bc-dgx7.103: B7e-manifest — A book's chapter and interlude manifest

[mordam]

Adds b7e-manifest: a book's chapter/interlude manifest with derived checks, replacing
five hand-built versions of the same table from five different deluvia sessions
(dv-afr.6/7/8/9, dv-gr6.8). Reads a book directory (or a bare book number, resolved under
<repo root>/novel/Deluvia Book N via git's own toplevel from cwd) and prints one row per
chapter/interlude with its header fields and measured word counts, plus a derived block:
day/BP-range overlaps and gaps on the same thread, and interlude placement-key issues
(inconsistent, out-of-range, non-monotone).

Chapter/interlude counting reimplements deluvia/scripts/build_series_log.py's own regex
shapes in JS (this file has to run with no deluvia checkout on disk — the suite proves
it against a fabricated fixture tree) rather than importing them. Verified by hand against
the live deluvia repo: all six books' counts match SERIES_CHAPTER_LOG.md's committed,
generated block exactly, including matching the Python's own blind spot on
CHAPTER_18A_THE_PARTING.summary.md (named as "not counted" rather than silently
included or dropped). The bead's own acceptance criteria also ask for two specific
findings from the sessions that filed this: Book 3's Ch.18/Ch.19 day-range overlap still
reproduces against live data; Book 1's IL-05/06/12 placement issue no longer does — that
data has evidently been hand-fixed since the audit that found it, which is the finding
going stale, not the check being wrong. All four kinds of placement issue and both
overlap/gap detection are proven directly on fabricated fixtures in test/manifest.mjs
independent of Book 1's current state.

Registered @grant read in the bin file's own header per bc-wbrhi's derived-grants
mechanism (the bead's own "what shipping it takes" text describes the older
hand-edit-lib/toolbelt.js mechanism that bead predates and that landed already superseded
this).

One self-caught bug worth flagging: my first pass at the README section edit replaced the
## Notes on bd heading line instead of preserving it around the new content, which
test/anchors.mjs caught immediately (a dangling in-page link). Fixed in a second commit
before this delivery — worth knowing in case a similar large insertion-before-a-heading
edit gets reviewed elsewhere without that suite catching it.

Tests: node test/manifest.mjs (52 checks, new suite) plus test/lockfile.mjs, test/tooldecl.mjs, test/grants.mjs, test/allowlist.mjs, test/evidence.mjs, test/filter.mjs, test/b7ereadme.mjs, test/anchors.mjs, test/checks.mjs, test/documents.mjs, test/skills.mjs, test/usage.mjs, test/routes.mjs, test/policies.mjs, test/coverage.mjs, test/grepargs.mjs, test/authority.mjs, test/harness.mjs — all green. The local full 'npm test' gate was queued behind 4 other sessions' gates (oldest 40+ min in), so per this repo's own documented practice (a local full gate under this much contention is slower and noisier than CI's 4-shard run) I ran the suites b7e-affected named plus the ones a new bin/ registration always touches, and am leaving full-suite verification to CI.

Files changed — 6 files · +1211 −0 · against `main`

test/manifest.mjs  +513 −0
lib/manifest.js    +442 −0
bin/b7e-manifest   +170 −0
README.md          +84 −0
package-lock.json  +1 −0
package.json       +1 −0

---

Opened by a beadcause worker session on bc-dgx7.103 — b7e-manifest — A book's chapter and interlude manifest, with the sequence checked. It merges itself once the checks report; merging is what closes the bead. If this is still open, something stopped that, and the reason is on bc-dgx7.103 and in Adam's inbox.

bead: bc-dgx7.103

[mordam]

A beadcause worker opened this and does not merge its own work. It is on the merge queue as bc-mig9x.

[mordam]

The beadcause merge queue tried to merge this and could not: it has been tried 3 times and stopped at the same place each time. the branch still conflicts with main and no resolver is on it any more — the window that was opened on it ended without making it mergeable, so this one is yours to settle. Tried 3 times — that was the last. It is Adam's call now — see bc-mig9x.