Answer prose round 1: 2 🔴, 7 🟡 (0 misread); retitle the decision, state its rule, and fix the carry claims

spec/flavour.md

@@ -582,7 +582,9 @@ open or close where the run sits (`un**-real**istic`), or one CommonMark's match
 the merged run's pairing to another delimiter — rides the inline carry whole. So does every node
 holding `em`, `strike` or `strong` in a line still read back wrong after 32 tries. An opaque carry inside
 a mark spelling is a named error in input: the carry restores its node exactly, marks included
-(`docs/decisions.md` §Unknown nodes ride the carry).
+(`docs/decisions.md` §Unknown nodes ride the carry). An inline card a link marks rides no carry:
+`adfToLosslessMarkdown` refuses it as `unsupported-node-shape`, and so does `losslessMarkdownToAdf`
+reading a carry that holds one, since Atlassian's schema gives a card no link mark.
 
 ```
 !adf:textColor[**Overdue**]{color="#ae2e24"}, H!adf:subsup[2]{type=sub}O, !adf:underline[signed].

spec/html.md

@@ -127,8 +127,8 @@ later release writes it as its own element.
   - a list holding a node that is no `listItem`, or an item holding a mark, text, an empty key or
     an attribute its type does not hold: the whole list rides the carry;
   - a `listItem` outside a list, an inline node in block position and a block node inline.
-- An inline card a link marks is refused as `unsupported-node-shape` wherever it stands, carried
-  or not: Atlassian's schema gives a card no link mark.
+- An inline card a link marks is refused as `unsupported-node-shape`, carried or not, by `adfToHtml`
+  and by `htmlToAdf` reading a carry element: Atlassian's schema gives a card no link mark.
 - A block carry's JSON nests at most 500 levels less the depth it stands at, an inline carry's at
   most 500, as in markdown.