# Assembled here, reviewed elsewhere · Tara · practise, chapter 4 · verification receipt · 11 September 2026

Scope: the fourth chapter of the _Tara · practise_ track. Thirty inherited
slides became thirty-eight: a cover, five section dividers, thirty-one content
slides — the thirty inherited and one new capture slide — and a close, in
sections of six, seven, five, six and seven. The
[brief](../../authoring/tara-ritual-assembly-chapter.md) records what the
chapter is for and why each treatment was chosen; the track brief is
[tara-practise-track.md](../../authoring/tara-practise-track.md);
[state.json](state.json) holds the exact counts, tool results and findings.

## What was checked

- **Build freshness.** `build-eve-oshun.py --check` passes on the delivered
  tree: 1,497 slides and 664 sources. The chapter is 38 slides running
  44.0 minutes, with three tables and nine diagrams.
- **Every value was executed.** The existing pinned fixture,
  `pinned-tara-ritual-assembly.test.mjs`, runs twenty cases against the pinned
  assembler, feature builder, contract, policy registry and tone reviewer, and
  every number on the slides comes from one of them or from a probe added for
  this rewrite; `state.json` lists the probes.
- **The capture.** `tara-studio-ritual-capture.mjs` renders the Lilith Studio
  Tara page from the pin over the three answers the pinned BFF sends when no
  member is signed in. It does not retype those answers: it evaluates
  `TARA_TODAY_FIXTURE`, `TARA_SITTING_CATALOG` and `TARA_RITUAL_DEFINITION` out
  of the pinned route file, quotes each route's anonymous branch, and fails if
  the declared fixture differs. The page's only replaced boundary is `bffGet`,
  which answers exactly the three declared paths and throws on any other.
  `test_product_captures.py` holds it to its pin, its bytes and its two claims.
- **Bindings.** `assignments/v1-tara-ritual-assembly.json` binds fifteen of
  these slides across two handbook units. All seventeen evidence entries carry
  fresh fingerprints, their explanations and both rationales are re-set to the
  rewritten slides, and `check-center-coverage.py --check` reports 347 reviewed
  and 0 stale.
- **Browser rendering.** `eve-edition-review.mjs --guide tara-ritual-assembly`
  sweeps all 38 slides at 1440×900, 1280×720 with audio chrome, 390×844 and in
  print — 152 measurements, no spill, no document overflow, no page errors, no
  findings.
- **Narration.** Thirty-eight tracks rendered on this host's CPU and verified
  with `--check --ids`; the library-wide `--check` verifies all 1,497 tracks,
  1,985.5 minutes, with nothing pending.
- **PDF.** 38 pages, exact authored visual text on every one, in order, read
  back from the downloadable `decks/tara-ritual-assembly.pdf`, and
  `check-pdf.py` passes, including its table-header size floor on all three
  tables.
- **Suites and lint.** Python 479 passed, 0 failed. Node 489 of 489.
  `ruff check` on the touched tests, ESLint on the new harness and Prettier on
  every touched file are clean.

## What executing the source found

- **A renamed template keeps a review its new policy would fail.** The review
  fingerprint covers the rendered step speech and not the title, but the title
  is one of the fields the policy resolver searches. A two-step breath template
  reviewed as _Morning grounding_ passes at 0.8355 under `breath-led-guide`.
  Renamed _Before sleep_, it still validates as active with that review, while
  a fresh review resolves `silent-witness` and fails at 0.6042. The active gate
  checks that a review covers the current speech, not that it was judged under
  the current policy.
- **The script numbers steps from `order + 1`.** The feature and contract
  renderers both assume steps are counted from zero; the domain assembler
  counts from one, and so does the review fixture, whose script opens "Step 2:
  Box breath". Renumbering identical speech from zero changes the fingerprint
  and makes the stored review stale.
- **Words pick the policy by substring.** "Presets for the morning" resolves to
  a transition because it contains "reset", and a passage step makes a ritual a
  dharma talk unless a sleep or transition word matched first — the one shape
  that requires tradition consent, judged under `bhakti-devotional` whatever
  tradition the metadata names.
- **Sleep descent drops a supplied journal without an error**, and a morning
  journal, optional for the moment, is emitted as a required step by default.

## What was fixed

- **Two policy identifiers.** The inherited policy-pack table named the
  breathing pack's teacher-safety and voice policies
  `tara-safe-teacher-boundary-v1` and `tara-breathing-voice-v1`; the pinned
  registry says `coach-practitioner` and `breathwork-coach`.
- **An overflow and two collisions.** The six-band tone-context stack ran 71px
  past the narration chrome at 1280×720 and became five bands, the last holding
  the three remaining outcomes in source order. In the contract-template graph
  two edge labels sat on each other; the steps moved above the hub. On the
  capture, the first pin placement covered the label it pointed at, and the
  pins now sit in empty space, checked at the size the slide shows them.

## What was not checked

- **No application path.** No converter between the three records, no caller
  of the assembler or the reviewer, no storage, publishing or launch. The
  bounded search that says so could miss a wrapper under another name.
- **No human review.** No editor, tradition authority, rights holder, member
  or listener took part; the tone verdict is a deterministic text heuristic
  with its pace axis skipped.
- **Only the anonymous Studio page.** The capture shows what the page renders
  from the BFF's anonymous answers. A signed-in author's continuation would
  differ, nothing in the web app links to the page, and the capture does not
  establish that no other authoring screen exists.
