Wrapper-unit insert seams should generalize from bulleted-list to compatible list containers before adding new geometry

What happened

The next sharp stress case was not a new document geometry at all.

It was a container variant:

• ordered-list
• with the same list-item unit contract already proved for bulleted-list

Plain copy/paste round-tripped, but explicit-at selection and range-ref rebasing fell through the wrong path because the seam was still special-cased to bulleted-list.

What fixed it

The durable fix was to generalize the wrapper-unit insert seam from:

• one hardcoded bulleted-list

to:

• compatible list containers that share the same list-item unit contract

For the current proved subset, that means:

• bulleted-list
• ordered-list

Why this works

This was not new geometry.

The unit contract was unchanged:

• same sibling unit type
• same wrapper-list behavior
• same post-insert selection/range semantics

Only the container type changed.

That should extend the existing seam, not create a new one.

Reusable rule

Before treating a new container as a new geometry problem, ask:

• does it reuse the same sibling unit contract?

If yes, widen the seam before designing a new model.

Related issues

• 2026-04-05-list-unit-fragment-proofs-should-treat-list-item-fragments-as-sibling-units-and-assert-real-paste-landings.md