manuscript-canon-ledgerlisted
Install: claude install-skill dinaf2026-web/manuscript-editorial-suite
# Manuscript Canon Ledger — The Living Spine of a Series
A snapshot tells you what canon *is* in one book. This ledger tells you what canon
*is across the series* and **which changes were on purpose.** That distinction is
the whole point: it lets `manuscript-continuity-audit` flag only unintentional
drift instead of drowning the author in intentional evolution (a character aging, a
relationship deepening, a secret shifting).
## STEP 0 — Resolve the project
From the current working directory, walk **up** for `.manuscript/`. The ledger
lives at the path the profile's `ledger_path` names (default
`.manuscript/canon-ledger.json`), relative to that project root. If there is no
`.manuscript/` yet, invoke `manuscript-profile-setup` first (it creates the folder
and seeds an empty ledger).
The JSON is the source of truth. The xlsx is a one-way view — never edit the xlsx
as the authority; edit the JSON, then re-render. Keep both inside the project's
`.manuscript/` folder so they travel with the manuscript.
## Schema (three layers)
1. **`canon`** — current state. Each entity has attributes; each attribute has:
- `value` — the current canonical fact
- `status` — `STABLE` (safe), `DISPUTED` (sources disagree; see `governed_by` OQ),
`EVOLVING` (changes on purpose; see `governed_by` delta)
- `baseline` / `since` — where it was established
- `conflict` / `note` — what diverges and where
- `governed_by` — the OQ or delta id that rules this attribute
2. **`deltas`**