docs(metadata-protocol): name each decline list's remit and close the JSDoc's flow-conflict gap (#9271) (#9556)

`migrateStoredMetadata` is described by two hand-written decline lists that
were not the same enumeration, and nothing enforced either count.

- `protocol.ts` JSDoc: add the missing "a flow whose rename the conflict
  guard refused" bullet. That path is internal to the function and reachable
  from every door, so its absence was a real omission rather than a scope
  choice.
- `protocol.ts` JSDoc: the remit sentence PR #9270 landed closed with "that
  does not by itself account for every difference ... see #9271". With the
  bullet above added, it now does, so the clause is restated as the finished
  fact it has become.
- `cli.mdx`: name the decline table's remit — the operator-observable
  surface — and say which single member the JSDoc carries that it does not,
  and why no operator door can reach it.

The two lists now differ by exactly one member, in one direction, for a
stated reason. Prose only; no behaviour change.


Claude-Session: https://claude.ai/code/session_017qYPmkKEsfbWY1yVg83p8F

Co-authored-by: Claude <noreply@anthropic.com>

Author: os-sam

content/docs/deployment/cli.mdx

@@ -856,7 +856,12 @@ is `migrate-stored`, so a later diff shows which changes were an upgrade and
 which were somebody's edit.
 
 What it deliberately declines, and names in the report rather than
-counting as done:
+counting as done. This table is the **operator-observable** surface — what a
+run can actually report, from the command above or the route below. The
+function's own JSDoc in `packages/metadata-protocol/src/protocol.ts` documents
+its full internal surface instead, and so lists one decline more: flow rows
+skipped for want of a reachable automation engine, which neither operator door
+can produce, because both supply a live one.
 
 | Not rewritten | Why |
 | :--- | :--- |

packages/metadata-protocol/src/protocol.ts

@@ -13183,20 +13183,31 @@ export class ObjectStackProtocolImplementation implements
      *
      * ## What it declines to touch, and says so
      *
-     * This section documents the function's internal surface, which is not
-     * always identical to what an operator running the CLI can observe:
+     * This section documents the function's FULL internal surface, which is
+     * not always identical to what an operator running the CLI can observe:
      * `os migrate meta --stored` always passes its own automation engine
      * (see `canonicalizeFlow` above), so the first bullet below is never
-     * observed from that door. That does not by itself account for every
-     * difference between this list and `content/docs/deployment/cli.mdx`'s
-     * decline table — see #9271.
+     * observed from that door. That is now the only difference between this
+     * list and `content/docs/deployment/cli.mdx`'s decline table, which
+     * documents the operator-observable surface and so carries the other
+     * four. #9271 ruled that split intentional: keep each list correct for
+     * its own audience rather than reconciling them.
      *
      * - **`flow` rows with no reachable automation engine.** Flow-node
      *   conversions carry ADR-0078's open-namespace conflict guard, which
      *   needs the engine's live executor registry. When one is reachable —
      *   passed as `canonicalizeFlow`, or resolved from the services registry
      *   (#4498) — flows are migrated like anything else (#4454); when none is,
      *   they are reported `skipped` with that reason, never counted done.
+     * - **A flow whose rename that guard refused.** With an engine reachable,
+     *   the same open-namespace check can still find the old node-type token
+     *   is a LIVE name something else owns here. The conversion is refused and
+     *   the row reported `failed`, naming the token and its path: rewriting
+     *   would clobber that owner, and a quiet skip would hide it. Unlike the
+     *   bullet above this one IS reachable from every door — the CLI boots an
+     *   inert engine to hold the registry, and `POST /meta/_migrate-stored`
+     *   resolves a live one from the services registry — so
+     *   `content/docs/deployment/cli.mdx` carries it too.
      * - **Types with no repository write path** (neither `allowOrgOverride` nor
      *   `allowRuntimeCreate`). This pass declines them, and since #5086 it
      *   would have no choice: `saveMetaItem` refuses a code-only type with