@@ -25,6 +25,13 @@ name: Docs Drift Check
 # on one day (#11180, #11262). The fold was right and stays; what changed is that a run
 # with an unanchored file now says so where a reviewer reads it, and withholds its ✅.
 #
+# Since #11356 the ✅ is withheld from the last arm that carried it: "N anchor(s) derived,
+# no hand-written page names any of them". That sentence reports the NAMING RELATION, and
+# a page is listed only when it already names a changed token — so a PR that widens an
+# enumerable vocabulary (a function table, an allowed-value set) is structurally invisible
+# to it, because the new members' absence from the page is precisely the defect. ⚠️ The
+# narrowing does NOT catch that class; it stops the run from claiming it did.
+#
 # And since #9519 it states WHICH TREE it read. On `pull_request`, actions/checkout gives
 # the MERGE of base and head, so the row set is a fact about a commit that exists on no
 # branch a reader can name — while the comment's own re-derivation command sent them to
@@ -303,11 +310,31 @@ jobs:
               // ⛔ The ✅ is DROPPED whenever a file went unanchored, deliberately: a green
               // tick is the clean-bill glyph, and "no page names the anchors I derived" is
               // not a clean bill when some of the diff derived no anchor to name.
+              //
+              // ⛔ And since #11356 the ✅ is gone from this arm ENTIRELY — including the run
+              // where nothing went unanchored, which is where it used to live. "No page names
+              // the anchors I derived" is a fact about the NAMING RELATION, not about the
+              // docs: a page is listed because it ALREADY names a changed token, so a change
+              // that WIDENS an enumerable vocabulary can never be caught here — the new
+              // members' absence from the page IS the defect, and an absence names nothing.
+              // Measured on #11347 (six new flow-expression functions): 9 anchors derived, 0
+              // pages named them, ✅ — while `content/docs/automation/flows.mdx` carried a
+              // table enumerating the available bindings that the PR had just made
+              // incomplete, and a reviewing seat almost passed the PR on that tick.
+              //
+              // ⚠️ This does NOT catch that class, and must not be read as doing so. Nothing
+              // in the anchor model can (three in-model routes were measured to fail on the
+              // one real instance; the authoring-mark route that would catch it is escalated
+              // as #11817). What it stops is the run CLAIMING it did: the verdict now says
+              // only what it measured — anchors were derived, no page named them — and leaves
+              // the clean-bill glyph to a run that earned one.
               const headline = anchorList.length === 0
                 ? (anchorless.length
                     ? `${notCovered} Nothing else in this diff resolved to a documentable surface (no symbol, route or SDK anchor derived from **${pkgs.length}** changed package(s)).`
                     : `Nothing in this diff resolved to a documentable surface (no symbol, route or SDK anchor derived from **${pkgs.length}** changed package(s)), so **this run has no opinion** about the docs.`)
-                : `**${anchorList.length}** anchor(s) derived from **${pkgs.length}** changed package(s); no hand-written page names any of them.${anchorless.length ? ` ${notCovered}` : ' ✅'}`;
+                : (anchorless.length
+                    ? `**${anchorList.length}** anchor(s) derived from **${pkgs.length}** changed package(s); no hand-written page names any of them. ${notCovered}`
+                    : `**${anchorList.length}** anchor(s) derived from **${pkgs.length}** changed package(s); no hand-written page names any of them, so this run has **nothing to list** — **not a clean bill of health**. This check sees only pages that NAME a derived anchor: one that documents this change in prose, or enumerates it in an authoring dialect, names none and stays invisible to it on every run.`);
               // The tree identity is rendered here only when anchors WERE derived: that
               // is the run with an opinion ("no page names any of them"), and a reader
               // whose own tree does yield a row is owed the reason. With no anchors the
 
@@ -209,6 +209,9 @@ reader has to expand a fold to find is one a reader does not find. The failure #
 records was never the tool lying; it was the tool never signalling its own limits at the
 point of use.
 
+Since #11356 the same posture reaches the ✅ itself — see **The verdict when nothing names
+the anchors** below.
+
 ### Measured, before and after
 
 Ten real PRs, each re-derived at its own merge base with its own docs corpus. `docs` rows:
@@ -241,6 +244,33 @@ run says so and points at the coarse set — where the old tool's 22 rows were e
 mentioning `@objectstack/cli`. A CLI **command name** (`os meta resync`) is exactly the
 recall class the shape guard costs us: it is a lowercase word, so it cannot anchor.
 
