LLL Content Snapshot Diff / 公開状態スナップショット差分

FAQ

Does it change entries? No. It reads and shows; it never writes, sends or stores anything. / 変更しません。読んで表示するだけで、書込・送信・保存はしません。

What is read? Collection, id, locale and slug of published entries in visible (not hidden) collections. Titles, body, SEO, dates, authors and drafts are not read into the snapshot or report. / 表示対象コレクションの公開記事のコレクション・id・言語・スラッグだけ。題名・本文・SEO・日時・作成者・下書きは対象外です。

Does "no longer published" mean deleted? No. The entry is simply not in the published set now: it may be deleted, unpublished, moved or in a collection that is now hidden. The tool cannot tell which. / 削除とは限りません。いま公開済みの中に見当たらないという意味で、削除・非公開化・移動・コレクションの非表示のどれかです。区別はできません。

What is "unverified"? One side was not read in full (a limit, a read failure, a hidden collection, or pasted text that was cut off). Those rows may or may not be real changes, so they are never called removed or added. / 片方を最後まで読めていない行です（上限・読取失敗・非表示コレクション・貼り付けの途中切れ）。実際の変化かは分からないので「消えた」「増えた」とは言いません。

Can I compare two different sites? No. Ids belong to one site. If no id is shared, the tool shows a warning instead of rows. / できません。id はサイトごとに異なります。共通の id が無いと、行の代わりに警告を出します。

How big can a snapshot be? About 1,000 entries (60 KB). Larger sites get an INCOMPLETE snapshot; entries that did not fit are never reported as removed. / 約1,000件（60KB）まで。それ以上は不完全なスナップショットになり、入りきらなかった分を「消えた」とは報告しません。

Drafts and scheduled posts? Not included; only published entries. / 含みません。公開済みだけです。

Security

LLL Inc. / MIT / contact@lx3.cafe. Version 0.1.0.

Capabilities: content:read and schema:read only. One POST admin route requiring plugins:manage and a signed-in caller. No network, hooks, cron, storage, settings or writes of any kind. No signing data or secrets are used.

Data: only collection slug, entry id, locale and slug of published entries in visible collections are read; titles, body, SEO, dates and personal data are not read into output. The snapshot is plain text in a code block; fields are percent-encoded so tabs, newlines, control characters and spreadsheet-formula leads (= + - @) cannot appear raw, and the pasted text is parsed strictly (fixed header, per-collection counts, end line; any malformed line is rejected without echoing its content). Report cells are sanitised the same way. Nothing is stored; copies in the browser are outside plugin control.

Limits: snapshot 60 KB, pasted text 62 KB, 3000 entries and 50 collections read; beyond these results are marked incomplete.

Report issues privately with version and a redacted reproduction. Do not send tokens or private exports.

Changelog

0.1.0

• First release: text snapshot of published entries (collection, id, locale, slug) and comparison with the live site: no longer published, newly published, slug changed, locale changed. Incomplete snapshots or reads are marked unverified instead of reported as removals. EN/JA.

Installation / 導入

Install from the EmDash Registry, then open Admin → Content Snapshot Diff. Requires permission to manage plugins. Capabilities: content:read, schema:read. No network access.

1. Press "Create snapshot" and copy the whole text into a file or note (this tool does not store it).
2. After an import, update or migration, open the page again, paste the text and press "Compare with now".

Use snapshots from the same site only: ids differ between sites.

Registry から導入し、管理画面の「Content Snapshot Diff」を開きます。プラグイン管理権限が必要です。権限は content:read / schema:read のみ。外部通信なし。

1. 「スナップショットを作る」を押し、出たテキスト全体をファイルやメモにコピーして保管します（このツールは保存しません）。
2. 取り込み・更新・移行のあとにもう一度ページを開き、テキストを貼り付けて「いまと比べる」を押します。

比較は同じサイトのスナップショット同士で行ってください。id はサイトごとに異なります。