Record that the schema checks change where they run, not what the library accepts

docs/decisions.md

@@ -775,6 +775,10 @@ and no spelling holds; nesting depth; output length; what a portable line can sp
 reader's refusal of markdown or HTML it cannot read. A schema refusal a converter gains later is
 made in the first step too.
 
+The maintainer's reason (2026-11-04): the library's behaviour in and out stays as it is; the
+checks run at the first step and answer the same whatever format the ADF converts to. The rule adds
+no refusal for a shape no writer refuses today.
+
 The earlier entries premised that the site a document is saved to validates the rest. A
 hand-built document has not reached the site yet, and the viewer/editor persona relies on a
 refusal arriving before the save.

todo.md

@@ -124,6 +124,12 @@ reading; `MIGRATION.md` says to convert it once with the last version that reads
 `CHANGELOG.md` and `MIGRATION.md`. Search the writers for any other refusal of a shape the schema
 forbids, and `carriedNode` for one the readers still build.
 
+The maintainer wants the behaviour in and out unchanged, with the checks alike in every format. The
+first step cannot tell which node one format's carry holds, so a shape that one writer refuses now
+is refused in all, a carried node included. That is the one change of behaviour: confirm it with
+the maintainer before the chunk ships, and file a `question` item if they want the carried case
+kept as it is.
+
 Stays where it is: a node holding one mark type twice (the schema allows it, no spelling holds it),
 nesting depth, output length, a portable task list holding what no list item spells, and a
 reader's refusal of markdown or HTML it cannot read. A node type, mark or attribute the schema does