# Tara selection and Eve release integration · 8 September 2026

The current library combines [Tara selection](../../decks/tara-selection.html),
[Tara–Lilith handoff](../../decks/tara-lilith-handoff.html), and the new
release-evidence page in [Eve operations](../../decks/eve-operations-detail.html).
It contains 70 guides, 791 unique slides, 814 guide-slide instances and 416 cited
paths. All 791 recordings remain unchanged; integration generates no speech.

Tara delivery: `e4351bde94eb4486c520b0cde9b0012c5cdcc364`. Incoming main:
`6c62dfaed8fbdfe5c439e60faae20ab7abcbe228`, including Eve release commit
`6065e42ad81a572eb06c7d915981fda9f2591e60` and the separate computer-use change.
Main advanced during narration. The merge preserves both complete contributions
and the unrelated incoming Docs Center/native-fixture work.

## Current and reused evidence

[Center review](center.json) covers all three guides and all 63 pages in two
embedded contexts, six audio boundaries, navigation, search, history, mobile
focus and PDF export. [PDF text](pdf-text.json) checks every ordered visible
field. The integrated presentation, preservation and affected Eve-assignment
suite passes 142 tests; freshness and source accounting also pass.

[Reuse](reuse.json) compares all 70 guide HTML files, three affected manuscripts,
four shared UI files, both parents’ scripts/manifest entries and all 791 MP3
hashes. The Tara parent has 790 recordings and the incoming parent has 747;
the 746 common entries agree. Ten Tara source/test files also match the new
baseline. Tara’s 316-check browser review and the incoming Eve visual review
remain evidence for their unchanged slide content. New center checks cover the
combined catalog. [Raster comparison](raster-comparison.json) binds the actual
exports; all 44 Tara pages retain identical pixels. The final 19 Eve pages are
separately covered by [optical review](optical-review.json).

The [Tara receipt](../v1-tara-selection-2026-09-08/README.md) and
[Eve receipt](../eve-operations-release-assignment-2026-09-08/README.md) are
historical at their owning commits. Their global catalog hashes intentionally
precede this integrated library. Use this receipt for the current catalog.

## Integration mechanics and limits

The first merge reported a missing object and entered a long partial-clone
promisor-cache scan. The object was verified present locally. Two task-owned
merge attempts were stopped before they changed the worktree. A command-local
`remote.origin.promisor=false` override allowed the third merge to expose its
ordinary content conflicts; repository configuration was unchanged. Both delivery
notices and narration sets were retained, generated indexes rebuilt, and the
preservation test advanced to the incoming baseline. Automatic Git maintenance
finished before the integration browser gate. These are repository mechanics,
not new product acceptance. Logs retain attempts separately from passing checks.

The normal commit hooks also needed sparse-checkout setup for the incoming
files: eight staged text files, omitted project descriptors, the native
binding sources and its configuration/logging dependencies were materialized
from the index. The existing configuration package dependency installation was
linked locally. No tracked application code was changed for this setup. The
affected native package typecheck then passed with parallelism one; this is
typechecking evidence, not native fixture or device acceptance. Failed setup
attempts are retained alongside the successful typecheck log.

Run `python3 docs/presentations/presentation-center/verification/v1-tara-selection-integration-2026-09-08/verify-receipt.py`
from the repository root. Later library edits invalidate the frozen hashes.
The [full V1 program](../../authoring/v1-completion.md) stays active. Human reader
or listening acceptance, live member/device behavior and clinical acceptance
were not performed. Native fixture acceptance belongs to its incoming commit;
this receipt covers the presentation integration.