+### The verdict when nothing names the anchors
+
+That state — anchors derived, nothing left unanchored, no page naming any of them — used
+to end the headline in a ✅. It no longer does (#11356). The sentence reports the **naming
+relation**, and the relation only ever lists a page that ALREADY names a changed token, so
+a PR that **widens an enumerable vocabulary** is invisible to it by construction: the new
+members' absence from the page is precisely the defect, and an absence names nothing.
+
+Measured on #11347 — six new flow-expression functions (`round`/`floor`/`ceil`/`abs`/
+`min`/`max`) — the run derived 9 anchors, matched 0 pages, and rendered the ✅, while
+`content/docs/automation/flows.mdx` carried a table enumerating the available bindings
+that the same PR had just made incomplete. A reviewing seat almost passed the PR on that
+tick.
+
+⚠️ **The narrowing does not catch that class.** Nothing in the anchor model can: anchoring
+the new members is impossible by construction, the siblings live in carriers the diff
+never touches, and the container symbol is named by no page. The authoring-mark route
+that would catch it is escalated as #11817. What the narrowing changes is that the run no
+longer **claims** it did — the verdict states only what it measured, and the clean-bill
+glyph is left to a run that earned one. The rendering is pinned in both directions by
+`check-drift-comment.mjs`, which asserts the verdict byte-exact on that state and asserts
+it ABSENT on all four neighbouring states.
+
+How often it renders, re-derived over the 40 first-parent commits ending at `e43b18fd9`:
+3 of the 17 package-touching runs (18%) — `20a452e664`, `f213793ddb`, `dd4113ec0b` — so it
+is a rare notice rather than a per-PR banner, which is what keeps it readable.
+
 **Cost** (the card's open question): the anchor derivation reads the same 178-page corpus
 the old one did, plus the 18 route-registrar/ledger sources (~875 KB) and one `git show`
 per changed file per side. Measured end-to-end on the ten PRs above, `node affected-docs.mjs`
 
@@ -35,6 +35,16 @@
  *
  *   diff → the real affected-docs.mjs → its real --json → the real comment script → bytes
  *
+ * Since #11356 it pins one more verdict the same way: the `anchors-derived-none-named`
+ * state — anchors derived, nothing unanchored, no page naming any of them — used to end
+ * in a ✅. That glyph read as "docs verified" on a change whose own page had just gone
+ * incomplete, because a page is listed only when it ALREADY names a changed token, so a
+ * PR that widens an enumerable vocabulary can never be caught by the relation (the new
+ * members' absence from the page IS the defect). The narrowing does not catch that class;
+ * it stops the run from claiming it did. Both readings are on the record below: the
+ * verdict is pinned byte-exact on that state and pinned ABSENT on all four neighbours,
+ * because a notice that renders everywhere reports nothing.
+ *
  * Each case declares the mapper facts it depends on (`anchors` / `anchorless` / `docs`)
  * and those are asserted BEFORE the rendered text is. Without that, a change in how the
  * mapper classifies a `.md` file would quietly move every case onto the same branch and
@@ -182,6 +192,15 @@ const NAMES_THE_ANCHOR = '# Guide\n\nThe cap is `MAX_WIDGETS`, and it bounds the
 const ADDS_ANCHOR = 'export function renderWidget(x: number) {\n  return x + 1;\n}\n\nexport const MAX_WIDGETS = 5;\n';
 const README_EDITED = '# @objectstack/demo\n\nRun `os demo` to start. Then run `os demo studio` for the UI.\n';
 
