docs(sharing): name the screener's score row where the section is consented

`POST /api/mental-health/assessments` is fenced on `mind` and writes one row
into `measurements`: the instrument's total, minted `COMPUTED` and keyed on the
administration it came from, so the trend reads through the chart and rollup
tier without the encrypted item answers going anywhere. A grant scoped to mood
and mind therefore adds a point to the owner's measurement series, and nothing
said so.

Left in place rather than fenced. Withholding the projection the way the
booster mint is now fenced would file a screener whose own total is missing
from the record — an inconsistency inside the section the grant does open, to
avoid a derived point in one it does not. So it is documented instead, in the
three places it is decided or agreed to: the declaration, the section's own
definition, and the consent copy in all seven locales, which is where a
widening belongs. The published operation says it too.

docs/api/openapi.yaml

@@ -14734,10 +14734,17 @@ paths:
       tags:
         - Mental health
       summary: Record a completed screener
-      description: Stores one administration (item answers encrypted) and writes the derived total as a server-owned
-        PHQ9_SCORE / GAD7_SCORE / WHO5_SCORE / SCI_SCORE measurement (source COMPUTED — never client-writable). WHO-5
-        and SCI totals run HIGHER = better; their follow-up thresholds point downwards (≤ 50 / ≤ 16). On a positive
-        PHQ-9 item-9 the response carries the locale-aware crisis-resource set.
+      description: >-
+        Stores one administration (item answers encrypted) and writes the derived total as a server-owned PHQ9_SCORE /
+        GAD7_SCORE / WHO5_SCORE / SCI_SCORE measurement (source COMPUTED — never client-writable). WHO-5 and SCI totals
+        run HIGHER = better; their follow-up thresholds point downwards (≤ 50 / ≤ 16). On a positive PHQ-9 item-9 the
+        response carries the locale-aware crisis-resource set.
+
+
+        The projection row is the one write this operation makes outside its declared section: the operation is fenced
+        on `mind`, and the measurement it derives lands in `measurements`. It is deliberate and unconditional — a
+        screener whose own total is missing from the record would be an inconsistency inside the section the grant does
+        open. The consent copy for the mood-and-mind section names it.
       parameters:
         - *a6
       requestBody:

messages/de.json

@@ -10484,7 +10484,7 @@
       },
       "mind": {
         "label": "Stimmung und Psyche",
-        "detail": "Stimmungseinträge und der Verlauf der psychologischen Fragebögen."
+        "detail": "Stimmungseinträge und der Verlauf der psychologischen Fragebögen. Ein ausgefüllter Fragebogen legt seinen Summenwert zusätzlich als Messwert ab."
       },
       "cycle": {
         "label": "Zyklus",

messages/en.json

@@ -10484,7 +10484,7 @@
       },
       "mind": {
         "label": "Mood and mind",
-        "detail": "Mood entries and the history of the mental-health questionnaires."
+        "detail": "Mood entries and the history of the mental-health questionnaires. A completed questionnaire also files its total score as a reading."
       },
       "cycle": {
         "label": "Cycle",

messages/es.json

@@ -10484,7 +10484,7 @@
       },
       "mind": {
         "label": "Ánimo y bienestar mental",
-        "detail": "Las entradas de ánimo y el historial de los cuestionarios de salud mental."
+        "detail": "Las entradas de ánimo y el historial de los cuestionarios de salud mental. Un cuestionario completado también registra su puntuación total como una medición."
       },
       "cycle": {
         "label": "Ciclo",

messages/fr.json

@@ -10484,7 +10484,7 @@
       },
       "mind": {
         "label": "Humeur et santé mentale",
-        "detail": "Les entrées d'humeur et l'historique des questionnaires de santé mentale."
+        "detail": "Les entrées d'humeur et l'historique des questionnaires de santé mentale. Un questionnaire rempli enregistre aussi son score total comme une mesure."
       },
       "cycle": {
         "label": "Cycle",

messages/it.json

@@ -10484,7 +10484,7 @@
       },
       "mind": {
         "label": "Umore e benessere mentale",
-        "detail": "Le voci sull'umore e lo storico dei questionari sulla salute mentale."
+        "detail": "Le voci sull'umore e lo storico dei questionari sulla salute mentale. Un questionario compilato registra anche il suo punteggio totale come una misurazione."
       },
       "cycle": {
         "label": "Ciclo",

