Ruff pass for both authored Python files. No stored PDF/source change was needed for this extractor choice.

The [owned-pack operation receipt](preservation-receipt.json) reports a completed read-only validation of this pack: 39 authored files, unchanged SHA/size/mtime and 17 genuinely reproduced outputs. This operation receipt is not learner attainment or classroom evidence and gives no future preservation guarantee. Author workspace snapshots are retained privately in the repository review archive; unrelated paths, raw file digests and concurrent-work records are excluded from this public teaching receipt.

**Open human/local review:** Queensland senior English educator and school course/assessment moderation; current cohort/programme/text-source rights and appropriate cultural authority for any chosen alternative; student-paced reading/composition, accessible modes, private storage and actual reader responses if used. Desk/render/reproduction passes do not prove classroom timing, practical uptake or achievement.

Original review and guidance ©2026 NeuroForgeIO Pty Ltd, [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/).

## Public receipt scope correction · 1 October 2026

Root retained the complete author snapshot privately and replaced the public receipt with an owned-pack-only record. The matching verifier checks that exact public schema and rejects extra workspace provenance. Complete original stories, held sources/tasks, separate descriptive key and all stored print/alternative bytes remain unchanged. The earlier source/desk/render/negative receipts describe their executed versions; the final public-source correction receives its own read-only validation before publication. No student data, human moderation, classroom outcome or public-site deployment is inferred.

Executed public-receipt validation: 35 local links, 17 genuine temporary byte-reproduced outputs, 38 refreshed-manifest substantive faults rejected, three cache controls passed and the unlisted authored control rejected; authored SHA/size/mtime unchanged. The added private-provenance injection fault is rejected by the public receipt schema.