+/** The #11356 verdict, byte-exact — rendered by exactly one of the five cases below. */
+const ANCHORS_DERIVED_NONE_NAMED =
+  '**1** anchor(s) derived from **1** changed package(s); no hand-written page names any of them,'
+  + ' so this run has **nothing to list** — **not a clean bill of health**. This check sees only'
+  + ' pages that NAME a derived anchor: one that documents this change in prose, or enumerates it'
+  + ' in an authoring dialect, names none and stays invisible to it on every run.';
+/** The phrase that tells that verdict apart from every neighbouring one. */
+const NARROWED_PHRASE = 'nothing to list';
+
 /**
  * `want` is the mapper contract each case rides on — asserted before any text is, so a
  * case that silently stopped exercising its branch fails here instead of passing there.
@@ -193,6 +212,7 @@ const CASES = [
     base: { 'content/docs/guide.mdx': NAMES_NOTHING },
     change: { 'packages/demo/README.md': README_EDITED },
     want: { anchors: 0, anchorless: ['packages/demo/README.md'], docs: 0 },
+    narrowed: false,
     expect: (headline, body) => {
       check('readme-only', 'the headline names the blind spot', true, headline.includes('NOT COVERED by this run'));
       check('readme-only', 'the headline names the file that went unanchored', true, headline.includes('packages/demo/README.md'));
@@ -203,15 +223,22 @@ const CASES = [
   },
   {
     id: 'anchored-source',
-    what: 'an anchorable source file, no page naming it — the ✅ arm',
+    what: 'an anchorable source file, no page naming it — the anchors-derived-none-named arm (#11356)',
     base: { 'content/docs/guide.mdx': NAMES_NOTHING },
     change: { 'packages/demo/src/widget.ts': ADDS_ANCHOR },
     want: { anchors: 1, anchorless: [], docs: 0 },
+    narrowed: true,
     expect: (headline) => {
       check('anchored-source', 'the blind-spot sentence does NOT render — it is a report, not boilerplate',
         false, headline.includes('NOT COVERED'));
-      check('anchored-source', 'the headline is the untouched ✅ text', true,
-        headline === '**1** anchor(s) derived from **1** changed package(s); no hand-written page names any of them. ✅');
+      check('anchored-source', 'the headline is the narrowed #11356 verdict, byte-exact', true,
+        headline === ANCHORS_DERIVED_NONE_NAMED);
+      // The point of the narrowing, asserted as its own reading: this state no longer
+      // carries the clean-bill glyph. Byte equality above would catch a ✅ appended to
+      // the end, but not one moved elsewhere in the line.
+      check('anchored-source', 'the clean-bill glyph is withheld', false, headline.includes('✅'));
+      check('anchored-source', 'and the verdict says what it could not see, not just what it found', true,
+        headline.includes('not a clean bill of health'));
     },
   },
   {
@@ -220,6 +247,7 @@ const CASES = [
     base: { 'content/docs/guide.mdx': NAMES_NOTHING },
     change: { 'packages/demo/src/widget.ts': ADDS_ANCHOR, 'packages/demo/README.md': README_EDITED },
     want: { anchors: 1, anchorless: ['packages/demo/README.md'], docs: 0 },
+    narrowed: false,
     expect: (headline) => {
       check('anchored-and-anchorless', 'the derived anchors are still reported', true, headline.startsWith('**1** anchor(s) derived'));
       check('anchored-and-anchorless', 'and the unanchored file is reported beside them', true, headline.includes('NOT COVERED by this run'));
@@ -232,6 +260,7 @@ const CASES = [
     base: { 'content/docs/guide.mdx': NAMES_THE_ANCHOR },
     change: { 'packages/demo/src/widget.ts': ADDS_ANCHOR },
     want: { anchors: 1, anchorless: [], docs: 1 },
+    narrowed: false,
     expect: (headline, body) => {
       check('docs-listed', 'the work-list headline is unchanged', true,
         headline === 'This PR changes **1** package(s): `@objectstack/demo`, touching **1** documentable anchor(s).');
@@ -244,6 +273,7 @@ const CASES = [
     base: { 'content/docs/guide.mdx': NAMES_NOTHING },
     change: { 'content/docs/guide.mdx': `${NAMES_NOTHING}\nOne more line.\n` },
     want: { anchors: 0, anchorless: [], docs: 0 },
+    narrowed: false,
     expect: (headline) => {
       check('no-package-change', 'the original headline is preserved byte-for-byte', true,
         headline === 'Nothing in this diff resolved to a documentable surface (no symbol, route or SDK anchor derived from **0** changed package(s)), so **this run has no opinion** about the docs.');
@@ -275,6 +305,11 @@ try {
     check(c.id, 'the comment keeps its heading', '### 📓 Docs Drift Check', lines[1]);
     const headline = lines[2];
     if (PRINT) console.log(`\n── ${c.id} — ${c.what}\n${headline}\n`);
+    // Non-vacuity for the #11356 verdict, asserted in the LOOP rather than per case, so a
+    // sixth case cannot be added without declaring which side of the line it sits on.
+    if (typeof c.narrowed !== 'boolean') throw new Error(`case ${c.id} does not declare \`narrowed\``);
+    check(c.id, `the #11356 narrowed verdict ${c.narrowed ? 'renders here' : 'does NOT render here'}`,
+      c.narrowed, headline.includes(NARROWED_PHRASE));
     c.expect(headline, body);
   }
 } finally {