messages/ko.json

@@ -10484,7 +10484,7 @@
       },
       "mind": {
         "label": "기분과 마음",
-        "detail": "기분 기록과 정신건강 설문 이력이에요."
+        "detail": "기분 기록과 정신건강 설문 이력이에요. 작성한 설문은 총점도 측정값으로 함께 저장돼요."
       },
       "cycle": {
         "label": "생리 주기",

messages/pl.json

@@ -10484,7 +10484,7 @@
       },
       "mind": {
         "label": "Nastrój i zdrowie psychiczne",
-        "detail": "Wpisy o nastroju i historia kwestionariuszy zdrowia psychicznego."
+        "detail": "Wpisy o nastroju i historia kwestionariuszy zdrowia psychicznego. Wypełniony kwestionariusz zapisuje też swój wynik łączny jako pomiar."
       },
       "cycle": {
         "label": "Cykl",

src/app/api/mental-health/assessments/route.ts

@@ -159,6 +159,17 @@ async function postAssessment(request: NextRequest): Promise<Response> {
   // into the record: the ordinary way this instrument is used when the person
   // it is about cannot run their own record. The row records what the person
   // in the room reported; the audit row says who typed it.
+  //
+  // It writes ONE row outside its declared section, and the crossing is
+  // deliberate rather than an oversight: the `*_SCORE` Measurement below is a
+  // projection of this administration, minted `COMPUTED`, keyed on
+  // `assessment:<id>`, and it exists only so the trend reads through the chart
+  // and rollup tier without ever touching the encrypted item content. Fencing
+  // it on `measurements` the way the booster mint is fenced would file a
+  // screener whose own total is missing from the record — an inconsistency
+  // inside the section the grant DOES open, to avoid a derived point in one it
+  // does not. The consent copy for this section says so, which is where a
+  // widening belongs.
   const { user, actor } = await requireRecordAuth("manage", "mind");
 
   // Opt-in module (default OFF) — same gate as GET.

src/lib/openapi/routes/mental-health.ts

@@ -135,7 +135,8 @@ export const mentalHealthPaths: NonNullable<ZodOpenApiObject["paths"]> = {
       tags: ["Mental health"],
       summary: "Record a completed screener",
       description:
-        "Stores one administration (item answers encrypted) and writes the derived total as a server-owned PHQ9_SCORE / GAD7_SCORE / WHO5_SCORE / SCI_SCORE measurement (source COMPUTED — never client-writable). WHO-5 and SCI totals run HIGHER = better; their follow-up thresholds point downwards (≤ 50 / ≤ 16). On a positive PHQ-9 item-9 the response carries the locale-aware crisis-resource set.",
+        "Stores one administration (item answers encrypted) and writes the derived total as a server-owned PHQ9_SCORE / GAD7_SCORE / WHO5_SCORE / SCI_SCORE measurement (source COMPUTED — never client-writable). WHO-5 and SCI totals run HIGHER = better; their follow-up thresholds point downwards (≤ 50 / ≤ 16). On a positive PHQ-9 item-9 the response carries the locale-aware crisis-resource set.\n\n" +
+        "The projection row is the one write this operation makes outside its declared section: the operation is fenced on `mind`, and the measurement it derives lands in `measurements`. It is deliberate and unconditional — a screener whose own total is missing from the record would be an inconsistency inside the section the grant does open. The consent copy for the mood-and-mind section names it.",
       parameters: [idempotencyKeyParameter],
       requestBody: {
         content: {

src/lib/sharing/scope.ts

@@ -68,7 +68,16 @@ export const SHARE_DOMAINS = [
   "profile",
   /** Illness episodes and the day logs under them. */
   "illness",
-  /** Mood entries and mental-health screener history. */
+  /**
+   * Mood entries and mental-health screener history.
+   *
+   * One write in this section lands outside it, and the consent copy names it:
+   * completing a screener also files the instrument's total as a `COMPUTED`
+   * measurement, so the trend reads through the chart and rollup tier without
+   * the encrypted item content going anywhere. It is a projection of the
+   * administration rather than a second reading, and withholding it would file
+   * a screener whose own total is missing from the record.
+   */
   "mind",
   /** Cycle tracking. */
   "cycle",