# Explore the product graph · chapter 2 · verification receipt · 10 September 2026

Scope: the second chapter of the _Use Eve_ track, rewritten in the Eve edition —
two manuscripts merged into one, sixteen slides re-treated and re-voiced, six
chrome slides added, and every binding re-reviewed. Chapter 1 also gained the
section divider it was waiting on. The [brief](../../authoring/eve-graph-chapter.md)
records what the chapter is for and why each treatment was chosen;
[state.json](state.json) holds the exact counts, hashes and tool results.

## What was checked

- **Build freshness.** `build-eve-oshun.py --check` passes on the delivered
  tree: 1,366 unique slides, 658 sources, 75 guides. The chapter is 22 slides
  and 30.7 minutes of narration.
- **Teaching crosswalk.** `check-eve-capability-teaching.py` passes across 105
  slides: every slide has at least three review questions and a hundred words of
  explanation, the guide's slide order matches its manuscript, and both
  product-graph assignments carry the rewritten titles and takeaways.
- **Reviewed assignments.** The two units in `assignments/eve-product-graph.json`
  were re-read against the rewritten planes slide rather than re-pointed. It
  still identifies the compiled artifact as knowledge authority, still calls the
  database knowledge stores rebuildable projections, still excludes an
  assistant-turn write from authoritative mutation, and still traces an accepted
  change through source, regeneration and independent checking. Each carries a
  dated re-review note and the new slide fingerprint.
- **Narration.** All 22 tracks rendered and verified: `--check --ids` over the
  whole chapter exits 0, which checks the fingerprint, byte count, SHA-256,
  probed duration and a full decode against the authored pause plan. The other
  1,344 tracks are untouched.
- **Browser rendering.** `eve-edition-review.mjs --guide eve-product-graph`
  sweeps all 22 slides at 1440×900, 1280×720 and 390×844 and in print with no
  findings: no text outside the viewport, no page overflow, no narration chrome
  pushed off screen, and every authored string painted rather than merely
  present in the document. Chapter 1's ten slides sweep clean too.
- **PDFs.** Both chapters exported and checked. The export asserts no print-page
  overflow and runs `check-pdf.py` on each file; `check-eve-specimen-pdf.py` then
  holds every page to the exact-text contract — every authored visual field on
  its own page, in order, with headers, takeaways and captions excluded from the
  body boundary. 22 pages and 10 pages pass.
- **Suites and lint.** Python 445 passed with one failure that is present on the
  unchanged baseline; Node 443 of 445 with two failures caused by an incomplete
  `node_modules` in this worktree; Ruff, ESLint and Prettier clean on everything
  changed.

## What was not checked

- No human reader session and no listening review at normal speed.
- The narration was rendered on this host's CPU rather than the Apple GPU the
  rest of the library used. A control re-render of an untouched slide produced an
  identical duration to the millisecond and a 0.99 envelope correlation over
  20 ms frames, so the model output matches; only the MP3 bytes differ.
- No new product capture was taken. A capture harness re-run on this host's
  Chromium produces different bytes from the committed assets, which would break
  the slide pins, so the chapter reuses the September fixture capture.
- `test_v2_rollback_session` still fails. It guards prior teaching against a
  published baseline, and both delivered chapter rewrites trip it. Re-baselining
  that guard is a decision for the user, so it was left alone rather than
  extended to bless the rewrites.
